Keywords: Robotics
Abstract: Robotic leg prostheses and exoskeletons have not been widely adopted despite being first commercialized two decades ago. Many design challenges related to weight, noise, and rigidity are now being resolved with better actuators, but control challenges limiting the versatility, reliability, and effectiveness of these devices remains a major roadblock to adoption. Recent advances in legged robots might suggest that sim-to-real Reinforcement Learning is the answer, but the human part of the system cannot be reliably simulated, measured, or controlled like an autonomous robot. This talk will describe how classical control approaches for legged robots—which have been effectively replaced by computational methods—are now providing the basis for versatile and reliable control of wearable robots. One key concept is the use of a phase variable to parameterize the gait cycle in a time-invariant manner, allowing continuous synchronization of a prosthetic leg to the amputee user. We extend this control paradigm to continuous variations of the primary activities of daily life, enabling translation to the Össur Power Knee for improved amputee outcomes. Another key concept is the use of energy shaping to augment body dynamics, allowing exoskeletons to enhance voluntary joint motion with safety guarantees. With this task-agnostic control approach, lower-limb exoskeletons can reduce muscle effort and joint moments across daily activities to mitigate muscular fatigue or manage osteoarthritic joint pain. While the field is beginning to see positive outcomes in diverse real-world contexts, many outcomes remain elusive and may require new methods to optimize versatile controllers for specific users.

Keywords: Control Applications, Optimal Control, Uncertain Systems and Robust Control
Abstract: This paper presents a direct data-driven predictive controller for layer-to-layer feedback control of direct-write additive manufacturing (DWAM). We compare our direct data-driven controller with three other control approaches: open-loop, ideal model-based, and indirect data-driven. The open-loop controller represents the current state-of-the-art for many printers that lacks layer-to-layer feedback. The ideal model-based predictive control assumes perfect knowledge of the DWAM dynamics to quantify best-case performance. The indirect data-driven predictive controller uses system identification to learn a model from data, which is then used in the model-based predictive controller. In contrast, our novel direct data-driven approach integrates data directly into the optimal control problem, bypassing the need for a model. We present a theorem showing that our direct approach achieves the same optimal performance as the ideal model-based approach. Furthermore, we conduct a numerical study to compare the robustness of each approach, with the results demonstrating the advantages of our direct data-driven controller.

Keywords: Cyber physical systems, Control Applications, Manufacturing Systems
Abstract: E-jet printing is a promising additive manufacturing technique for micro/nano manufacturing. It enables manufacturing complex geometries that are challenging for traditional manufacturing processes. Several processes and environmental factors influence printing quality. Maintaining consistent, high-quality printing under varying vibrations and environmental conditions is a significant challenge due to the lack of reliable sensing techniques for real-time E-jet printing control. A sensing approach has been developed to accurately quantify the nozzle-to-substrate standoff distance during E-jet using OpenCV to address this. Then real-time controller is integrated with the sensing feedback to compensate for the environmental factors to perform reliable printing with uniform line width. Experimentally, the proposed machine vision sensing approach was integrated with PID control for E-jet printing subject to external vibrations, and reliable printing performance was achieved.

Keywords: Manufacturing Systems, Modeling and Validation
Abstract: Electrohydrodynamic jet (e-jet) printing has emerged as a versatile micro/nano scale additive manufacturing (μ/n-AM) technique known for its precision and compatibility with a wide array of materials, enabling advanced applications in optics and flexible electronics. Despite its capabilities, a critical gap remains in linking droplet deposition dynamics to final printed feature quality, particularly for drop-on-demand (DoD) processes. In this study, high- fidelity numerical simulations using COMSOL Multiphysics are employed to bridge this gap by accurately capturing the underlying physics governing droplet deposition and subsequent structural formation. Moreover, in this study we investigate the influence of key ink properties, including viscosity, density, and surface tension, on the post-deposition morphology of ink material on the substrate. Simulation results demonstrate the direct correlation between viscosity and the evolution of surface roughness over time, as well as the significant role surface tension plays in determining both planar and vertical printing resolutions. These findings offer critical insights, enabling more informed selection of ink parameters tailored to specific structural and precision manufacturing objectives.

Keywords: Manufacturing Systems, Optimal Control
Abstract: The manufacturing industry is under growing pressure to enhance sustainability while preserving economic competitiveness. As a result, manufacturers have been trying to determine how to integrate onsite renewable energy and real-time electricity pricing into manufacturing schedules without compromising profitability. To address this challenge, we propose a bi-level model predictive control framework that jointly optimizes product prices and production scheduling with explicit consideration of renewable energy integration. The higher level determines the product price to maximize revenue and renewable energy usage. The lower level controls production scheduling in runtime to minimize operational costs and respond to the product demand. Market response is incorporated through price elasticity, enabling strategic pricing to align the product demand with the availability of renewable energy. Results from a lithium-ion battery pack manufacturing system case study demonstrate that our approach enables manufacturers to reduce grid energy costs while increasing profit.

Keywords: Robotics, Adaptive and Learning Systems, Manufacturing Systems
Abstract: Interoperability and data integration are significant challenges in production planning and control (PPC) due to heterogeneous data formats and fragmented software systems. This paper introduces aas-middleware, an open-source software system leveraging Asset Administration Shells (AAS) to enable interoperable, real-time integration for smart manufacturing. By realizing service-orientation, aas-middleware addresses data heterogeneity and system integration challenges, facilitating automation and information orchestration in PPC. This research details the middleware's architecture and showcases its application and evaluation in a modular assembly station, demonstrating its ability to bridge IT and OT systems effectively in real-time and to orchestrate complex planning tasks automatically.

Keywords: Sensors and Actuators, Electromechanical systems, Manufacturing Systems
Abstract: The development of hybrid-process multi-material additive manufacturing with conductive nanoparticle inks and electrically-insulating composites enables the design of novel electromechanical actuators. In this work, we present the design, fabrication, and characterization of a novel fully-additively manufactured cylindrical three-phase radial-flux stator for an electric motor. This stator is manufactured using a unique four-axis (three linear, one rotary) 3D printer. While prior work has focused on additively manufacturing planar axial motors, the stator in this work is printed on a rotating axis. This enables the creation of a radial stator with nine slots, each consisting of five layers of direct ink write silver nanoparticle ink coils encased in fused filament fabricated nylon-fiberglass insulation. After this rotary electric motor is assembled, the performance and dynamics of the motor are characterized, demonstrating 0.51 N·mm/A peak torque, 1.02% peak efficiency, 0.67 Hz switching bandwidth, and continuous operation at up to 150°C for a 20.7 W input. The motor is used as a drill in order to demonstrate a realistic application. Printing the motor with the rotary axis not only enables more complex designs, but also enables manufacturing of previously unprintable 3D geometries for embedded high-power electric machines.

Keywords: Nonlinear Control Systems, Robotics
Abstract: Mobile sampling platforms are increasingly being used in marine studies. Accurate control of such platforms is often essential to study design, but is also difficult as the dynamics of mobile platforms are often nonlinear, nonholonomic, and underactuated. Event-triggered control where the applied control is updated, often aperiodically, at discrete times, addresses practical limitations on computational power and actuator response; however, most of the existing results ignore nonholonomic constraints and assume input-to-state stability of the closed-loop system. This work proposes a novel event-trigger design that relaxes these limitations. The efficacy of the proposed event-triggering mechanism is further shown in simulation.

Keywords: Path Planning and Motion Control, Transportation Systems, Automotive Systems
Abstract: Accurate detection of road edges is a key step in generating high-definition (HD) maps for autonomous driving. This work presents a road edge detection strategy that finds pavement drop-offs, curbs, or similar vertical deviations from the road surface by processing LiDAR data within individual scan lines, approximating each scan line as a road cross section in transverse-height coordinates. The key insight of this work is to exploit optimal extrema filtering to find maximum and minimum values in the first and second derivatives of the road profile, along with finding the corresponding strength of each extrema. The road boundary points, e.g., locations of sudden deviation from the road surface, are shown to correspond to extrema with the highest correlation strengths. The method was evaluated using real-world data collected from the Penn State mapping van operating at Penn State’s Larson Transportation Institute test track. Results demonstrate that the proposed approach yields consistent boundary detection, even in challenging overlapping scenarios. As well, special road mapping cases - such as when there are vertical overlapping features over the road's edge that block driving access (guardrails, vegetation, signage, etc.) - can be handled in the processing steps via simple modifications. A key advantage of the result is that it can process LiDAR data into road-surface estimates at the 2D scan-line level, allowing very rapid data processing and avoiding 3D point cloud processing.

Keywords: Sensors and Actuators, Intelligent Autonomous Vehicles, Unmanned Ground and Aerial Vehicles
Abstract: High-precision measurement of roadway spatial data is essential for applications such as high-definition mapping, advanced driver assistance systems (ADAS), and autonomous driving. These applications require accurate alignment between sensors and the vehicle coordinate frame to ensure consistent localization and perception. Among various sensing modalities, 3D Light Detection and Ranging (LiDAR) has emerged as a key technology due to its ability to capture dense and accurate point clouds in complex environments. However, the quality of LiDAR-based measurements depends critically on precise extrinsic calibration, aligning the LiDAR frame with the vehicle and other onboard sensors such as the Global Positioning System (GPS), Inertial Measurement Unit (IMU), and cameras.

A wide range of extrinsic calibration techniques has been developed for LiDAR sensors, with target-based methods remaining widely adopted due to their geometric stability and reproducibility. The proposed method focuses specifically on spherical targets, leveraging their geometric stability and view invariance to enable accurate calibration under diverse environmental and sensor configurations.

The calibration pipeline comprises three main steps: (1) static deployment of spherical targets with known geometry and GPS-surveyed centers; (2) robust sphere detection and center estimation from LiDAR scans using a constrained RANSAC followed by least-squares refinement; and (3) point-to-point registration using singular value decomposition (SVD) to compute the rigid-body transformation between the LiDAR frame and the vehicle’s GPS reference frame. Experimental validation using road lane markers demonstrates a mean positional accuracy of 3.6 cm, with a standard deviation of 1.3 cm and a maximum error of 7.1 cm. The methodology can achieve calibration either with single-scan data or by aggregating data across scans. These results confirm the method's effectiveness in improving spatial data accuracy for autonomous navigation, HD map generation, and multi-sensor fusion applications.

Keywords: Path Planning and Motion Control, Unmanned Ground and Aerial Vehicles, Uncertain Systems and Robust Control
Abstract: Autonomous drones have gained considerable attention for applications in real- world scenarios, such as search and rescue, inspection, and delivery. As their use becomes ever more pervasive in civilian applications, failure to ensure safe operation can lead to physical damage to the system, environmental pollution, and even loss of human life. Recent work has demonstrated that motion planning techniques effectively generate a collision-free trajectory during navigation. However, these methods, while creating the motion plans, do not inherently consider the safe operational region of the system, leading to potential safety constraints violation during deployment. In this paper, we propose a method that leverages run time safety assurance in a kinodynamic motion planning scheme to satisfy the system’s operational constraints. First, we use a sampling-based geometric planner to determine a high-level collision- free path within a user-defined space. Second, we design a low-level safety assurance filter to provide safety guarantees to the control input of a Linear Quadratic Regulator (LQR) designed with the purpose of trajectory tracking. We demonstrate our proposed approach in a restricted 3D simulation environment using a model of the Crazyflie 2.0 drone.

Keywords: Machine Learning in modeling, estimation, and control, Intelligent Autonomous Vehicles, Transportation Systems
Abstract: Perception is a critical component of autonomous vehicles (AVs), yet onboard sensors often suffer from occlusions and blind spots caused by other vehicles, infrastructure, or environmental elements. This paper presents an enhanced perception framework for autonomous vehicles that addresses occlusion challenges by integrating Vehicle-to-Infrastructure (V2I) data through a two-step Kalman Filter fusion strategy. Unlike traditional onboard sensor-only approaches, the proposed method combines LiDAR data from the ego vehicle with 3D object detections from roadside unit (RSU) camera, leveraging late fusion to fuse multisource inputs. The framework employs a Kalman Filter with dual update steps to merge the input data, optimizing object detection and tracking accuracy in occlusion-prone scenarios. In CARLA, a simulation scenario involving occlusion at an intersection is reconstructed to evaluate the proposed approach in terms of perception accuracy. Experimental results show a 21.5% improvement in Mean IoU and a reduction in RMSE to 2.02m compared to the baseline.

Keywords: Unmanned Ground and Aerial Vehicles, Control Applications, Multi-agent and Networked Systems
Abstract: This paper presents a hierarchical control strategy for a team of quadrotor unmanned aerial vehicles (UAVs) engaged in cooperative region-following tasks, as encountered in inspection missions. In the proposed approach, an outer-loop guidance law uses potential functions to drive each UAV toward a moving target region while avoiding inter-UAV collisions. The uncertainty and disturbance estimator (UDE) reinforces the guidance law with a disturbance compensation term to counteract modeling uncertainties and external disturbances, such as unmodeled aerodynamic effects or wind disturbance. The inner-loop control is a geometric nonlinear controller that tracks the outer-loop's commanded force vector by generating appropriate thrust and orientation commands in a globally stable manner. This two-tier design is fully decentralized, requiring only local sensing of each UAV's state and relative distances to neighbors. We validate the approach through simulation studies, which demonstrate robust region tracking and effective collision avoidance under various test scenarios.

Keywords: Modelling and Control of Environmental Systems, Power and Energy Systems, Control of Smart Buildings and Microgrids
Abstract: Borehole thermal energy storage (BTES) can reduce the operation of fossil fuel-based heating, ventilation, and air conditioning systems for buildings. With BTES, thermal energy is stored via a borehole heat exchanger in the ground. Model predictive control (MPC) may maximize the use of BTES by achieving a dynamic interaction between the building and BTES. However, modeling BTES for MPC is challenging, and a trade-off between model accuracy and an easy-to-solve optimal control problem (OCP) must be found. This manuscript presents an accurate numerical model yielding an easy-to-solve linear-quadratic OCP.

Keywords: Modelling, Identification and Signal Processing, Distributed Parameter Systems, Power and Energy Systems
Abstract: The Krylov subspace method is an efficient physics-based approach for reduced-order modeling (ROM) of large-scale linear systems common in computational fluid dynamics (CFD). It works by matching moments between a full-order model (FOM) and its ROM counterpart at specific interpolation points or shifts. However, its adoption is hindered by the need for high-dimensional operators from the CFD model. Recently, several methods have emerged to bypass this requirement for diffusive and convective transport problems. This paper interprets such attempts as a new branch of ROM techniques based on alter-and-excite (A&E) for improved identification, in contrast with traditional system-identification techniques relying solely on excitation for extraction of system information. In an A&E approach, the CFD model is iteratively both altered and excited to identify Krylov subspace basis vectors directly from system responses. The A&E approach, thus, allows building a physics-based ROM without requiring the high-dimensional system operator of a CFD model. There exists, therefore, a great potential for industrial adoptions of ROMs via employment of the A&E approach. However, previous A&E methods are confined to interpolation around real shifts. A novel A&E method for generating the Krylov subspace with complex conjugate shifts in presented in this paper. We prove, by adopting an induction process, that reformulating the CFD model as a multiphysics problem provides the necessary information for constructing a ROM based on real matrices when complex shifts are adopted. The study details the implementation and evaluation of the A&E method with complex conjugate shifts for a mass-transport problem in turbulent flows.

Keywords: Modelling, Identification and Signal Processing, Estimation, Sensors and Actuators
Abstract: This paper examines the problem of modeling the dynamics of oxygen transport through the endotracheal tube of a mechanical ventilator. Such modeling can be potentially valuable for ventilator control, patient monitoring, and estimating airway oxygen transport during experimental studies on hypoxia. The paper presents an airway sensing package that provides collocated measurements of oxygen concentration and total gas flow rate. In theory, multiplying these two measurements can provide an estimate of tracheal oxygen flow rate. Unfortunately, substantial differences in dynamics between the above two sensors necessitate a more sophisticated modeling and estimation approach. Towards this goal, the paper presents a dynamic model of advective tracheal oxygen transport that accounts for measurement dynamics and delays. Parameters of this model are estimated from a hypoxia experiment on a Yorkshire swine, leading to good fitting accuracy for oxygen concentration measurements

Keywords: Modelling, Identification and Signal Processing, Machine Learning in modeling, estimation, and control, Estimation
Abstract: Electromyography (EMG) signals are essential for neuromuscular diagnosis and assistive device control but suffer from noise contamination that obscures clinically important transient features. We present an interpretable, automated framework that enhances EMG quality by integrating low rank Hankel matrix reconstruction with Bayesian regularized Koopman regression and optimized smoothing, all tuned through multi objective Bayesian optimization (MOBO). Raw EMG traces are embedded in a time delay Hankel matrix and decomposed via truncated singular value decomposition (SVD) to isolate underlying dynamics. A ridge regularized Bayesian regression learns a predictive transition operator, whose forecasts are optimally combined with the reconstructed signal. A windowed Makima median smoother provides final refinement. Key hyperparameters including embedding dimension, SVD rank, regularization variances, and fusion weight are selected to simultaneously maximize signal to noise ratio (SNR) and structural fidelity using Gaussian process surrogates and an expected hypervolume improvement (EHVI) criterion. When evaluated on walking and squatting EMG data from ten healthy adults across four muscle groups, our method achieves a 108.7% increase in a squared geometric mean (SGM) metric (p = 0.0282) and an 11.4% gain in normalized joint fidelity (p = 0.0011), while decoupling noise reduction from signal preservation (Pearson r ≈ 0). These results establish a powerful, generalizable approach for real time clinical and rehabilitation applications.

Keywords: Modelling, Identification and Signal Processing, Manufacturing Systems, Machine Learning in modeling, estimation, and control
Abstract: Extracting pulsive temporal patterns from a small, possibly one-shot dataset shows significant importance in manufacturing applications but does not sufficiently attract scientific attention. We propose to extract such temporal patterns based on their length, namely, quantifying how long patterns appear in a small dataset without relying on their repetition or singularity. Inspired by the celebrated time delay embedding and data-driven Hankel matrix analysis, we introduce a linear dynamical system model on the time-delay coordinates behind the data to derive the discrete-time bases each of which has a distinct exponential decay constant. The derived bases are fitted onto subsequences that are extracted with a sliding window in order to quantify how long patterns are dominant in the set of subsequences. We call the quantification method Data-driven Exponential Framing (DEF). A toy model-based experiment shows that DEF can identify multiple patterns with distinct lengths. DEF is also applied to electric current measurement on a punching machine, showing its possibility to extract multiple patterns from real-world oscillatory data.

Keywords: Optimal Control, Modeling and Validation, Manufacturing Systems
Abstract: This paper introduces a fuzzy logic-inventory model using blockchain technology to assist manufacturers in managing uncertainty in supply chains and enhancing profitability. Based on the previous work, we utilize three categories of fuzzy numbers—triangular, trapezoidal, and heptagonal —to model vagueness in such critical parameters as demand, price sensitivity, and blockchain deployment expenses. The model considers two categories of retailers: those embracing blockchain (who pay transparency costs) and those that do not. Our numerical and analytical results indicate that the incorporation of blockchain always enhances profit in all fuzzy conditions, with the heptagonal model being particularly sensitive to uncertainty in price and levels of information-sharing. Heptagonal works best under turbulent markets, trapezoidal provides a compromise under moderate risk, and triangular is best suited for stable environments. The research also finds important thresholds (e.g., α > 5%) affecting optimal choices. In general, this study fills the existing gap between abstract blockchain models and real-world uncertainty by providing a flexible and responsive framework for supply chain decision-making.

Keywords: Control Design
Abstract: This presentation introduces a direct approach for constructing a feedforward Rectified Linear Unit (ReLU) neural network that exactly represents the explicit solution of a class of multi-parametric Quadratic Programs (mp-QPs). The weights and biases of the neural network are determined analytically for a given mp-QP, removing the need for training. While the number of layers of the neural networks equals the number of inequality constraints in the mp-QP and the weight matrices are sparse, the total number of neurons grows exponentially. The proposed approach enables the optimal feedback control policy associated with linear Model Predictive Control (MPC) to be represented as a feedforward neural network. Several numerical examples show that the proposed neural network representation can be significantly faster to generate and evaluate than region-based explicit solutions to mp-QPs for problems with a small number of inequality constraints.

Keywords: Manufacturing Systems
Abstract: Recent advances in manufacturing system technology and increasing customer expectations have led to a growing interest in mass individualization - the ability of manufacturing systems to create customized products on a large scale. For consumers, manufacturing systems with mass individualization capabilities will create enhanced customer experience and enable the development of tailored solutions based on their unique requirements. For manufacturers, developing mass individualization capabilities could offer significant competitive advantages over manufacturers offering a single product or product family. However, since control strategies for manufacturing systems have been primarily developed for mass production, there is a lack of fundamental knowledge about the impact of individualized control strategies and a lack of performance guarantees for these control strategies. This Faculty Early Career Development (CAREER) research project focuses on the integration of "intelligent products" in the production process to enable more customized and individualized production. An intelligent product consists of a physical part and a software component that contains the customer specifications, manufacturing capabilities, and production plans for the individual part. The developed intelligent products will adapt to various specifications, while ensuring that the manufacturing environment remains reliable and safe. In addition, through hands-on workshops and activities, the project will also focus on establishing a manufacturing technology education program to provide opportunities for the development of an informed manufacturing workforce.

Keywords: Optimal Control
Abstract: In this presentation, we introduce a generalizable framework for the automated tuning of parameter-dependent predictive control laws, using Bayesian Optimization (BO) and Gaussian Processes (GP). The methods we propose require minimal measurements to be collected from the system and iteratively learn the parameter values, maximizing a user-defined performance index. Three novel acquisition functions – one inspired by particle filters and two by gradient-based methods – are presented and benchmarked against state-of-the-art, showing favorable performances under small-data regimes, while maintaining low computational overhead. The framework is applied to the autotuning of a Data-Enabled Multi-Step Predictive Controller (DeMuSPc) designed for the personalized automated insulin delivery in people with Type 1 Diabetes (T1D), and validated in-silico exploiting a high-fidelity metabolic simulator of T1D dynamics.

Keywords: Robotics
Abstract: We will present a learning-based framework that enables robots to "learn from the experience" of other robots, even when their dynamics differ. This approach to robot planning and control significantly reduces the effort required for design, testing, evaluation, and certification, while allowing robots to be uniquely and efficiently customized for their operating environments. Central to this framework is an architecture that automatically adjusts the outputs of baseline planners and controllers by incorporating feedforward learning signals to enhance performance. Rather than replacing existing planning and control methods or competing for the most highly optimized performance, this framework provides an elegant learning mechanism that is highly adaptable, reasonably efficient, and requires minimal hardware modification and software reconfiguration.

Keywords: Robotics
Abstract: As generative AI and robotics increasingly integrate into daily life, ensuring their safe interaction with humans remains a critical challenge. Safety concerns extend beyond physical interactions—such as preventing collisions—to conversational safety, where AI must avoid exchanging harmful or dangerous information. In this talk, I will discuss how we may frame these challenges as constraint satisfaction problems and address them using forward invariance principles from control theory. The first line of work I will discuss focuses on Physical Safety. I will introduce SPARK, a comprehensive toolbox and benchmark designed to ensure safety in humanoid autonomy and teleoperation. The second line of work addresses Conversational Safety. Large language models (LLMs) are highly vulnerable to multi-turn jailbreaking attacks, where contextual drift gradually leads them away from safe behavior. To mitigate this, we propose a safety steering framework grounded in control theory to maintain invariant safety in multi-turn dialogues. Lastly, as many safety certificates are learned via neural networks, a critical question arises: how can we certify Neural Safety Certificates? I will discuss formal verification methods designed to provide guarantees on their reliability. By applying control-theoretic safety principles across diverse domains—from physical robot safety to conversational AI—we aim to build AI systems that interact with humans in both safe and trustworthy ways.

Keywords: Automotive Systems, Estimation, Intelligent Autonomous Vehicles
Abstract: This paper presents a model-free approach to compensate for time-varying output delay in networked control systems. The proposed architecture combines a model-free observer and the Smith predictor. The model-free observer estimates the current state while handling modeling errors and uncertainties of the system. The Smith predictor moves the effect of time delay outside the control closed-loop using the estimated delayed output and the actual output of the plant. The proposed method is applied to a cooperative on-ramp merging problem. First, an ultra-local model predictive control is implemented to provide a computationally efficient online speed planner agnostic to the vehicle dynamics. After that, a model-free observer is designed to estimate the current state. Finally, the proposed architecture is tested against a time-varying output delay with an upper bound of 200 milliseconds. The results demonstrate the effectiveness of the proposed method with improved tracking of intervehicle distance.

Keywords: Human-Machine and Human-Robot Systems, Path Planning and Motion Control, Robotics
Abstract: We propose a framework that enables autonomous vehicles (AVs) to proactively shape the intentions and behaviors of interacting human drivers. The framework employs a leader-follower game model with an adaptive role mechanism to predict human interaction intentions and behaviors. It then utilizes a branch model predictive control (MPC) algorithm to plan the AV trajectory, persuading the human to adopt the desired intention. The proposed framework is demonstrated in an intersection scenario. Simulation results illustrate the effectiveness of the framework for generating persuasive AV trajectories despite uncertainties.

Keywords: Control Design, Optimal Control, Machine Learning in modeling, estimation, and control
Abstract: Connected and Automated Vehicles (CAVs) offer significant potential for improving energy efficiency and lowering vehicle emissions through eco-driving technologies. Control algorithms in CAVs leverage look-ahead route information and Vehicle-to-Everything (V2X) communication to optimize vehicle performance. However, existing eco-driving strategies often neglect macroscopic traffic effects, such as upstream traffic jams, that occur outside the optimization horizon but significantly impact vehicle energy efficiency. This work presents a novel Neural Network (NN)-based methodology to approximate the terminal cost within a model predictive control (MPC) problem framework, explicitly incorporating upstream traffic dynamics. By incorporating traffic jams into the optimization process, the proposed traffic-aware approach yields more energy-efficient speed trajectories compared to traffic-agnostic methods, with minimal impact on travel time. The framework is scalable for real-time implementation while effectively addressing uncertainties from dynamic traffic conditions and macroscopic traffic events.

Keywords: Intelligent Autonomous Vehicles, Control Design, Path Planning and Motion Control
Abstract: In this paper, we design a safe and efficient cruise control for the connected automated vehicle with access to motion information from multiple vehicles ahead via vehicle-to-vehicle (V2V) communication. Position and velocity data collected from a chain of human-driven vehicles are systematically leveraged to design a connected cruise controller that smoothly responds to traffic perturbations while maximizing energy efficiency. A safety filter derived from a control barrier function provides the safety guarantee. We investigate the proposed control design's energy performance against real traffic datasets and quantify the safety filter's energy impact. It is shown that optimally utilizing V2V connectivity reduces energy consumption by more than 10% compared to standard non-connected adaptive cruise control. Meanwhile, interesting interplays between safety filter and energy efficiency design are highlighted, revealing future research directions.

Keywords: Automotive Systems, Estimation, Mechatronic Systems
Abstract: Snowplows play a vital role in maintaining road safety during winter, yet they frequently suffer collisions with other vehicles due to reduced visibility, drivers’ inattention, and hazardous conditions during snowplowing. Existing collision avoidance systems primarily focus on forward object detection, whereas research on rear-approaching vehicles is limited. This paper presents a snowplow protection system that continuously tracks the trajectories of vehicles approaching from the rear using a low-cost radar and a nonlinear observer. Once a potential collision is detected, an audiovisual warning system is activated from the snowplow to alert the approaching driver. Unlike conventional vehicle tracking models that rely on global coordinate systems, this study introduces a relative motion model that accounts for the ego vehicle’s rotational motion effects, improving tracking performance during turning maneuvers. Additionally, a nonlinear observer is implemented to enhance estimation accuracy across varying vehicle speeds using a single observer gain. The proposed system offers a cost-effective and practical solution for snowplow safety in harsh weather conditions, with its performance validated through both simulations of various scenarios and real-world vehicle experiments.

Keywords: Estimation, Modelling, Identification and Signal Processing, Robotics
Abstract: The skidding and slipping motion of skid-steered wheel mobile robots (SSWMRs) is highly influenced by the complex nature of tire-terrain interactions. Lack of reliable terrain friction models cascade into unreliable motion models, especially the reduced-ordered variants used for state estimation and robot control. Ensemble modeling is an emerging research direction where the overall motion model is broken down into a family of local models to distribute the performance and resource requirement and provide a fast real-time prediction. To this end, a Gaussian Mixture Model (GMM) based modeling approach for identification of model clusters is adopted and implemented within an Interactive Multiple Model (IMM) based state estimation framework. The methodology is applied and investigated for estimating angular velocity for a mid-scale skid-steered wheel mobile robot platform.

Keywords: Control Applications, Nonlinear Control Systems, Power and Energy Systems
Abstract: Safety concerns associated with lithium-ion batteries (LiBs) have become increasingly prominent due to their widespread use in electronic devices. Physical or electrical damage can lead to short circuits (SC), potentially triggering thermal runaway (TR) if not properly managed. TR is particularly hazardous at high states of charge (SoC) due to the presence of residual energy. This research aims to develop a method for estimating unmeasurable states and safely discharging LiBs based on these estimates. A Lyapunov-based observer is proposed to estimate the SoC, SC current, and core temperature. To ensure safe discharge, an input-to-state exponentially stabilizing control Lyapunov function and a perturbation control barrier function are formulated and implemented via quadratic programming. These tools are used to design a controller that accounts for estimation errors while utilizing the estimated SoC and core temperature. The proposed approach is validated using a simulation model of a 111Ah cell. The observer yielded accurate state estimates, enabling a safe discharge from 100% to 50% SoC within 14 minutes while maintaining all safety constraints.

Keywords: Machine Learning in modeling, estimation, and control, Estimation, Power and Energy Systems
Abstract: High-energy-density batteries are gaining popularity with the rise of electric vehicles (EVs). State-of-charge (SOC) estimation is crucial for informed decision-making, operational safety, and the longevity of these batteries; however, model-based SOC estimation often struggles due to the nonlinear dynamics of batteries, unpredictable measurement noise, and dynamic loading conditions. Model-free data-driven methods provide an alternative solution. However, they often struggle to process a long sequence of temporal dependencies in the input data. Furthermore, uncertainty awareness associated with the estimated SOC is often unavailable.

This work presents an uncertainty-aware SOC estimation framework integrating a Gated Recurrent Unit (GRU) network with deep evidential regression. Our framework processes sequential time-series data to estimate SOC and its associated uncertainty in a single forward pass. Validation results on real-world battery cycling datasets show that, for in-distribution data, our method achieves predictive performance comparable to computationally intensive ensemble methods.

Keywords: Nonlinear Control Systems, Optimal Control, Control Applications
Abstract: This paper presents a novel application of control barrier functions (CBFs) for calculating an estimate of state of power (SOP) during charging and discharging cycles of lithium-ion batteries. We define SOP as the maximum amount of power that can be maintained over a specified time period. The proposed algorithm predicts the maximum achievable power level within a constraint set defining the operational boundaries of the cell, namely, state of charge (SOC), voltage, and core temperature.

The estimation algorithm uses two‐stages, whereby an outer bisection loop rapidly converges to the highest feasible power level within a user‐specified tolerance and an inner loop solves a quadratic program to adjust computed current levels so that the battery tracks constant‑power contours while satisfying CBF constraints. The algorithm effectively solves a minimax problem with tunable coefficients which allow users to balance assertiveness and conservatism in the constraint satisfaction.

The novelty of this algorithm lies in providing a computationally manageable approach to compute state of power, as opposed to previously proposed algorithms which focus on the total energy delivery over a specified time horizon.

To demonstrate the efficacy of this approach, we simulate battery performance under the Urban Dynamometer Driving Schedule (UDDS), a representative profile of city driving conditions. Comparisons with i) a model predictive control (MPC) and ii) a bisection-based approach illustrate the merits of the CBF-based method in terms of practicality for real-world automotive applications. The aim of this work is to help advance efforts to create safer and more effective battery management solutions for electric vehicles and energy storage technologies.

Keywords: Power and Energy Systems, Control Applications, Control Design
Abstract: An increasing number of smart devices controlling loads opens a potential pathway for false data attacks which could alter the loads. The presence of energy storage with its ability to quickly respond to discrepancies in loads offers a promising solution for security by preventing further instabilities and potential blackouts. This paper proposes a control methodology for secure predictive energy management that uses batteries to mitigate the impact of load-altering attacks. To that extent, we develop a microgrid model along with the primary control for microgrid. The developed models and the optimization algorithm are validated through a real-time numerical simulation of a modified IEEE 9 bus system involving a battery as one of the energizing sources. The results show the effectiveness of the battery in countering the load alterations.

Keywords: Power and Energy Systems, Estimation, Sensors and Actuators
Abstract: Abstract Accurate estimates of cell state of power (SOP) are critical to maximize battery-pack performance and safety. Since SOP is not directly measurable, algorithms having varying complexity are implemented to compute SOP estimates. Input to these algorithms are the cell's measurable quantities, acquired with sensors whose characteristics are defined by precision, accuracy, and synchronicity. This paper provides an evaluation of the performance of SOP estimation algorithms versus the integrity of the measurements provided by the cell voltage, current, and temperature sensors. Overviews of state-of-charge and cell-resistance estimation, required by SOP, are also shown. We employ model-based simulation to compare the ideal case having zero sensor measurement error against real-life sensor performances which exhibit measurement offset, noise and nonsynchronicity. We consider typical usage scenarios in electric-vehicle and ESS applications, cell chemistry, estimation method, and sensor performance.

Keywords: Power and Energy Systems, Transportation Systems, Control Applications
Abstract: An electrified roadway system that transfers power from the roadway infrastructure to electric vehicles (EVs) while the vehicle is in motion can downsize the required energy-storage capacity of EV battery packs, reduce costs, and promote adoption. The battery packs in EVs driving on electrified roads are charged via wireless inductive or capacitive transmitters embedded in the road, coupled with matched receivers in the vehicle. They are discharged by vehicle loads such as propelling demand, and randomly receive charge through regenerative braking. The control of the charging process must consider vehicle dynamics, available power, safety limits, and battery status. This work evaluates the effectiveness of different power-management strategies for in-road wireless charging by examining their effect on the state of the battery pack. The investigation is carried out using a simulation framework based on a battery equivalent-circuit model and simulates EV operation with different power-control configurations.

Keywords: Intelligent Autonomous Vehicles, Machine Learning in modeling, estimation, and control, Modeling and Validation
Abstract: We present a modular, language-guided scenario generation and evaluation pipeline for autonomous vehicle simulation using CARLA and Autoware. The pipeline forms a closed-loop system that translates natural language input into OpenSCENARIO scripts, executes them via full-stack autonomy, and uses performance feedback to inform subsequent generations. Our framework generates diverse and progressively adaptive scenarios at scale by leveraging prompt engineering and chain-of-thought reasoning. Through iterative tests of adaptive cruise control (ACC) and cut-in scenarios, we demonstrate that the pipeline adjusts scenario difficulty according to performance metrics and supports robust ego agent evaluation. This work establishes a foundation for intelligent, scalable, and fully automated AV scenario testing workflows.

Keywords: Intelligent Autonomous Vehicles, Machine Learning in modeling, estimation, and control, Unmanned Ground and Aerial Vehicles
Abstract: Performance and reliability analyses of autonomous vehicles (AVs) can benefit from tools that "amplify" small datasets to synthesize larger volumes of plausible samples of the AV's behavior. We consider a specific instance of this data synthesis problem that addresses minimizing the AV's exposure to adverse environmental conditions during travel to a fixed goal location. The environment is characterized by a threat field, which is a strictly positive scalar field with higher intensities corresponding to hazardous and unfavorable conditions for the AV. We address the problem of synthesizing datasets of minimum exposure paths that resemble a training dataset of such paths. The main contribution of this paper is an inverse reinforcement learning (IRL) model to solve this problem. We consider time-invariant (static) as well as time-varying (dynamic) threat fields. We find that the proposed IRL model provides excellent performance in synthesizing paths from initial conditions not seen in the training dataset, when the threat field is the same as that used for training. Furthermore, we evaluate model performance on unseen threat fields and find low error in that case as well. Finally, we demonstrate the model's ability to synthesize distinct datasets when trained on different datasets with distinct characteristics.

Keywords: Intelligent Autonomous Vehicles, Modeling and Validation, Robotics
Abstract: The engineering community currently encounters significant challenges in the systematic development and validation of autonomy algorithms for off-road ground vehicles. These challenges are posed by unusually high test parameters and algorithmic variants. In order to address these pain points, this work presents an optimized digital engineering framework that tightly couples digital twin simulations with model-based systems engineering (MBSE) and model-based design (MBD) workflows. The efficacy of the proposed framework is demonstrated through an end-to-end case study of an autonomous light tactical vehicle (LTV) performing visual servoing to drive along a dirt road and reacting to any obstacles or environmental changes. The presented methodology allows for traceable requirements engineering, efficient variant management, granular parameter sweep setup, systematic test-case definition, and automated execution of the simulations. The candidate off-road autonomy algorithm is evaluated for satisfying requirements against a battery of 128 test cases, which is procedurally generated based on the test parameters (times of the day and weather conditions) and algorithmic variants (perception, planning, and control sub-systems). Finally, the test results and key performance indicators are logged, and the test report is generated automatically. This then allows for manual as well as automated data analysis with traceability and tractability across the digital thread.

Keywords: Intelligent Autonomous Vehicles, Nonlinear Control Systems, Robotics
Abstract: Ground vehicles maneuvering in limited and congested areas, such as intersections, can become challenging when a hazard condition is triggered. Facing an imminent collision, manned or unmanned vehicles have reduced time and space to react. A decentralized response complicates the situation, allowing for coupled unsafe behaviors and mutual interference, increasing the chances of accidents. For these cases, a centralized optimal approach has the potential to achieve the best possible outcome, reducing the likelihood of a collision. This work proposes a centralized generation of optimal trajectories for all vehicles involved in a potential future hazard, maximizing obstacle and collision avoidance, which will be tracked individually by predictive control.

Keywords: Intelligent Autonomous Vehicles, Path Planning and Motion Control, Robotics
Abstract: We formulate a vector cost alternative to the scalarization method for weighting and combining multi-objective costs. The algorithm produces solutions to bimatrix games that are simultaneously pure, unique Nash equilibria and Pareto optimal with guarantees for avoiding worst case outcomes. We achieve this by enforcing exact potential game constraints to guide cost adjustments towards equilibrium, while minimizing the deviation from the original cost structure. The magnitude of this adjustment serves as a metric for differentiating between Pareto optimal solutions. We implement this approach in a racing competition between agents with heterogeneous cost structures, resulting in fewer collision incidents with a minimal decrease in performance. Code is available at https://github.com/toazbenj/race_simulation.

Keywords: Intelligent Autonomous Vehicles, Uncertain Systems and Robust Control, Robotics
Abstract: This paper proposes a control-oriented optimization platform for autonomous mobile robots (AMRs), focusing on extending battery life while ensuring task completion. The requirement of fast AMR task planning while maintaining minimum battery state of charge, thus maximizing the battery life, renders a bilinear optimization problem. McCormick envelop technique is proposed to linearize the bilinear term. A novel planning algorithm with relaxed constraints is also developed to handle parameter uncertainties robustly with high efficiency ensured. Simulation results are provided to demonstrate the utility of the proposed methods in reducing battery degradation while satisfying task completion requirements.

Keywords: Control Applications, Control Design, Unmanned Ground and Aerial Vehicles
Abstract: The paper proposes two Active Disturbance Rejection Control (ADRC) schemes for quadrotors’ angular rate under wind gusts. Firstly, it introduces a Linear Active Disturbance Rejection Control (LADRC) and analyzes the effects of tuning parameters on the control performances. Secondly, it presents a Generalized Active Disturbance Rejection Control (GADRC) designed with an actual quadrotor’s model identified from experimental data using the Optimized Predictor-Based Subspace Identification (PBSIDopt) method and discusses the design variables. The designed controllers are implemented for attitude control in the inner loop of a cascade architecture which incorporates a proportional controller in the outer loop. The study compares in simulation the performance of the designed control schemes with the classical P/PID controller architecture by subjecting them to a generated artificial wind disturbance that simulates actual wind conditions.

Keywords: Control Applications, Estimation, Automotive Systems
Abstract: This paper addresses a new observer-based robust fault-tolerant control strategy (OFTC) for steering guidance (SG) and lane keeping assistance system (LKAS) of highly automated vehicle equipped with Steer-by-Wire (SbW) system considering multiple actuator faults. The main contribution of this work is to propose a novel co-design of a robust adaptive simultaneous estimation of both system state and multiple actuator faults associated with an adaptive control law for stability purposes to ensure both lane keeping and steering guidance performance even in faulty situations. A Linear Parameter Varying (LPV) observer architecture is proposed to estimate both the vehicle state and unknown actuator faults considering real-time unmeasurable variations in longitudinal and lateral velocities, represented within a polytope with finite vertices. Subsequently, a robust and adaptive state feedback active fault-tolerant controller is proposed using the Takagi-Sugeno (T-S) approach. An optimization problem is formulated in terms of linear matrix inequalities (LMI) to guarantee system stability and the asymptotic convergence of state and fault estimation errors. Lyapunov stability arguments are used to allow more relaxation and additional robustness against immeasurable nonlinearities. Hardware validation is carried out using the SHERPA dynamic car simulator in real driving situations, demonstrating the performance and the effectiveness of the proposed OFTC scheme.

Keywords: Control Applications, Multi-agent and Networked Systems, Transportation Systems
Abstract: Highway merges present difficulties for human drivers and automated vehicles due to incomplete situational awareness and a need for a structured (precedence, order) environment, respectively. In this paper, an unstructured merge algorithm is presented for Connected and Automated Vehicles (CAVs). There is neither precedence nor established passing order through the merge point. The algorithm relies on Control Barrier Functions for safety (collision avoidance) and for coordination that arises from exponential instability of gridlock-causing equilibria in the inter-agent space. A Monte Carlo simulation comparison to a first-in-first-out approach shows improvement in traffic flow and a significant energy efficiency benefit.

Keywords: Control Design, Nonlinear Control Systems, Unmanned Ground and Aerial Vehicles
Abstract: A tethered multirotor autogyro can function as an unmanned aerial vehicle for energy-efficient and prolonged deployment, as it uses the available wind energy to sustain flight. This article presents an adaptive altitude control strategy for such a device. At a constant wind speed, the equilibrium altitude can be approximated by a quadratic function of the pitch angle. The proposed adaptive control estimates the coefficients of this quadratic function. The estimates are used for altitude control and to attain the maximum altitude (and minimum horizontal drift) for a given wind speed. A feedback controller based on regenerative differential rotor braking is used as the actuation to modulate the autogyro’s pitch angle. Implementation of the controller using a control-oriented, higher-order dynamic model demonstrates the controller’s capability to regulate the altitude and maintain stable flights under varying wind speeds. Based on the system’s maximum altitude tracking performance, the adaptive control is adjusted to improve performance under substantial changes in wind speeds.

Keywords: Transportation Systems, Intelligent Autonomous Vehicles, Automotive Systems
Abstract: Due to their cost efficiency, safety improvements, and repeatability, traffic simulators are an integral and widely used tool to study the performance of a traffic network with connected and/or autonomous vehicles (CAVs). Such simulations depend significantly on the choice of simulation parameters, and in most cases, it is up to the user to identify the appropriate setting for even a basic usage question: what duration should the simulation be run? It is generally understood that a simulation's duration should be sufficient to ensure that the simulations are fully initialized wherein the user-chosen initial conditions of the simulation do not affect assessments of steady behavior. This duration is generally chosen based on the user's practical experience. This paper presents a method for determining a simulation's initialization interval by examining the settling time for traffic simulations, e.g. the time the simulation requires to reach steady-state conditions. In this work, steady-state behavior is measured via the cumulative means of the speeds of edges on a traffic network. The completion times of vehicle trips were also evaluated. Two examples of virtual and real-world networks were tested with Simulation of Urban MObility (SUMO) traffic simulator, and the results show that: i) for the given networks, the settling times were 0.63 and 0.73 hours for virtual and real-road networks respectively, and ii) the ratio between the mean of settling times and the mean of trip completion times for the real-world network is higher than the one for the virtual network, approximately 20 times versus 10 times longer respectively. An important insight of this work is that the required simulation time for the traffic simulation to converge to steady-state behavior is a large multiplier of the average trip time, and this duration is far longer than what is often seen in the literature.

Keywords: Transportation Systems, Intelligent Autonomous Vehicles, Automotive Systems
Abstract: Cellular-vehicle-to-everything (C-V2X) technology is a promising solution for enhancing vehicular communication and augmenting road safety. Its efficacy depends greatly on its deployment within the built infrastructure and the environmental and operational conditions in the traffic network. This study uses experimental data collected from a test track and road sites to investigate the performance and coverage of C-V2X in work zones. The findings indicate that C-V2X needs an unobstructed line of sight between the vehicle's onboard unit (OBU) and the roadside unit (RSU) for maximum functionality. The results show substantial coverage deterioration due to the presence of signal-blocking elements including structures, elevation changes, work zone vehicles, and vegetation. A key challenge in the live work zone deployments was lack of access to high-elevation locations for antenna mounting, and a general lack of power sources within active work zones. Across the test sites in the field, the typical blockages in the work zone that resulted in an effective range of C-V2X was approximately 500 meters, roughly one-fifth of the maximum measured line of sight range of 2260 meters. These findings are useful for estimating the number of CV2X units needed to achieve full coverage within work zone sites.

Keywords: Human-Machine and Human-Robot Systems, Machine Learning in modeling, estimation, and control, Robotics
Abstract: A critical goal in robotics and autonomy is to teach robots to adapt to real-world collaborative tasks, particularly in automatic assembly. The ability of a robot to understand the original intent of an incomplete assembly and complete missing features without human instruction is valuable but challenging. This paper introduces 3D combinatorial assembly completion, which is demonstrated using combinatorial unit primitives (i.e., Lego bricks). Combinatorial assembly is challenging due to the possible assembly combinations and complex physical constraints (e.g., no brick collisions, structure stability, inventory constraints, etc.). To address these challenges, we propose a two-part deep reinforcement learning (DRL) framework that tackles teaching the robot to understand the objective of an incomplete assembly and learning a construction policy to complete the assembly. The robot queries a stable object library to facilitate assembly inference and guide learning. In addition to the robot policy, an action mask is developed to rule out invalid actions that violate physical constraints for object-oriented construction. We demonstrate the proposed framework’s feasibility and robustness in a variety of assembly scenarios in which the robot satisfies real-life assembly with respect to both solution and runtime quality. Furthermore, results demonstrate that the proposed framework effectively infers and assembles incomplete structures for unseen object types.

Keywords: Linear Control Systems, Mechatronic Systems, Motion and Vibration Control
Abstract: In this paper, we exclusively address the problem of controlling the second derivative (namely acceleration in a mechanical system) of a simple 2nd order Linear Time Invariant State Space (LTISS) system. We argue that for controlling (i.e. assuring asymptotic convergence to zero) of the acceleration, it is not sufficient to adopt a proportional plus integral and a derivative (PID) combination controller, as is currently being done in the current literature control systems design practice. It is shown that such PID type controller can never promise asymptotic convergence to the origin. It is then shown that it is necessary to employ a controller that promises asymptotic convergence of at least 4 (out of the six) real state variables that exist in the original 2nd order state space system. In other words, in addition to the 2 real state variables of the original state space system, another 2 real state variables belonging to the symmetric matrix space and possibly 1 more real state variable belonging to the skew-symmetric part of the system have to be promised to be convergent as well. If the the skew-symmetric space has a single complex conjugate (or pure imaginary) set of state variables, then at least all the 4 real state variables belonging to the original and symmetric part state space have to made convergent. Examples are provided to illustrate such controller design, which is labeled as control design for Convex Stability.

Keywords: Mechatronic Systems, Modeling and Validation, Electromechanical systems
Abstract: Lorentz actuators use magnetic interactions between coils and magnets to create forces and torques for actuation. Especially for applications that enable actively controlled 6-DoF levitation it is necessary to model these Lorentz forces and torques and evaluate these models in real time. Generally, Lorentz forces are evaluated using numerical integration of the Lorentz force integrals. These calculations become computationally expensive for increasing number of magnets and, therefore, deriving analytical or semi-analytical models for the integrals are important to reduce computation time while maintaining model accuracy. For the studied superconductor-based 6-DoF levitation system a semi-analytical model of the Lorentz actuators forces/torques is derived by exploiting the periodicity of the actuator magnets magnetization. Using Fourier series an analytical expression of the magnetic flux density of the actuator magnets is found by solving Maxwell’s equations under magneto-static conditions. Using this analytical expression a semi-analytical formulation of the Lorentz force integrals is given. In this publication the model derivation is presented, and the model is validated and discussed based on single coil experiments and compared to the fully numeric surface charge model. It was possible to show that the semianalytical model can closely match the measured forces and torques even with a reduced set of approximation elements to reduce computational effort by over 70% with respect to a fully numeric implementation.

Keywords: Modeling and Validation, Sensors and Actuators, Mechatronic Systems
Abstract: This paper investigates analytically and experimentally the mechano-chemo-electrical behavior of ionic polymer composite (IPMC) and engineered IPMC (eIPMC) sensors under extensional loading. To predict the sensing response of eIPMCs, a detailed model is proposed incorporating a composite layer (CL) for the abraded interface between polymer and electrode. We present open circuit voltage and short circuit current sensing predictions derived from this model and we validate them via experiments on anisotropic extensional loading of IPMCs. Experimental results demonstrate that our sensors' electrical outputs align well with theoretical predictions, thereby validating our findings and enhancing our understanding of eIPMC strain sensor behavior.

Keywords: Nonlinear Control Systems, Control Design, Control Applications
Abstract: This paper investigates extremum seeking control for a torque-controlled antenna pointing system without direct angular measurements. We consider a two-degree-of-freedom (2-DOF) antenna system that receives an unknown signal from its environment, where the signal strength varies with the antenna’s orientation. It is assumed that only real-time measurements of the signal are available. We develop an extremum seeking control strategy that enables the antenna to autonomously adjust its direction to maximize the received signal strength based on the symmetric product approximation. Under suitable assumptions on the signal function, we prove local practical uniform asymptotic stability for the closed-loop system.

Keywords: Estimation
Abstract: This study focuses on identification of the model parameters of an artificial cochlea. For this purpose, a least squares estimator incorporated with a filtering approach, is proposed to accurately determine the parameters of the artificial hair cell model. The proposed approach leverages readily available standard signals from the cell, thereby eliminating the need for direct acceleration measurements. Utilization of the proposed method enables efficient and practical parameter identification. To validate the effectiveness of the proposed approach, numerical simulations are provided to illustrate the successful identification of uncertain model parameters.

Keywords: Stochastic Systems, Nonlinear Control Systems
Abstract: This paper establishes Lyapunov-like sufficient conditions for the ultimate boundedness of discrete-time stochastic nonlinear dynamical systems. The concept of ultimate boundedness in probability is established, along with the definition of a stochastic ultimate bound time to capture the average convergence behavior of the system. Sufficient conditions for ultimate boundedness in probability is developed using a stochastic comparison function. The upper bound of ultimate bound time, characterizing the practical finite time convergence, is characterized for nondecreasing concave functions. Specifically, two distinct types of nondecreasing concave convergence results are introduced to guarantee ultimate boundedness. Finally, a numerical example is provided to illustrate global ultimate boundedness in probability.

Keywords: Uncertain Systems and Robust Control, Manufacturing Systems, Control Applications
Abstract: Laser Powder Bed Fusion (LPBF) is a metal additive manufacturing process that uses a high-power laser to melt a predefined shape in a bed of metal powder, layer by layer. The size of the melted pool throughout the process can significantly affect the mechanical properties of the final part; too small of a melt pool may result in poor fusion, too large will cause porosity. The size of the melt pool is governed by inherently complex multi physical interactions. Complex models have been developed and simplified in the literature, and in this paper, a nonlinear first order single state energy transfer model is used to simulate the size of the melt pool transverse surface area. The error is defined as the difference between the melt pool area and a desirable reference value, and a sliding mode control (SMC) law is developed to use input laser power to drive the system to a zero-error manifold in finite time. Since the model used takes advantage of potentially unrealistic geometrical assumptions about the melt-pool shape, the control law is further developed to be robust to inaccuracies and real-time changes in the system parameters related to this assumption. The performance of the controller is compared with other control strategies in the presence of bounded parameter uncertainty.

Keywords: Uncertain Systems and Robust Control, Motion and Vibration Control, Mechatronic Systems
Abstract: This paper presents a case study of the recently proposed nonlinear indirect adaptive robust control (IARC) approach to disturbance observer (DOB) design. Specifically, the recently proposed lumped uncertainties based IARC approach makes full use of available prior process information, such as disturbance bounds, to construct a controlled nonlinear estimation/compensation that enables the complete separation of feedback control designs with guaranteed robust stability and performance from the lumped uncertainty estimation. As a result, the well-established robust control theory of mu-synthesis can be used for the underlying feedback control designs, while the severe performance limitation of Q-filer design in traditional DOB for robust stability is naturally overcome. This paper focuses on the application of the approach to the precision motion control of a motor-driven servo-table with significant flexible modes and bounded uncertain nonlinearities. Comparative simulations are presented to demonstrate the effectiveness of the proposed approach.

Keywords: Uncertain Systems and Robust Control, Nonlinear Control Systems, Assistive and Rehabilitation Robotics
Abstract: People with hemiparesis following a stroke can experience balance deficits, which can lead to increased risks of falls. Most falls in people post-stroke occur due to loss of balance while walking. Lower-limb exoskeletons have been developed to assist and restore leg function during walking. However, existing exoskeleton controllers have primarily focused on improving gait energetics and providing assistance in the sagittal plane. People control balance during walking by adjusting the step width, using mediolateral (ML) leg control, i.e., control in the frontal plane. To prevent falls and injuries, recent efforts in the robotics community have focused on providing ML assistive and training strategies, albeit mainly in an open-loop fashion or using neuromuscular perturbation controllers. In this paper, a closed-loop robust sliding-mode controller is developed for a two-degree-of-freedom (DoF) robotic exoskeleton that assists hip motion in both the sagittal and frontal planes (i.e., customized frontal-plane assistance) during the swing phase of walking. A nonlinear Euler-Lagrange system with parametric uncertainty and exogenous disturbances is introduced to model the paretic leg and device. Exponential tracking of the hip frontal and sagittal trajectories is guaranteed via a Lyapunov-based stability analysis. Experimental results in one able-bodied individual were obtained in three treadmill walking trials at 0.5 m/s with 0%, 15%, and 30% frontal deviations from baseline. RMS joint tracking errors are 0.02±0.01 and 0.021±0.01 rad for frontal assistance, and 0.048±0.026 and 0.039±0.022 rad for sagittal assistance in the 15% and 30% deviation trials, respectively, demonstrating the feasibility of the developed approach.

Keywords: Uncertain Systems and Robust Control, Nonlinear Control Systems, Optimal Control
Abstract: The performance of autonomous vehicles degrades significantly in the presence of uncertainties, which is even more pronounced in high-speed off-road scenarios. It is essential to ensure robust performance in the presence of uncertainties. Model predictive control (MPC) is a popular control method used to control vehicles; however, MPC depends heavily on the accuracy of a model. Often, when there are model inaccuracies, MPC fails. Making MPC more robust against model errors is essential. Model error control synthesis (MECS) addresses the challenge of compensating for model errors by using a nonlinear estimator to determine model error corrections with a point minimization, in a one-step-ahead manner and compensates for uncertainties through a control input. In this paper, we incorporate MECS into an MPC formulation, where MECS compensates for model errors while the nominal MPC handles trajectory tracking. The resulting framework offers a robust and computationally efficient solution for trajectory tracking in highly dynamic and uncertain off-road autonomous driving.

Keywords: Uncertain Systems and Robust Control, Nonlinear Control Systems, Optimal Control
Abstract: Robust control problems have significant practical implications since external disturbances can significantly impact the performance of control methods. Existing robust control methods excel at control-affine systems but fail at neural network dynamic models. Developing robust control methods for such systems remains a complex challenge. In this paper, we focus on robust tracking methods for neural network dynamic models. We first propose a reachability analysis tool designed for this system and then introduce how to reformulate a robust tracking problem with reachable sets. In addition, we prove the existence of a feedback policy that bounds the growth of reachable sets over an infinite horizon. The effectiveness of the proposed approach is validated through numerical simulations of the tracking task, where we compare it with a standard tube MPC method.

Keywords: Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems, Multi-agent and Networked Systems
Abstract: A networked battery system (NBS) consists of multiple batteries linked to form a battery energy storage system (BESS). Intelligent NBS are formed by smart battery agents that communicate and coordinate together to optimize transients, energy usage, storage, and power allocation. Popular applications of these include DC microgrids. Traditionally, a decentralized voltage-current (V-I) droop control strategy is deployed in these systems to regulate energy distribution. However, V-I droop control operates as a power-averaging distributed algorithm, which may not be suitable for battery systems with heterogeneous batteries of varying capacities and state-of-charge (SoC) dynamics. During battery discharge, a battery with a lower SoC level or smaller installed capacity will deplete first and will no longer be able to contribute to the battery system. Therefore, maintaining SoC balancing is essential to prevent these issues. To overcome these limitations, we propose a multi-agent reinforcement learning control algorithm to optimize SoC balancing. This approach ensures that each battery maintains the same SoC level. Reinforcement learning (RL) enables data-driven learning of the optimal charging/discharging control policies even in the presence of uncertainties in the battery capacities and SoC dynamics. This multi-agent RL framework ensures optimal balancing of SoC levels with the ultimate goal of improving battery life and reducing battery maintenance costs.

Keywords: Machine Learning in modeling, estimation, and control, Power and Energy Systems, Automotive Systems
Abstract: Optimal charging of electric vehicle (EVs) depends heavily on reliable sensor measurements from the battery pack to the cloud-controller of the smart charging station. However, an adversary could corrupt the voltage sensor data during transmission, potentially causing local to wide-scale disruptions. Therefore, it is essential to detect sensor cyberattacks in real-time to ensure secure EV charging, and the developed algorithms must be readily adaptable to variations, including pack configurations. To tackle these challenges, we propose adaptable fine-tuning of an XgBoost-based cell-level model using limited pack-level data to use for voltage prediction and residual generation. We used battery cell and pack data from high-fidelity charging experiments in PyBaMM and ‘liionpack’ package to train and test the detection algorithm. The algorithm’s performance has been evaluated for two large-format battery packs under sensor swapping and replay attacks. The simulation results also highlight the adaptability and efficacy of our proposed detection algorithm.

Keywords: Power and Energy Systems, Control Applications, Adaptive and Learning Systems
Abstract: The charging of batteries presents an inherent conflict — the charge rate should be maximized, while the degradation should be minimized. To balance these competing objectives, optimal control strategies have been applied to the charging of lithium-ion batteries. However, most optimal control strategies require a model, which can be costly to develop and prohibitive to rapid battery technology development. In this work, we propose a behavior-informed, direct data-driven control approach using Data-enabled Predictive Control (DeePC). DeePC allows for an optimization that exploits directly learned behaviors that are guided by known system interactions, such as capacity loss due to interface failures, as a function of the inputs without an explicit model provided. Numerical simulations using high-fidelity, mature, publicly available models and frameworks validate the optimality of the proposed approach.

Keywords: Power and Energy Systems, Estimation
Abstract: Detecting internal short circuits (ISCs) in a single cell connected in parallel with multiple other cells is challenging due to unmeasured internal currents among the parallel-connected cells, which can obscure measurable outcomes, such as charge loss and associated voltage drop from the faulty cell. In this work, we propose a new method for detecting ISCs in parallel-connected cells based on the interacting multiple model (IMM) estimation technique. This method can provide a probability for the occurrence of an ISC and, simultaneously, estimate the short-circuit resistance, which indicates the severity of the ISC. Our algorithm relies on models of the dynamic electrothermal behavior of parallel-connected cells (nP), in which one of the cells is modeled in both healthy and faulty modes. In the faulty mode, the ISC is represented as a resistor in parallel to the cell's internal resistance, and this parallel combination is in series with a tab resistance. The proposed methodology utilizes interacting unscented Kalman filters (UKFs) to detect the faulty mode and estimate the ISC resistance based on the voltage drop and the temperature increase caused by the unknown ISC. The efficacy of the methodology is demonstrated across various synthetic data sets that are corrupted by realistic levels of voltage and temperature sensor noise, assuming Gaussian characteristics. We simulate fifty short-circuit scenarios in a 46P parallel cell group during a discharging dynamic drive cycle. Short-circuit resistances ranged from hard shorts (0.5 Ohm) to soft shorts (100 Ohm), tested across ten different SOCs and corresponding voltage levels. Our IMM-based method successfully detected and estimated the ISC in all fifty cases. In all simulated ISC scenarios, the temperature rise remained below 6◦C before detection, which is well before the onset of thermal runaway conditions.

Keywords: Power and Energy Systems, Machine Learning in modeling, estimation, and control, Automotive Systems
Abstract: Accurate State-of-Charge (SOC) estimation is critical for the safe and efficient operation of energy storage systems such as Li batteries, yet it remains challenging due to their nonlinear dynamics and the dependability of SOC behavior on operating conditions. This paper introduces an ensemble data-driven framework to create a single parsimonious temperature-dependent SOC model using experimental data collected at various operating conditions. The modeling approach is based on the Sparse Identification of Nonlinear Dynamical Systems (SINDy) framework to construct interpretable and parsimonious SOC models. Starting with a baseline model at 25°C, the method incorporates terms inspired by electrochemical principles (e.g., Fick’s law and Butler-Volmer kinetics) into a library of candidate functions. The baseline model is constructed by minimizing a novel cost function that balances accuracy and complexity using training and validation errors and the number of model terms. The ensemble method utilizes bootstrapping aggregate data to capture varying behaviors within the chosen datasets by choosing appropriate library terms. The bootstrapping process creates multiple datasets by randomly selecting samples with replacement to enhance robustness and accuracy within limited data. For each generated bootstrap dataset, a new SOC model is developed. The final ensemble model aggregates these models via the median of their sparse vector of coefficients. This approach effectively filters out the outlier terms and the noise in the data, while preserving physical interpretability. The ensemble also enables uncertainty quantification through coefficient distribution analysis. The framework was validated experimentally on a 21700 Li-ion cell. The experiments involved custom and standard driving profiles across six different temperatures from -10°C to 40°C. The ensemble model demonstrated superior accuracy and generalizability across all temperatures compared to a model trained on a combined dataset. Notably, the bootstrap-derived model reduced the SOC estimation RMSE by over 50% at low temperatures and maintained an estimation accuracy below 2.5% error on unseen test data.

Keywords: Modeling and Validation, Automotive Systems, Modelling, Identification and Signal Processing
Abstract: A battery management system (BMS) relies on accurate battery models to perform predictions, where, equivalent circuit models (ECMs) are most practical. However, batteries possessing significant hysteresis cannot be effectively modeled with conventional ECMs, limiting BMS accuracy. Therefore, this study improves ECM accuracy by addressing hysteresis, a key lithium-ion (Li-ion) cell characteristic. In this context, a second-order Thevenin ECM, incorporating dynamic and instantaneous hysteresis (Plett, 2015), was developed and tested on high-nickel-ternary-cathode (HNTC) Li-ion cells. When working with hysteresis models, the hysteresis tuning rate (gamma) is an estimated parameter often defined as a constant and is overlooked in adjusting the rate of hysteresis decay. In this work, a variable gamma was identified and examined for different state of charge (SOC) intervals and temperatures. An n-way analysis of covariance (ANCOVAN) was applied to the variable response of gamma and indicated a significant gamma-hysteresis voltage relationship for specific SOC windows. Comparatively through validation of dynamic and total hysteresis modeling, results also indicate that incorporating instantaneous hysteresis does not necessarily improve overall model performance.

Keywords: Machine Learning in modeling, estimation, and control, Power and Energy Systems
Abstract: Battery charging protocols play a critical role in battery management as they directly impact operational efficiency, safety, and aging. While charging protocol optimization requires systematic exploration of diverse charging patterns, conventional charging protocol generation methods face limitations in comprehensiveness, adaptability, and optimization tractability due to restricted pattern variability and high-dimensional parameter spaces. To address these challenges, this paper proposes a variational autoencoder (VAE)-based framework for universal battery charging protocol generation. By leveraging a VAE model trained on a randomized dataset, the framework encodes any given protocol into a compact 3-dimensional latent space, enabling efficient sampling and optimization. A merge-based data augmentation strategy synthesizes heterogeneous protocols for training to enhance the decoder’s ability to accurately reconstruct known charging protocols and generate novel, feasible ones. These capabilities are demonstrated using battery charging protocols from four different electric vehicles (EVs). This work offers a promising approach for both universal battery charging protocol generation and optimization in battery management systems.

Keywords: Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems, Power and Energy Systems
Abstract: Accurate modeling of lithium-ion battery (LiB) packs is essential for optimizing performance, longevity, and safety. This involves capturing the complex electro-thermal dynamics within individual cells (or pack segments) as well as the interactions between these segments. We present a novel hybrid modeling approach that combines structurally identifiable distributed equivalent circuit model (ECM) and lumped thermal (LT) model of each segment with a pack level Transformer-based neural residual. Leveraging the Transformer's self-attention mechanism, the residual enables data-based high-resolution modeling of dynamical complexities by capturing temporal dependencies and the spatial interactions between pack segments, which are difficult to capture with interconnections of simple ECMs. The training scheme for this hybrid model is designed to maximize explainability by prioritizing learning of the distributed ECM< parameters. Considering a demonstrative high-fidelity model of an LiB pack, the learned hybrid model is shown to successfully predict state trajectories of each cell for long sequences up to voltage and temperature RMSEs of 0.011V and 0.2°C, respectively. This paves the way for adaptive and accurate predictive pack control, including battery balancing and thermal management.

Keywords: Power and Energy Systems, Modeling and Validation, Estimation
Abstract: All-Solid-State Lithium-Sulfur Batteries (ASLSBs) offer a superior alternative to traditional lithium-ion batteries, given their potential for higher energy densities and enhanced safety. This paper explores ASLSBs under controlled stress and constant strain scenarios, develops data-driven models that predict cell voltage given input current, and correlates observations with cell states. The empirical data collected under varying operational pressure conditions are used to perform system identification of equivalent circuit model parameters that account for variations in the externally applied force resulting from dynamic operational pressure. Cell capacity fades rapidly, but correlations are discovered between usable energy, cell impedance, and applied pressure. Most notable are observations of capacity recovery, which show that reduced capacity due to operations at lower pressure is reversible. This work also demonstrates that variations in battery stress over time in a constant volume provide a means to estimate the cell's State of Charge. This study highlights the crucial role of pressure in enhancing the performance of ASLSBs.

Keywords: Control Design, Large Scale Complex Systems, Control Applications
Abstract: DC microgrids offer cost-effective solutions to provide reliable electricity to remote and disadvantaged communities, presenting a promising path toward energy equity. However, when powered by renewable energy sources, these microgrids face stability challenges due to DC bus voltage fluctuations from renewable energy uncertainty due to weather conditions and ripple effects from AC/DC converters. Such disturbances, particularly in DC microgrids with various converter types and constant power loads (CPLs), can compromise system stability. This paper introduces a sliding-mode control approach that minimizes DC bus voltage deviations and maintains stability in DC microgrids with CPLs. The research expands this methodology to multi-area DC microgrids, ensuring stability across larger systems with multiple power resources.

Keywords: Cyber physical systems, Distributed Parameter Systems, Security and Privacy
Abstract: Cyber-Physical Distributed Parameter Systems (DPS) governed by Partial Differential Equations (PDEs) face unique security challenges due to their spatially distributed dynamics and reliance on sensors placed only at discrete-points. While studies on ODE-modelled systems address the impact of temporal-only cyber-attacks, PDE-modelled systems require additional analysis of spatio-temporal vulnerabilities that can arise from cyber-attacks on sensors and actuators in DPS. Motivated by this need, this paper analyzes the effects of coordinated actuator-sensor attacks on a class of DPS, modelled by linear parabolic PDEs – with distributed actuation and boundary measurement. First, we propose a framework for defining adversarial metrics that can capture the effect of attacks in a DPS setting. Subsequently, we derive mathematical conditions under which adversarial objectives can be achieved in terms of the previously defined metrics. Finally, a case study on battery module thermal behavior, based on the proposed adversarial metrics and mathematical conditions, illustrates how cyber-attacks can potentially affect the operation of DPS.

Keywords: Power and Energy Systems
Abstract: This paper presents an advanced optimization framework for the optimal dispatch of electric vehicle (EV) charging stations integrated with second-life batteries (SLBs) and photovoltaic (PV) generation. The proposed formulation explicitly models EV demand dynamics using arrival and departure times, energy targets, and plug constraints, alongside laxity-based charging policies. The framework incorporates detailed operational constraints for battery energy storage systems, including charging/discharging limits, round-trip efficiency, and energy balance. A mixed-integer linear programming model is developed to maximize station profitability by optimizing grid energy purchases, solar power utilization, EV charging revenues, and battery operations. Through case studies with different levels of SLB degradation and varying numbers of EV charging ports, we demonstrate the effectiveness of the proposed optimization model in improving economic performance, enhancing energy efficiency, and meeting EV demand under realistic operational constraints.

Keywords: Automotive Systems, Optimal Control, Power and Energy Systems
Abstract: The lack of charging infrastructure has become one of the most significant barriers for widespread adoption of battery electric vehicles (BEVs). Mobile charging stations (MCSs) with integrated batteries is a promising charging alternate due to their flexible deployment and enhanced grid stability. In addition, using retired BEV batteries as a batter energy storage system (BESS) for MCSs further reduces equipment cost, which makes MCSs economically viable. MCSs discharge energy from the BESS therein to BEVs, and the discharging strategy needs to be well designed to address the safety risks posed by the degradation of second-life battery (SLB) capacity and performance during secondary use. Although the optimal charging strategies from the chargers to BEVs have been well studied, the optimal discharge strategy from the BESS-integrated MCS side has received little attention. To fill in this gap, this paper develops an optimal aging-aware discharging strategy specifically for SLB-integrated MCS applications using Pontryagin’s Minimum Principle (PMP). The discharging time and current profile of the SLBs are optimized to extend the battery life and reduce the risk of thermal runaway. The simulation results show that the proposed strategy effectively reduces capacity loss without significantly extending discharging time, while maintaining temperature within the optimal range compared to conventional constant current discharging methods. Additionally, the performance of the method is proven on different SLB-integrated MCSs with different State of Health (SoH).

Keywords: Automotive Systems, Optimal Control, Power and Energy Systems
Abstract: Range anxiety caused by moderated battery performance and accelerated capacity loss at low temperatures is regarded as one of the most significant barriers to the adoption of battery electric vehicles (BEVs). Preheating the battery using grid power has been proven to be a convenient and effective method for BEV users to maintain driving range and extend battery life. Traditional preheating strategies typically heat the batteries right before the vehicle departs, which helps improve operating conditions during battery discharge. However, the impact of battery temperature during charging is often overlooked. Charging batteries at extreme sub-zero temperatures can lead to consequences similar to those of discharging under such conditions, but simply preheating before charging may shift energy usage to peak hours under time-of-use electricity pricing and increase the charging cost. To address this challenge, this paper proposes an optimal coordinated charging and preheating strategy for BEVs to minimize operational costs and extend battery life during level 2 overnight charging. An electro-thermal-aging coupled model is used to describe the battery dynamics under different conditions. Results of a case study under -20 °C show that over 80% saving on the operational cost is achieved by the proposed strategy comparing to the conventional strategy and the optimized strategy neglecting accelerated battery aging under low temperature.

Keywords: Automotive Systems, Power and Energy Systems, Modeling and Validation
Abstract: Lithium-ion batteries are the enabling power source for transportation electrification. However, in real-world applications, they remain vulnerable to internal short circuits (ISCs) and the consequential risk of thermal runaway (TR). This paper develops BattBee, the first equivalent circuit model to specifically describe the onset of ISCs and the evolution of subsequently induced TR. Drawing upon electrochemical modeling, the model can simulate ISCs at different severity levels and predict their impact on the initiation and progression of TR events. By design, this model offers strong physical interpretability and predictive accuracy, while maintaining structural simplicity to allow fast computation. Validation based on experimental data demonstrates the BattBee model’s effectiveness, showing its potential for practical battery safety risk management in electric vehicles and other high-stakes applications.

Keywords: Control Applications, Estimation, Linear Control Systems
Abstract: This paper presents a method for estimating and compensating the transformation angle error in Indirect Rotor Flux-Oriented Control (IRFOC) of an induction machine (IM), caused by inaccuracies in electrical parameters. This method aims to improve the robustness of torque control in IM against variations in motor parameters. The proposed approach leverages an Extended Kalman Filter (EKF) to efficiently estimate the rotor flux. The estimated flux is then used to compute the flux position error from an analytical relation. This method operates on multiple reference frames, including alphabeta, dq, and a deviated rotating reference frame known as the d'q’ reference frame. The approach presented in this paper eliminates the need for online estimation of motor electrical parameters such as rotor resistance and mutual inductance. Simulation results are provided to demonstrate the effectiveness of the proposed approach.

Keywords: Control Applications, Power and Energy Systems, Optimal Control
Abstract: This paper addresses the challenge of pressure constraint violations in water electrolysis systems operating under dynamic power conditions, a problem common to both Proton Exchange Membrane and alkaline technologies. To investigate this issue, a control-oriented model of an alkaline electrolyzer is developed, capturing key pressure and flow dynamics. To manage rapid power fluctuations that may cause pressure to exceed manufacturer-defined operational boundaries, a model-based constraint-aware power governor based on the Reference Governor (RG) framework is proposed. Simulation results show that the strategy effectively maintains pressure within the specified operating range, outperforming conventional filtering methods while enhancing hydrogen production and reducing auxiliary energy consumption.

Keywords: Control Applications, Robotics, Estimation
Abstract: Handling objects with unknown or changing masses is a common challenge in robotics, often leading to errors or instability if the control system cannot adapt in real-time. In this paper, we present a novel approach that enables a six-degrees-of-freedom robotic manipulator to reliably follow waypoints while automatically estimating and compensating for unknown payload weight. Our method integrates an admittance control framework with a mass estimator, allowing the robot to dynamically update an excitation force to compensate for the payload mass. This strategy mitigates end-effector sagging and preserves stability when handling objects of unknown weights. We experimentally validated our approach in a challenging pick-and-place task on a shelf with a crossbar, improved accuracy in reaching waypoints and compliant motion compared to a baseline admittance-control scheme. By safely accommodating unknown payloads, our work enhances flexibility in robotic automation and represents a significant step forward in adaptive control for uncertain environments.

Keywords: Estimation, Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems
Abstract: In this paper, we present an unscented Kalman filter (UKF) and an unscented Kalman smoother (UKS) algorithm for wind estimation using a quadcopter. The presented algorithms utilize a learning framework for modeling the drag force, which combines a parametric model to capture the nominal behavior and a nonparametric Gaussian Process Regression (GPR) to learn the residual dynamics. Based on different combinations of this learning approach, four variants of the UKF and UKS are evaluated. A comparative performance analysis of the variants with flight test data shows that using the parametric model as a nominal drag model and the nonparametric GPR to account for uncertainties in the learning process of the drag model provides the best performance in both wind estimation and quantifying its associated uncertainty.

Keywords: Estimation, Underwater Vehicles
Abstract: Underwater localization and mapping using scanning sonar is critical for autonomous underwater vehicles (AUVs) where Global Positioning System (GPS) and vision are unreliable. Single-beam scanning sonar provides a cost-effective sensing option, but its low update rate and sensitivity to vehicle motion often result in distorted sonar images and poor mapping performance. In this paper, we present a low-cost sonar map reconstruction framework that tightly integrates a single-beam scanning sonar with an onboard inertial measurement unit (IMU). Our method employs an Error-State Kalman Filter (ESKF) for motion estimation and compensation, and utilizes Generalized Iterative Closest Point (GICP) with RANdom Sample Consensus (RANSAC) for point-cloud registration and outlier rejection. To mitigate sonar distortion, we introduce a motion-compensated mapping approach that reconstructs accurate and consistent sonar maps in real time. Experimental results conducted in a controlled pool environment demonstrate the effectiveness of our method in enabling reliable pipeline localization and navigation using a low-cost remotely operated vehicle (ROV), offering a practical solution for underwater tasks in GPS-denied and visually degraded settings.

Keywords: Control of Smart Buildings and Microgrids, Multi-agent and Networked Systems, Automotive Systems
Abstract: Effective microgrid coordination relies on consensus algorithms, but network delays create asynchrony, complicating their performance. While the Uniform, Metropolis, and Mean Metropolis iteration matrices are standard for improving synchronous consensus, their effectiveness under asynchronous conditions remains underexplored. This paper systematically evaluates the impact of asynchrony on these matrices by simulating a 16-agent network across dense, sparse, and intermediate topologies. By analyzing convergence speed, error bounds, and error variance, we find the Uniform matrix yields the highest accuracy, while the Metropolis and Mean Metropolis matrices offer faster convergence but with greater error variability. These results clarify the accuracy-efficiency trade-off for designing asynchronous microgrids.

Keywords: Path Planning and Motion Control, Human-Machine and Human-Robot Systems, Unmanned Ground and Aerial Vehicles
Abstract: Use of Autonomous Guided Vehicles (AGV) in factory and warehouse environments have become very popular recently, as these vehicles offer tremendous opportunity to optimize the logistics and transportation of materials. However, use of AGVs in a particular region of a factory restricts the use of other Manually Operated Vehicles (MOVs) or presence of other personnel. Hence, it is desired to develop algorithms that can assist plan paths of MOVs that would avoid any intersection with AGVs operating in the same space. Also, in many cases involving the use of a commercial system for AGV fleet management, the entire AGV paths are not accessible. In these cases, MOV path planning needs to be done with partial information that requires prediction of AGVs' paths. In this paper, we present algorithms for planning paths of MOVs in an actual companies warehouse with predetermined but not fully known Autonomous Ground Vehicle (AGV) paths. A shortest and a fastest method based on Dijkstra is compared to methods using A*, distance to goal and computing time are used as criteria for comparison. The actual time required to reach the goal is also analyzed. The paper also presents results based on simulation of vehicles' paths in 3D digital twin environment developed using NVIDIA Isaac Sim based on real-world warehouse.

Keywords: Path Planning and Motion Control, Intelligent Autonomous Vehicles, Robotics
Abstract: Autonomous vehicle navigation is a basic functionality for ground robots, with applications within research and real-world scenarios. The paper discusses the configuration and utilization of ROS2's native tools and packages to deploy an entire indoor path planning system on a differential-drive robot, in a simulation setup. The BumperBot platform, obtained from an open-source GitHub repository, receives simulated odometry and lidar sensor data, and employs Graph SLAM for mapping, and Adaptive Monte Carlo Localization (AMCL) for pose estimation through the slam_toolbox package and configuration files. Path planning and control are done with the Nav2 stack, with the addition of a custom Python node to follow predefined waypoints. Experimental results on the Gazebo simulation platform confirm the usefulness of ROS2's modular navigation stack for enabling autonomous point-to-point navigation using a minimum of custom code. One of the key advantages of this approach is that the same software framework can be easily adapted to other differential-drive robots with menial changes, thanks to ROS’s abstraction layers and standardized interfaces–eliminating the need to start from scratch when upgrading or swapping hardware. The framework presented provides a reproducible and extensible foundation for future work, such as integration with aerial waypoint generation and object detection modules.

Keywords: Path Planning and Motion Control, Robotics, Intelligent Autonomous Vehicles
Abstract: The ability to update a path plan is a required capability for autonomous mobile robots navigating through uncertain environments. This paper proposes a re-planning strategy using a multilayer planning and control framework for cases where the robot's environment is partially known. A medial axis graph-based planner defines a global path plan based on known obstacles, where each edge in the graph corresponds to a unique corridor. A mixed-integer model predictive control (MPC) method detects if a terminal constraint derived from the global plan is infeasible, subject to a non-convex description of the local environment. Infeasibility detection is used to trigger efficient global re-planning via medial axis graph edge deletion. The proposed re-planning strategy is demonstrated experimentally.

Keywords: Path Planning and Motion Control, Robotics, Modelling and Control of Biomedical Systems
Abstract: Geodesic path planning, which computes the shortest paths on curved surfaces, is critical in robotics, computer graphics, and autonomous navigation. In medical applications, such as minimally invasive procedures, precise robotic guidance through complex anatomical structures is essential to reduce procedure time, minimize tissue damage, and improve patient outcomes. This paper presents a geodesic path planning framework for a miniature robot tasked with monitoring and treating multiple tumor sites on the liver surface. The proposed method first calculates geodesic distances between all tumor pairs based on the liver’s three-dimensional geometry, then solves a Traveling Salesman Problem (TSP) to determine the optimal visitation sequence. Simulations demonstrate that this approach generates smooth, efficient trajectories adhering to the liver’s curvature, offering a viable solution for multi-target robotic navigation in surgical applications.

Keywords: Path Planning and Motion Control, Transportation Systems, Modeling and Validation
Abstract: Emergency evacuation planning is a critical task that ensures individuals' swift and safe movement from hazardous locations to designated safe zones. This paper proposes a novel textit{controllable equivalent circuit model} (cECM) to capture the flow of traffic in a network. The model includes switches for flow routing and additional circuit elements to represent the time lost due to recharging, which is critical for (electric) vehicles with limited driving range. We embed this cECM into a mixed-integer linear program (MILP) that determines optimal evacuation routes and recharging strategies to minimize evacuation time while considering road capacity, limited vehicle energy, and the availability of charging stations. The effectiveness of the proposed approach is demonstrated through simulations based on an evacuation case study using the Sioux Falls and Anaheim networks.

Keywords: Path Planning and Motion Control, Unmanned Ground and Aerial Vehicles, Robotics
Abstract: With an accurate map of the environment, a map-based planner seeks a global path that avoids collisions with obstacles. However, if map errors in obstacle size and position are discovered during path execution where corridor width is insufficient, replanning may be necessary to divert from the original path plan, yet this may be impossible or may incur severe path costs. A common method to plan paths uses visibility graphs, and the algorithm herein leverages these to estimate corridor width by employing geometric ray projections implicit in determining point-to-point connectivity. Thus, if the visibility graph is already being calculated, this method requires little additional computational effort to enable path planning in consideration of corridor width. Simulations are presented that use county-scale floodplain maps in central Pennsylvania to analyze the ability to incentivize corridor width to navigate between start-goal pairs. The results show that, with corridor incentives, replanning is 30% more likely to be feasible for intermediate obstacle dilation values. Further, pre-planned corridor-aware paths are feasible with approximately twice as much size dilation versus nominal visibility graph planning methods.

Keywords: Automotive Systems, Control Design
Abstract: This paper presents an optimization-based framework for sizing hybrid battery systems in electric vehicles, combining high-energy (HE) and high-power (HP) chemistries. The framework aims to minimize the purchase cost, weight, and environmental footprint of the hybrid battery pack, while accounting for practical power and energy constraints—modeled via Ragone plots—as well as thermal, electrical, mass, and volume limits. A Savitzky-Golay filter is employed to perform the power split between the two batteries, allocating average power demands to the HE battery and pulsed power to the HP battery. The proposed methodology is applied to size a hybrid battery pack, composed of Nickel Manganese Cobalt (NMC) and Lithium Titanium Oxide (LTO) cells, across multiple driving cycles. Results demonstrate that the hybrid configuration can reduce overall mass by up to 60% when compared to single-chemistry batteries.

Keywords: Power and Energy Systems, Modeling and Validation, Sensors and Actuators
Abstract: Proton exchange membrane fuel cells (PEMFCs) have become a practical and promising alternative energy source for the automotive industry, offering high efficiency and zero-carbon emissions compared to internal combustion engines. Sustainable fuel cell (FC) operation in vehicles is dependent on effective water management. This study investigates impact of local, ambient relative humidity (RH) level variation on the water content and overall performance of a 5-kW open-cathode PEMFC. A Simcenter Amesim model of the FC was first developed, calibrated, and validated using experimental data. The model was tested under various scenarios to comprehensively analyze how changes in local RH affect membrane water content (𝜆), reflected in the polarization curve, FC power, and ohmic losses. The results show that FC efficiency improves by 23%, while ohmic resistance drops from 0.72 Ω·cm² to 0.26 Ω·cm² as the local RH increases from 10% to 90%.

Keywords: Optimal Control, Automotive Systems, Control Design
Abstract: Effective control of anode relative humidity is essential for maintaining the balance required for optimal membrane hydration in proton exchange membrane (PEM) fuel cells. This is particularly challenging in open-cathode PEM fuel cells, where the input air must simultaneously supply reactant oxygen, provide cooling and regulate water content. The strong coupling between these processes, especially between temperature and humidity, makes conventional control strategies less effective. This study employs a Koopman model predictive control (KMPC) approach to regulate anode humidity in a 5kW open-cathode PEM fuel cell stack. Unlike traditional controllers, MPC can account for multivariable interactions and constraints. The performance of the KMPC is compared to a baseline proportional-integral (PI) controller. Results indicate that humidity control has a greater influence on cell performance than temperature, especially for medium to high current density regions. While both controllers achieve similar reference tracking performance, KMPC effectively adjusts control effort based on operating conditions, improving overall efficiency. Additionally, the linear nature of the Koopman MPC allows for the use of computationally efficient linear MPC solvers, making it well-suited for real-time PEM fuel cell control while allowing adaptable weight tuning in varying operating conditions.

Keywords: Adaptive and Learning Systems, Machine Learning in modeling, estimation, and control, Linear Control Systems
Abstract: Accurately estimating battery State of Charge (SOC) is essential for safe and optimal electric vehicle operation and requires fast and adaptive battery models. Physics-based battery models capture the electrochemical battery dynamics leading to a higher fidelity of state prediction but are computationally expensive to solve. Empirical or data-driven models can be implemented online as they do not explicitly model the complex reaction mechanism but require extensive data for calibration. This paper presents a Koopman operator based algorithm that combines data with physics-based features to bridge the gap between the physics-based and abstract models in accurately estimating the battery SOC during the lifetime of the electric vehicle. The proposed approach produces a fast interpretable linear model in state space form. A derivative-free Unscented Kalman Filter is integrated with the developed Koopman-based model for battery SoC estimation. The developed model continuously adapts via the streamed battery data and the trained linear model would be pushed to the onboard BMS for real-time battery state estimation. Results show the designed algorithm can adapt to the aging battery dynamics and consistently provide accurate SOC estimation without additional laboratory tests.

Keywords: Automotive Systems, Optimal Control, Nonlinear Control Systems
Abstract: One of the main reasons electric vehicles (EVs) struggle to completely replace fuel-powered cars is their limited driving range, which can be significantly reduced during extreme cold environment. Existing study on battery thermal management often treats EV batteries as a singular unit and ignores the thermal gradient among battery cells. To overcome this issue, a nonlinear model predictive control (NMPC)-based optimal strategy for cell-level thermal management of EV batteries is proposed. The considered thermal management system consists of a heat pump to warm up coolant fluid, three-way valves to divert coolant between different cell branches, and a flow-reversible pump to change coolant direction in real-time. A mixed-integer optimization problem is then formulated, where the integer variable corresponds to the flow direction and the continuous variables include heat pump compressor speed, flow speed, and coolant flow distributions among different cell branches. Compared to non-flow-reversible thermal management system, simulation results suggest that the proposed system can reduce the cell thermal gradient by 16.6% throughout the simulation and by 42.1% during steady state.

Keywords: Adaptive and Learning Systems, Automotive Systems, Optimal Control
Abstract: This paper presents a Model Reference Adaptive Control (MRAC) framework for lateral trajectory tracking in autonomous vehicles under uncertainty. It extends our prior work on MRAC for steer-by-wire systems by introducing a formulation that directly compensates for unknown dynamics and external disturbances without requiring system identification. A standard adaptive Model Predictive Control (MPC) is used solely as a benchmark for comparison. Both controllers are evaluated on a high-fidelity vehicle model under aggressive maneuvers such as Closed Circuit and Figure Eight, where lateral accelerations exceed 0.3g. Results show that MRAC achieves competitive tracking performance while reducing computational cost by more than 100 times, demonstrating its potential for real-time control in autonomous driving.

Keywords: Control Design, Optimal Control
Abstract: In this paper, an optimal data-driven direct input synthesis control (ODD-DISC) is proposed for output tracking of linear time-invariant (LTI) systems under input constraints consideration. Data-driven control (DDC) has recently gained popularity for its potential to address the limitations of model-based control in modeling, controller design and implementations. However, challenges still exist in current DDC methods to achieve precision output tracking in practice, due to the stringent assumption on the system’s prior knowledge, the lack of account of input energy and amplitude constraint, and the requirement of full system state measurement. Therefore, we propose a DISC approach using historical input-output data while considering input energy minimization and input amplitude. Specifically, the optimal control input to account for the input energy minimization is formulated as a data-driven linear quadratic regulator (LQR) problem and solved analytically, and the input amplitude constraint is addressed by converting the (amplitude) constrained tracking error minimization into an unconstrained problem, and solved via the numerical optimization algorithm. The proposed ODD-DISC method is illustrated through an output tracking example in the simulation.

Keywords: Control Design, Uncertain Systems and Robust Control, Optimal Control
Abstract: This paper presents a Model-Inspired Distributionally Robust Data-enabled Predictive Control (MDR-DeePC) framework for systems with partially known and uncertain dynamics. The proposed method integrates model-based equality constraints for known dynamics with a Hankel matrix-based representation of unknown dynamics. A distributionally robust optimization problem is formulated to account for parametric uncertainty and stochastic disturbances. Simulation results on a triple-mass-spring-damper system demonstrate improved disturbance rejection, reduced output oscillations, and lower control cost compared to standard DeePC. The results validate the robustness and effectiveness of MDR-DeePC, with potential for real-time implementation pending further benchmarking.

Keywords: Linear Control Systems, Optimal Control, Control Design
Abstract: Dual-mode model predictive control (DMPC) offers a scalable approach to constrained control by combining a finite-horizon MPC phase with a terminal stabilizing policy. While classical DMPC assumes symmetric constraint structures, real-world systems often exhibit directional asymmetry—especially in safety-critical applications like battery management. This paper analyzes how such asymmetric input and state constraints fundamentally alter the geometry of the maximal positively invariant (χMPI ) set, often leading to directional compression and loss of feasibility. Through scalar and two-dimensional examples, we highlight how controller tuning (via the LQR weight R) interacts with constraint asymmetry to distort invariant sets. A practical battery charging case study reveals that even under fixed physical bounds, aggressive control gains can shrink reachable sets and hinder terminal invariance. These findings motivate new care in DMPC design under realistic constraint regimes.

Keywords: Optimal Control
Abstract: Incorporating a notion of cost of sensing, or sensing-cost, within the optimal control framework is beneficial in controlling systems where the duration of sensing, and/or the cost of sensors themselves, have a considerable impact on the overall cost. In this regard, this paper presents multiple methods for incorporating an integral sensing-cost into the optimal control framework for Linear Time-Invariant (LTI) systems. Sensing-cost is traded off against the conventional costs of control and stabilization. In this paper, emphasis is placed on determining optimal sensing intervals, derived by applying the Pontryagin's Minimum Principle. Other formulations of the sensing-cost problem, and extension to nonlinear systems, are possible. The theoretical developments of this paper are validated through numerical solutions and demonstrated through simulations. Analytical results are derived for the sensing-cost problem applied to first-order systems. Additionally, a Shrinking Horizon method is demonstrated for practical implementation of the proposed theory and as a means to address uncertainties.

Keywords: Optimal Control, Linear Control Systems, Control Design
Abstract: This paper presents a MATLAB toolbox for implementing robust-to-early termination model predictive control, abbreviated as REAP, which is designed to ensure a sub-optimal yet feasible solution when MPC computations are prematurely terminated due to limited computational resources. Named REAP-T, this toolbox is a comprehensive, user-friendly, and modular platform that enables users to explore, analyze, and customize various components of REAP for their specific applications. Notable attributes of REAP-T are: (i) utilization of built-in MATLAB functions for defining the MPC problem; (ii) an interactive and intuitive graphical user interface for parameter tuning and visualization; (iii) real-time simulation capabilities, allowing users to observe and understand the real-time behavior of their systems; and (iv) inclusion of real-world examples designed to guide users through its effective use.

Keywords: Optimal Control, Machine Learning in modeling, estimation, and control, Intelligent Autonomous Vehicles
Abstract: In this paper, we present a new policy iteration approach to solve the linear-quadratic (LQ) optimal control problem for infinite horizon discrete-time dynamic systems under the assumption that the system input matrix is known and the state variables are accessible. The adaptive dynamic programming (ADP) methodology has been applied and widely used in many areas on reinforcement learning of corresponding continuous-time problems. However, the built-in constraint has limited its generalization to discrete-time problems as the prior knowledge of the system matrix is required. With our newly proposed method, the system state matrix can be recovered from the iterative reinforcement learning procedure so that the newly developed partial model-free policy iteration approach can also serve as a linear system identification technique. A realistic data-driven example, based on an online tracking and planning navigation mathematical model of a non-holonomic tracked vehicle under a leader-follower formation, is included to demonstrate the efficiency and robustness of the newly proposed methodology.