| |
Last updated on October 21, 2024. This conference program is tentative and subject to change
Technical Program for Tuesday October 29, 2024
|
TuAT1 |
Avenue Ballroom E |
Model Predictive Control and Optimization Techniques |
Regular Session |
Chair: Anubi, Olugbenga | Florida State University |
Co-Chair: Yi, Jingang | Rutgers University |
|
09:15-09:18, Paper TuAT1.1 | |
Autoencoder-Based Metamodeling for Structural Design Optimization |
|
Schneider, Fabian | Universtiy Siegen, Siegen |
Hellmig, Ralph Jörg | University of Siegen, Siegen |
Nelles, Oliver | University of Siegen |
Keywords: Machine Learning in modeling, estimation, and control, Modeling and Validation, Manufacturing Systems
Abstract: a data-driven metamodel approximates the computationally expensive simulation results of first principle models, e.g., finite element analyses. A significant drawback of typical metamodels is the limited amount of information that can be predicted due to their generally low-dimensional model output. Consequently, the metamodel usually does not predict the distribution of the desired quantity. This work presents a metamodel approach capable of predicting the spatial and temporal distribution of quantities for structural processes. This increases the modeling capability and makes more information available for the optimization. The autoencoder compresses the spatial distribution into a couple of features. The proposed methodology is applied to a three-stage forming process.
|
|
09:18-09:21, Paper TuAT1.2 | |
Optimization of Vibration-Based Reciprocating Compressor Valve Health Classification |
|
Chesnes, Jacob | Rochester Institute of Technology |
Nelson, Daniel | Novity, Inc |
Kolodziej, Jason | Rochester Institute of Technology |
Keywords: Machine Learning in modeling, estimation, and control, Modelling, Identification and Signal Processing, Power and Energy Systems
Abstract: The goal of this work is to improve vibration-based condition monitoring for reciprocating compressor valves by optimizing the parameters that convert the vibration data into the time-frequency domain. The continuous wavelet transform is used for this transformation, specifically the generalized Morse wavelet which has two parameters, gamma and P^2. The optimization starts with a grid search to examine the patterns and finishes with a rotating convergence pattern search to find the optimal parameters. For the outlet valve, low P^2 values performed well while high P^2 values performed better for the inlet valve.
|
|
09:21-09:24, Paper TuAT1.3 | |
Optimal Design of an Off-Grid PV Charger System with Second-Life Batteries Considering Performance Degradation |
|
Su, Zifei | Tennessee Technological University |
Dunlap, Caleb | Tennessee Technological University |
Chen, Pingen | Tennessee Technological University |
Keywords: Automotive Systems, Control of Smart Buildings and Microgrids, Optimal Control
Abstract: As the first-generation Battery Electric Vehicles (BEVs) reach the end of use life, the disposal of retired batteries raised significant economic and environmental concerns. To alleviate these problems, reusing retired BEV batteries on other applications such as off-grid photovoltaic (PV) systems with integrated energy storage system is a promising direction to give them a “second-life”. Although reuse Second-life Batteries (SLB) reduces cost of PV systems, significant challenges such as battery performance degradation caused by aging still need to be tackled. This paper utilizes Genetic Algorithm (GA) to minimize total cost of an off-grid PV system by optimizing the solar array size, SLB size, and starting state of health of SLBs simultaneously. The battery dynamics, aging, and performance degradation are modeled and integrated into the problem. Additionally, a real-world BEV charging dataset from 39 state parks in Tennessee is used to validate the performance of the algorithm in off-grid PV charger applications. The simulation results exhibit the algorithm can minimize total cost while ensuring battery performance. An economic analysis is performed comparing with the use of new batteries, and it is found that utilizing SLB for off-grid PV charger gives an average of 49.19% cost saving.
|
|
09:24-09:27, Paper TuAT1.4 | |
Sensor-Fusion-Based Optimal Multi-Disturbance Filtering in Atomic Force Microscope Imaging |
|
Chen, Jiarong | Rutgers, the state university of New Jersey |
Zou, Qingze | Rutgers, the State University of New Jersey |
|
|
09:27-09:30, Paper TuAT1.5 | |
Inverse Optimal Parametric QP: From the Construction of Constraints to the Selection of the Cost Function |
|
Yang, Songlin | CentraleSupele, Paris Saclay University |
Olaru, Sorin | CentraleSupelec |
Rodriguez-Ayerbe, Pedro | Supelec |
Hovd, Morten | Norwegian University of Technology and Science |
Keywords: Optimal Control
Abstract: The paper presents theoretical conditions and effective selection procedures for quadratic cost functions capable of completing the construction of an inverse optimal solution for a given piecewise affine (control) function, such as those arising from linear model predictive control (MPC) synthesis. Building on convex-concave liftings, previously proven capable of situating the explicit control function on the faces of a polyhedron in an extended optimization space, this work exploits the visibility of the constrained optimum from a generic unconstrained point to demonstrate that a quadratic cost can be used for inverse optimality, at least locally. The optimality conditions can then be aggregated to select a global quadratic cost function. These represent convex restrictions on the coefficients of the quadratic cost function, rendering the selection a computationally tractable problem. In cases of infeasibility, piecewise quadratic cost functions can be employed to address inverse optimality.
|
|
09:30-09:33, Paper TuAT1.6 | |
On NMPC-Based Rollover Avoidance Methods for Semi-Autonomous Forest Machines |
|
Badar, Tabish | Aalto University |
Knuutinen, Jere | Aalto University, ELEC School |
Backman, Juha | Natural Resources Institute Finland |
Visala, Arto | Aalto University, ELEC School |
Keywords: Unmanned Ground and Aerial Vehicles, Nonlinear Control Systems, Agricultural Systems
Abstract: The objective is to establish autonomy for the forest machine chain, which includes a harvester and a forwarder. In this study, we focus on building models and methodologies for autonomous forwarder operations. Rollover of autonomous forwarders owing to uneven terrain is a possible danger that must be identified and prevented during real-time forest harvesting. For a system model, a high-fidelity (augmented 6-DOF) vehicle model is proposed, which incorporates a 3D map of the terrain. A hybrid (reduced-order) vehicle model is proposed for the nonlinear model-predictive control (NMPC) method, which is employed for 3D path tracking while accounting for the local height variations to predict and prevent vehicle rollover. Furthermore, the aided-height odometry approach is employed to generate a reference 3D path that the autonomous forwarder can follow. A MATLAB-based high-fidelity simulation platform is developed, consisting of the identified 6-DOF vehicle model, actuator models, and 3D map of the terrain for evaluating the hybrid model-based NMPC method. The effectiveness of the two models is then shown using vehicle dynamics simulations to evaluate NMPC-based path tracking and roll predictions.
|
|
09:33-09:36, Paper TuAT1.7 | |
Safe Net-Recovery of Fixed-Wing Unmanned Aerial Vehicles Using Safety-Critical MPC |
|
Nejatbakhsh Esfahani, Hossein | Clemson University |
Mohammadpour Velni, Javad | Clemson University |
Keywords: Aerospace, Nonlinear Control Systems, Unmanned Ground and Aerial Vehicles
Abstract: To address the problem of safe net-recovery landing for small fixed-wing Unmanned Aerial Vehicles (UAVs), we propose to leverage the concept of Control Barrier Functions (CBFs) in a Model Predictive Control (MPC) scheme. In the proposed safety-critical control design, we address safety issues involved in the net-recovery, where a UAV aims at performing a safe landing with a minimum airspeed while safely navigating around unsafe zones even if the models used cannot perfectly capture the real-world environment. In this scenario, we formulate Robust CBFs (RCBFs) with adaptive safety margins to tackle the CBF model mismatch due to an imperfect wind model. Finally, to demonstrate the efficacy of the proposed safety-critical net-recovery, simulation studies are conducted for fixed-wing UAVs, and the results demonstrate the superiority of the proposed method compared to the state-of-the-art.
|
|
09:36-09:39, Paper TuAT1.8 | |
Closed-Loop Model Identification and MPC-Based Navigation of Quadcopters: A Case Study of Parrot Bebop 2 |
|
Amiri, Mohsen | Washington State University |
Hosseinzadeh, Mehdi | Washington State University |
Keywords: Modeling and Validation, Optimal Control, Unmanned Ground and Aerial Vehicles
Abstract: The growing potential of quadcopters in various domains, such as aerial photography, search and rescue, and infrastructure inspection, underscores the need for real-time control under strict safety and operational constraints. This challenge is compounded by the inherent nonlinear dynamics of quadcopters and the on-board computational limitations they face. This paper aims at addressing these challenges. First, this paper presents a comprehensive procedure for deriving a linear yet efficient model to describe the dynamics of quadrotors, thereby reducing complexity without compromising efficiency. Then, this paper develops a steady-state-aware Model Predictive Control (MPC) to effectively navigate quadcopters, while guaranteeing constraint satisfaction at all times. The main advantage of the steady-state-aware MPC is its low computational complexity, which makes it an appropriate choice for systems with limited computing capacity, like quadcopters. This paper considers Parrot Bebop 2 as the running example, and experimentally validates and evaluates the proposed algorithms.
|
|
09:39-09:42, Paper TuAT1.9 | |
Locomotion Control of Robot Walking on Moving Surface with Contingency Model Predictive Control (MPC) |
|
Chen, Kuo | Motional |
Huang, Xinyan | Rugters Universtity |
Chen, Xunjie | Rutgers University |
Yi, Jingang | Rutgers University |
Keywords: Uncertain Systems and Robust Control, Control Applications, Robotics
Abstract: It is challenging for bipedal robots to achieve balancing and walking on horizontally shifting surfaces such as on moving trains. This paper proposes an approach handling the uncertainty of the surface movement with contingency model predictive control (CMPC) by predicting not only robot states but also future surface movement. The contingency MPC predicts robot states based on two extreme cases of future surface movement. The input trajectories corresponding to the two extreme cases share the same initial portion without extra effort to select between them. Swing foot touch down location strategy is proposed based on the contingency framework. Zero moment point (ZMP) trajectories are optimized to satisfy the foot placement. A whole body locomotion controller then stablize the robot on the desired ZMP trajectories and swing foot trajectories. Simulation results validate the contingency MPC framework of bipedal robot locomotion on horizontally moving surfaces.
|
|
09:42-09:45, Paper TuAT1.10 | |
Model Predictive Control for Battery Electric Vehicles Considering Energy Efficiency, Battery Degradation and Tire Wear |
|
Su, Zifei | Tennessee Technological University |
Abdullah Eissa, Magdy | Tennessee Technological University |
Qari, Marwan | Harbinger Motors Inc |
Chen, Pingen | Tennessee Technological University |
Keywords: Automotive Systems, Optimal Control, Machine Learning in modeling, estimation, and control
Abstract: Eco-driving of Battery Electric Vehicles (BEVs) has been extensively studied in the past decade because of its potential of enhancing the energy efficiency of individual BEVs without significantly increasing the hardware investment. In this study, we propose a model predictive control (MPC)-based eco-driving control scheme which simultaneously considers the vehicle efficiency, battery degradation, and tire wear of BEVs in the optimization of speed profile. An electro-chemical battery degradation model is deployed to account for different aging factors, and a regression model is utilized to quantify the tire wear based on non-exhaust particle matter emissions. Furthermore, a deep neural network-based velocity prediction model is trained and integrated to the control framework to accommodate the requirements of forecasting future speed due to the nature of MPC. Comparative studies have been performed in a real-world driving cycle. Optimization results show that tire particulate matter (PM) emissions, battery degradation, and energy consumption can be reduced by 44.15%, 2.88%, and 0.73%, respectively, when compared to the baseline controller.
|
|
09:45-09:48, Paper TuAT1.11 | |
Nonlinear Model Predictive Control for Directional Drilling Applications |
|
Xu, Jiamin | The University of Texas at Austin |
Demirer, Nazli | Halliburton |
Pho, Vy | Halliburton |
Zhang, He | Halliburton |
Tian, Kaixiao | Halliburton |
Bhaidasna, Ketan | Halliburton |
Darbe, Robert | Halliburton |
Chen, Dongmei | UT Austin |
Keywords: Nonlinear Control Systems, Optimal Control, Control Applications
Abstract: In directional drilling, the nonlinear Delay Differential Equation (DDE) model is recognized for its high precision in predicting borehole trajectory but has rarely been integrated into Model Predictive Control (MPC) frameworks due to its inherent complexity. To address this challenge, this paper proposes a novel method to transform the nonlinear DDE model into discretized Ordinary Differential Equation (ODE) model. Following this transformation, a novel optimization framework is proposed to concurrently determine optimal control inputs and solve the linear complementarity problem (LCP). The validity of both the discretized model and the optimization strategy is verified through comparison with results from existing literature. Subsequent closed-loop simulations demonstrate the ability of the proposed MPC to maintain the drill string's alignment with the planned well trajectory, even in the presence of disturbances and noise. These results highlight the effectiveness and potential applicability of the proposed MPC strategy in real-world directional drilling operations.
|
|
09:48-09:51, Paper TuAT1.12 | |
Fast Model Predictive Control of Input-Affine Systems: Application to the Hindmarsh-Rose Neuron Model |
|
Grancharova, Alexandra | University of Chemical Technology and Metallurgy |
Xie, Junhong | University of Chemical Technology and Metallurgy |
Olaru, Sorin | CentraleSupelec |
Keywords: Optimal Control, Control Design, Control Applications
Abstract: The paper presents a low complexity nonlinear MPC design for the class of constrained input-affine systems. Essentially, it builds on the idea of adding a contractive constraint in the NMPC problem formulation, which would ensure the closed-loop system stability when using a small prediction horizon. In particular, the one-step ahead NMPC problem with contractive constraint is considered and an approach to obtain an efficient online solution of the associated convex quadratically constrained quadratic programming problem is developed. The proposed technique is shown to be effective for embedded convex NMPC of input-affine systems, since it will reduce the computational complexity of the online NMPC and simplify the software and hardware implementation. The methodological developments are illustrated with simulations on the Hindmarsh-Rose neuron model.
|
|
09:51-09:54, Paper TuAT1.13 | |
Distributed Model-Predictive Energy Management Strategy for Shipboard Power Systems Considering Battery Degradation |
|
Vedula, Satish | Florida State University |
Alaviani, Seyyed Shaho | Florida State University |
Anubi, Olugbenga | Florida State University |
Keywords: Power and Energy Systems, Control Applications, Control Design
Abstract: With the integration of loads such as pulse power loads, a new control challenge is presented in meeting their high ramp rate requirements. Existing onboard generators are ramp rate limited. The inability to meet the load power due to ramp rate limitation may lead to instability. The addition of energy storage elements in addition to the existing generators proves a viable solution in addressing the control challenges presented by high ramp rate loads. A distributed energy management strategy maximizing generator efficiency and minimizing energy storage degradation is developed that facilitates an optimal adaptive power split between generators and energy storage elements. The complex structure of the energy storage degradation model makes it tough for its direct integration into the optimization problem and is not practical for real-time implementation. A degradation heuristic to minimize absolute power extracted from the energy storage elements is proposed as a degradation heuristic measure. The designed strategy is tested through a numerical case study of a consolidated shipboard power system model consisting of a single generator, energy storage element, and load model. The results show the impact of the designed energy management strategy in effectively managing energy storage health.
|
|
09:54-09:57, Paper TuAT1.14 | |
Nonlinear Model Predictive Control for Mitigating Epidemic Spread Using a Partial Differential Equation Based Compartmental Dynamic Model |
|
Street, Logan | University of Cincinnati |
David, Deepak Antony | University of Cincinanti |
Liu, Chunyan | Cincinnati Children's Hospital Medical Center |
Ehrlich, Shelley | Cincinnati Children’s Hospital Medical Center |
Kumar, Manish | University of Cincinnati |
Ramakrishnan, Subramanian | University of Dayton |
Keywords: Control Applications, Modeling and Validation, Nonlinear Control Systems
Abstract: We present a Nonlinear Model Predictive Control (NMPC) framework for epidemic spread mitigation using a Partial Differential Equation (PDE) based Susceptible-Latent- Infected-Recovered (SLIR) epidemiological dynamic model. The spatio-temporal epidemic spread predictions of the model were numerically validated in our previous work using empirical COVID-19 data for Hamilton County, Ohio, employing a single-objective Genetic Algorithm (GA) for training model parameters. The validated model serves as the basis for the NMPC prediction and control framework developed to support the design of optimal Non-Pharmaceutical Interventions for spread mitigation. We consider a cost function comprising the infection spread density and the cost of applied control, with the latter representing socioeconomic effects. With a prediction horizon (Tp) of 30 days and a control horizon (Tu) of 15 days. The NMPC investigates a uniformly distributed control scheme across the entire spatial domain for three different time periods of the COVID-19 pandemic with distinct infection trends. In summary, the article presents one of the first efforts towards developing an NMPC framework based on a spatio-temporal epidemic dynamic model. The results provide an analytical basis for improved spread mitigation of future epidemics.
|
|
09:57-10:00, Paper TuAT1.15 | |
Enhancing Traffic Flow Via Feedback Linearization and Model Predictive Control under Input Constraints |
|
Rahmanidehkordi, Arash | University of North Carolina at Charlotte |
Ghasemi, Amirhossein | University of North Carolina Charlotte |
Keywords: Control Applications, Control Design, Nonlinear Control Systems
Abstract: This paper introduces a novel algorithm that combines feedback linearization (FL) with model predictive control (MPC) for managing highway traffic as an over-actuated, constrained nonlinear system. FL converts the non-linear traffic flow dynamics of the METANET model into a linear form, but it does not inherently handle control command constraints. To address this, an MPC will be integrated that takes the linearized output from the FL controller and produces the virtual control commands for the FL controller. Followed by that, a novel constraint mapping algorithm will be presented to determine these virtual control commands, ensuring all input constraints are met. The algorithm also selects the most cost-effective command for optimal reference tracking. Simulations validate the approach, showing significant improvements in traffic flow and reductions in average travel times.
|
|
10:00-10:03, Paper TuAT1.16 | |
Adversarially Robust Reduced Order Model Predictive Control: Balancing Performance and Efficiency |
|
Ahmed, Akhil | Imperial College London |
del Rio-Chanona, Ehecatl Antonio | Imperial College London |
Mercangöz, Mehmet | Imperial College London |
Keywords: Chemical Process Control, Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems
Abstract: Model Predictive Control (MPC) is a powerful technique for controlling process systems, but it faces a notable trade-off between control performance and computational efficiency. At one extreme, one may choose to employ accurate nonlinear models and solve a nonlinear MPC problem to gain in performance at the expense of efficiency. At the other extreme, one may opt for simple linear models to solve a linear MPC problem sacrificing performance for efficiency gains. This trade-off is exacerbated for large-scale systems, which are ubiquitous in plant-wide process control. In response to this challenge, Reduced Order Models (ROMs) emerge as a promising solution, as they preserve model fidelity (and thus control performance) but with reduced dimensionality allowing for a smaller and thus more efficient optimal control problem to be solved. However, state-of-the-art data-driven ROMs, such as autoencoders, often exhibit sensitivity to disturbances and noise when applied in control settings like MPC. This diminishes control performance, limiting their practical use. This work explores the integration of adversarial machine learning techniques to fortify ROMs against such perturbations, aiming to achieve enhanced robustness without compromising performance or efficiency.
|
|
10:03-10:06, Paper TuAT1.17 | |
Decentralized Droop-Based Finite-Control-Set Model Predictive Control of Inverter-Based Resources in Islanded AC Microgrid |
|
Olajube, Ayobami | Florida State University |
Omiloli, Koto Andrew | Florida State University |
Vedula, Satish | Florida State University |
Anubi, Olugbenga | Florida State University |
Keywords: Control of Smart Buildings and Microgrids, Control Design, Linear Control Systems
Abstract: This paper presents an improved droop control method to ensure effective power sharing, voltage regulation, and frequency stabilization of inverter-based resources (IBRs) connected in parallel in an islanded AC microgrid. In the contemporary droop control algorithm, the distance between connected inverters affects the effectiveness of the active power-frequency and the reactive power-voltage droop characteristics which results in poor power sharing at the primary level of the microgrid. That is, high impedance emanating from long transmission lines results in instability, poor voltage tracking, and ineffective frequency regulation. Hence, the inner current control loop of the inverters is replaced by the finite-control-set model predictive controller (FCS-MPC) which gives efficient voltage tracking, good frequency regulation, and faster performance response. FCS-MPC is easy to implement in fast switching converters and does not suffer from computational burden unlike the continuous-set MPC and is also devoid of issues of multiple-loop, parameter variation, and slow response associated with conventional droop control methods. We derived the condition for bounded stability for FCS-MPC and the proposed method is tested via a numerical simulation on three IBRs. The results show effective power sharing, capacitor voltage tracking, and frequency regulation with reduced oscillations to changes in load.
|
|
10:06-10:09, Paper TuAT1.18 | |
A New Parameter Tuning Method of Flocking Control Based on a Dual-Loop Model Predictive Control Structure |
|
Wang, Gang | Arizona State University |
Chen, Yan | Arizona State University |
Keywords: Control Design, Automotive Systems, Intelligent Autonomous Vehicles
Abstract: Although flocking control of multi-agent mobile systems is extensively studied in the literature, the control performance is often limited or varied by the tuning of control parameters. Inappropriate parameter selection can greatly impact the convergence speed of flocking and may even result in flocking failure. Because of the conflicting features of flocking control parameters, such as those for attractive and repulsive forces, and the increased control parameters with the number of agents, simultaneous control parameter tuning toward an optimized flocking behavior of a large group of multi-agent systems faces real challenges. A dual-loop model predictive control (MPC) structure is proposed to address this challenge in a systematic manner. Unlike common MPC that directly gives control design, the proposed approach optimizes the flocking control parameters within the MPC framework and then feeds these optimized parameters into the flocking controller to generate control inputs. This novel method enhances the performance of flocking controllers through a dual-loop MPC structure. Simulation results demonstrate that the proposed approach maintains the inherent properties of the designed flocking controllers while leveraging the adaptability and optimality of MPC for parameter tuning.
|
|
10:09-10:12, Paper TuAT1.19 | |
Battery Aging-Aware Optimal Charging Threshold for EV Range Estimation Based on Depth of Discharge (DoD) Modeling |
|
Abdullah Eissa, Magdy | Tennessee Technological University |
Chen, Pingen | Tennessee Technological University |
Keywords: Automotive Systems, Estimation, Modeling and Validation
Abstract: The depth of discharge (DOD) significantly influences electric vehicle (EV) battery lifespan, as higher DoD levels accelerate chemical reactions, imposing greater stress on battery electrodes and electrolytes. Maintaining the DOD within a specific range is crucial for enhancing energy efficiency and maximizing battery longevity. Consequently, optimizing charging and discharging schedules to account for aging effects becomes imperative in managing DOD. This paper proposes a methodology for determining the adaptive threshold of the optimal charging area to estimate the driving range while considering aging effects and expanding the driving range within the recommended DOD boundaries. The performance of this methodology is validated using real-world driving data, demonstrating its capability to enhancing battery lifespan, while also promoting correct charging behaviors. These findings underscore the challenges associated with addressing range anxiety and mitigating battery degradation across diverse user profiles. Implementing the optimized charging strategies proposed in this study enables stakeholders and EV users to improve driving range predictability and optimize battery performance.
|
|
10:12-10:15, Paper TuAT1.20 | |
Conflict-Aware Data-Driven Safe Linear Quadratic Control |
|
Esmzad, Ramin | Michigan State University |
Niknejad, Nariman | Michigan State University |
Modares, Hamidreza | Michigan State University |
Keywords: Linear Control Systems, Optimal Control, Uncertain Systems and Robust Control
Abstract: This paper presents a new data-driven conflict-aware safe Linear Quadratic Regulator (LQR) with dual safety measures. During design, the LQR control gain is optimized solely from data to minimize costs and enlarge a conflict-free zone ensuring safe optimal trajectories. In execution, a control barrier certificate (CBC) verifies the safety of controller actions. The design-time intervention implicitly aligns LQR weights with safety constraints, preventing harmful conflicts and reducing the need for frequent CBC interventions. To achieve this, the LQR gain is parameterized with a lambda-contractive safe set. Simulation results on the vehicle steering model demonstrate the effectiveness of this approach.
|
|
TuAT3 |
Prime 1 |
Estimation and Control for Energy Storage Systems |
Invited Session |
Chair: Lucero, Joseph N. E. | Stanford University |
Co-Chair: Shahbakhti, Mahdi | University of Alberta |
|
09:15-09:30, Paper TuAT3.1 | |
An Investigation into the Viability of Cell-Level Temperature Control in Lithium-Ion Battery Packs (I) |
|
Abadie, Preston | Texas Tech University |
Docimo, Donald | Texas Tech University |
Keywords: Power and Energy Systems, Linear Control Systems, Control Applications
Abstract: This paper focuses on the thermal management and temperature balancing of lithium-ion battery packs. As society transitions to relying more heavily on renewable energy, the need for energy storage rises considerably, as storage facilitates power regulation between these sources and the grid. Lithium-ion batteries are leading the market for energy storage options, but their properties are temperature sensitive, with thermal abuse resulting in shortened pack lifetime and possible safety issues. Current battery thermal management systems (BTMS) are implemented in a number of ways to ensure consistent and reliable operation. However, they are typically limited in architecture and restricted in ability to attend to temperature gradients. This work proposes a BTMS topology which permits control of the individual cooling received by a cell in a pack. First, an analysis is done using timescale separation to confirm that cell-level temperature control is capable of extending the lifetime of a pack as compared to pack-level control. The analysis is used to guide gain tuning of a state feedback controller, which directs more cooling effort to cells of higher temperatures. Validation of the BTMS topology and control is performed through simulation of a battery pack, with variations in total cooling power and resistance heterogeneity. The outcome of the validation studies indicates that the proposed BTMS configuration is better equipped to reduce temperature differences and extend pack life. This benefit increases as total input power increases, giving the controller more freedom to cool unhealthy cells while remaining within power constraints.
|
|
09:30-09:45, Paper TuAT3.2 | |
Transformer-Based Capacity Prediction for Lithium-Ion Batteries with Data Augmentation (I) |
|
Modekwe, Gift | Texas Tech University |
Al-Wahaibi, Saif | Texas Tech University |
Lu, Qiugang | Texas Tech University |
Keywords: Power and Energy Systems, Machine Learning in modeling, estimation, and control
Abstract: Lithium-ion batteries are pivotal to technological advancements in transportation, electronics, and clean energy storage. The optimal operation and safety of these batteries require proper and reliable estimation of battery capacities to monitor the state of health. Current methods for estimating the capacities fail to adequately account for long-term temporal dependencies of key variables (e.g., voltage, current, and temperature) associated with battery aging and degradation. In this study, we explore the usage of transformer networks to enhance the estimation of battery capacity. We develop a transformer-based battery capacity prediction model that accounts for both long-term and short-term patterns in battery data. Further, to tackle the data scarcity issue, data augmentation is used to increase the data size, which helps to improve the performance of the model. Our proposed method is validated with benchmark datasets. Simulation results show the effectiveness of data augmentation and the transformer network in improving the accuracy and robustness of battery capacity prediction.
|
|
09:45-10:00, Paper TuAT3.3 | |
Optimal Fast Charging of Lithium-Ion Batteries through Continual Hybrid Model Learning (I) |
|
Hailemichael, Habtamu | Clemson University |
Ayalew, Beshah | Cemson University |
Keywords: Power and Energy Systems, Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems
Abstract: Towards the goal of addressing the critical challenge of extended charging times for lithium-ion batteries (LiBs), this study introduces a novel learning-based fast charging control framework that optimizes charging schedules throughout the LiB's lifespan. This is achieved by first continually learning a virtual hybrid model, which is then utilized to generate data via latent imagination for fast charging policy training with deep reinforcement learning (DRL). Unlike traditional heuristic methods, which are often conservative, or purely physics model-based approaches that struggle to capture the complex dynamics of LiB operation and degradation, our hybrid model continuously adapts with operational data, enabling the generation of customized fast charging policies as the battery degrades. Through high-fidelity simulations and comparisons with standard CCCV charging protocols, we find that the proposed framework achieves a significant charging speed improvement at different ambient temperatures and cooling efforts while ensuring battery health.
|
|
10:00-10:15, Paper TuAT3.4 | |
Diagnosing and Decoupling the Degradation Mechanisms in Lithium Ion Cells: An Estimation Approach (I) |
|
Appana, Raja Abhishek | The Ohio State University |
El Idrissi, Faissal | The Ohio State University |
Ramesh, Prashanth | The Ohio State University |
Canova, Marcello | The Ohio State University |
Kang, Chun Yong | Hyundai Motor Company |
Um, Kimoon | Hyundai Motor Company |
Keywords: Power and Energy Systems, Estimation
Abstract: Understanding battery degradation in electric vehicles (EVs) under real-world conditions remains a critical yet under-explored area of research. Central to this investigation is the challenge of estimating the specific degradation modes in aged cells with no available information on usage history, bypassing the invasive yet conventional method of tear-down tests. Using an electrochemical model, this study pioneers a methodology to decouple and isolate the aging mechanisms in batteries sourced from EVs with varying mileages. A robust correlation is established between the model parameters and distinct degradation processes, enabling the diagnosis and estimation of each mechanism’s impact on the battery’s parameters. This paper sheds light on battery degradation in real-world scenarios and demonstrates the feasibility of their identification, isolation, and approximate quantification of their effects.
|
|
10:15-10:30, Paper TuAT3.5 | |
Comparing Mass-Preserving Numerical Methods for the Lithium-Ion Battery Single-Particle Model |
|
Lucero, Joseph N. E. | Stanford University |
Xu, Le | Stanford University |
Onori, Simona | Stanford University |
Keywords: Modeling and Validation, Automotive Systems
Abstract: The single particle model (SPM) is a reduced electrochemical model that holds promise for applications in battery management systems due to its ability to accurately capture battery dynamics; however, the numerical discretization of the SPM requires careful consideration to ensure numerical stability and accuracy. In this paper, we present a comparative study of two mass-preserving numerical schemes for the SPM: the finite volume method and the control volume method. Using numerical simulations, we systematically evaluate the performance of these schemes, after independently calibrating the SPM discretized with each scheme to experimental data, and find a tradeoff between accuracy (quantified by voltage root- mean-square error) and computational time. Our findings provide insights into the selection of numerical schemes for the SPM, contributing to the advancement of battery modeling and simulation techniques.
|
|
10:30-10:45, Paper TuAT3.6 | |
Control-Oriented Modeling of a Solid Oxide Fuel Cell Affected by Redox Cycling Using a Novel Deep Learning Approach |
|
Mohamadali, Tofigh | University of Alberta |
Fakouri Hasanabadi, Masood | University of Alberta |
Smith, Daniel J. | Cummins |
Ali, Kharazmi | Cummins Inc |
Amir Reza, Hanifi | University of Alberta |
Koch, Charles Robert | University of Alberta |
Shahbakhti, Mahdi | University of Alberta |
Keywords: Machine Learning in modeling, estimation, and control, Power and Energy Systems, Modelling, Identification and Signal Processing
Abstract: A Solid Oxide Fuel Cell (SOFC) is a multi-physics system that involves heat transfer, mass transport, and electrochemical reactions to produce electrical power. Reduction and re-oxidation (Redox) cycling is a destructive reaction that can occur during SOFC operation. Redox induces various degradation mechanisms, such as electrode delamination, nickel agglomeration, and microstructural changes, which should be mitigated. The interplay of these mechanisms makes a post-Redox SOFC a nonlinear, time-varying, nonstationary dynamic system. Physics-based modeling of these complexities often leads to computationally expensive equations that are not suitable for the control and diagnostics of SOFCs. Here, a data-driven approach based on dilated convolutions and a self-attention mechanism is introduced to effectively capture the dynamics underlying SOFCs affected by Redox. Controlled Redox cycles are designed to collect appropriate experimental data for developing deep learning models, which are lacking in the current literature. The performance of the proposed model is validated on diverse unseen data sets gathered from different fuel cells and benchmarked against state-of-the-art models, in terms of prediction accuracy and computation complexity. The results indicate 31% accuracy improvement and 27% computation speed enhancement compared to the benchmarks.
|
|
TuAT4 |
Prime 2 |
Recent Advances in Vehicle Motion Planning and Controls |
Invited Session |
Chair: Orosz, Gabor | University of Michigan |
Co-Chair: Karbowski, Dominik | Argonne National Laboratory |
|
09:15-09:30, Paper TuAT4.1 | |
Trajectory Shaper: A Solution for Disrupted Cooperative Adaptive Cruise Control (I) |
|
Zhou, Anye | Oak Ridge National Laboratory |
Wang, Zejiang | The University of Texas at Dallas |
Cook, Adian | Oak Ridge National Laboratory |
Keywords: Intelligent Autonomous Vehicles, Transportation Systems, Control Applications
Abstract: Cooperative adaptive cruise control (CACC) can effectively reduce energy consumption, alleviate traffic congestion, and enhance safety. However, communication-related constraints and uncooperative vehicle users can disrupt CACC during real-world operations, significantly undermining the putative benefits of CACC. To alleviate the negative impacts of disrupted CACC, this study develops the trajectory shaper (TS) methods as backup solutions for two scenarios: (i) communication between vehicles is infeasible, and vehicles execute adaptive cruise control (ACC) using local sensor measurements; (ii) follower vehicles reject forming a cooperative platoon and execute their local distributed controllers using the information attained via communication. When communication is infeasible, a distributed TS is devised on each vehicle to modify the sensor measurements, enabling safe and efficient ACC operations. When communication is available but uncooperative agents are involved, the lead vehicle of the platoon executes a centralized TS to modify the information shared with uncooperative agents, achieving optimal platoon-level performance. The centralized and distributed TSs are implemented based on the model predictive control algorithms to yield optimal modifications on input information. Robustness is also factored to tackle model uncertainties during TS operations to ensure safety and efficiency. Numerical experiments validate the control performance of the proposed TSs.
|
|
09:30-09:45, Paper TuAT4.2 | |
Enhancing Eco-Driving Control in Connected and Automated Vehicles through Reinforcement Learning Optimization (I) |
|
Zhang, Yaozhong | Argonne National Laboratory |
Ammourah, Rami | Argonne National Laboratory |
Han, Jihun | Argonne National Laboratory |
Moawad, Ayman | Argonne National Laboratory |
Shen, Daliang | Argonne National Laboratory |
Karbowski, Dominik | Argonne National Laboratory |
Rousseau, Aymeric | Argonne National Laboratory |
Keywords: Machine Learning in modeling, estimation, and control, Intelligent Autonomous Vehicles, Optimal Control
Abstract: This paper introduces an innovative approach to augmenting the energy efficiency of eco-driving controllers in Connected and Automated Vehicles (CAVs) by integrating Reinforcement Learning (RL) algorithms to optimize key control parameters. Leveraging continuous action space RL algorithms, specifically Deep Deterministic Policy Gradient (DDPG) and Proximal Policy Optimization (PPO), we focus on the real-time optimization of the control parameters that guide the vehicle’s speed in various driving modes. These parameters are integral to the eco-driving controller, which is based on a model predictive control framework and the optimal control algorithm known as Pontryagin’s maximum principle (PMP). In our study, Argonne National Laboratory’s RoadRunner, a Simulink-based simulation tool is used to conduct experiments in a controlled setting featuring realistic traffic elements for CAV controls. By training RL agents to adjust these parameters in response to environments, we achieve significant energy savings without compromising travel time. Our results from five distinct scenarios reveal that the RL-enhanced eco-driving controller outperforms both non-RL eco-driving control (by 12%) and the human driver model (by 24.2%) in terms of energy savings. This work demonstrates the potential of RL with continuous action spaces to further improve eco-driving efficiency as one example of a novel application of RL algorithms aimed at maximizing the performance of the existing optimization-based vehicle controls.
|
|
09:45-10:00, Paper TuAT4.3 | |
Maneuvering with an Autonomous Unicycle (I) |
|
Vizi, Mate Benjamin | Budapest University of Technology and Economics |
Orosz, Gabor | University of Michigan |
Takacs, Denes | MTA-BME Research Group on Dynamics of Machines and Vehicles |
Stepan, Gabor | Budapest Univ of Technology and Economics |
Keywords: Path Planning and Motion Control, Control Design, Robotics
Abstract: The path-following task of an autonomous unicycle is considered in the three-dimensional space. The equation of motion is derived using the Appellian aproach of nonhonolomic dynamics. The resulting nonlinear model is transformed into the path-reference frame. Using pole placement, a linear feedback controller is designed that takes into account the velocity of the maneuver. The resulting controller is tested on the nonlinear model via numerical simulations; lane change maneuvers are carried out successfully at different speeds.
|
|
10:00-10:15, Paper TuAT4.4 | |
Consideration of Vehicle Characteristics on the Motion Planner Algorithm (I) |
|
Ahmed, Syed Adil | University of Michigan Dearborn |
Shim, Taehyun | Univ of Michigan-Dearborn |
Keywords: Path Planning and Motion Control, Automotive Systems, Nonlinear Control Systems
Abstract: Autonomous vehicle control is generally divided in two main areas; trajectory planning and tracking. Currently, the trajectory planning is mostly done by particle or kinematic model based optimization controllers. The output of these planners, since they do not consider CG height and its effects, is not unique for different vehicle types, especially for high CG vehicles. As a result, the tracking controller may have to work hard to avoid vehicle handling and comfort constraints while trying to realize these sub-optimal trajectories. This paper tries to address this problem by considering a planner with simplified double track model with estimation of lateral and roll based load transfer using steady state equations and a simplified tire model to reduce solver workload. The developed planner is compared with the widely used particle and kinematic model planners in collision avoidance scenarios in both high and low acceleration conditions and with different vehicle heights
|
|
10:15-10:30, Paper TuAT4.5 | |
Energy-Optimal Trajectory Planning for Semi-Autonomous Hydraulic Excavators |
|
Cupo, Alessandro | Politecnico Di Milano |
Cecchin, Leonardo | Robert Bosch GmbH |
Demir, Ozan | Ruhr-Universitaet Bochum |
Fagiano, Lorenzo | Politecnico Di Milano |
Keywords: Path Planning and Motion Control, Optimal Control, Mechatronic Systems
Abstract: An optimal trajectory planning approach for hydraulic excavator arms is presented, where the goal is to create trajectories that trade-off energy consumption and completion time. We develop a physics-based model of the excavator, which describes both the dynamics and the hydraulic system's behavior. Further investigation of the Optimal Control Problem, used to create the trajectory, allows for discussion regarding the trade-off between power and time recovering a wide range of solutions based on the designer's choice. Lastly, the problem is extended to include obstacle-avoidance constraints, creating a collision-free and efficient path.
|
|
10:30-10:45, Paper TuAT4.6 | |
Aerial and Ground Vehicles Collaboration for Automated Target Tracking Using Reinforcement Learning |
|
Zanone, R. Oliver | Clemson University |
Mohammadpour Velni, Javad | Clemson University |
Keywords: Unmanned Ground and Aerial Vehicles, Machine Learning in modeling, estimation, and control
Abstract: This paper presents a real-world application and implementation of collaborative target tracking between a drone and a ground vehicle, leveraging deep reinforcement learning (DRL) and a fiducial marker vision system. The work utilizes custom-designed modular platforms, combining the drone's aerial advantages with the ground vehicle's differential control capabilities. The drone here serves as the centralized controller, while the ground vehicle, with a cost-efficient custom design, features wheel encoders and orientation sensors, controlled by a single board computer. A comprehensive investigation of DRL using Proximal Policy Optimization (PPO) in a Unity-based simulated environment is used to refine the RL model. Field experiments demonstrate effective target tracking, with results showcasing improved stability and accuracy, validating the applicability of this collaborative system in practical scenarios such as surveillance and reconnaissance.
|
|
TuAT5 |
Streeterville E |
Healthcare and Medical Systems |
Invited Session |
Chair: Radisavljevic-Gajic, Verica | Ajman Univeristy |
Co-Chair: Ahuja, Nitisha | The Pennsylvania State University |
|
09:15-09:30, Paper TuAT5.1 | |
Soft Inflatable Knee Exosuit for Flexion Assistance in Swing Phase (I) |
|
Rokalaboina, Jahnav | Arizona State University |
Nibi, Tolemy | Arizona State University |
Tao, Weijia | Arizona State University |
Zhang, Wenlong | Arizona State University |
Keywords: Soft Robotics, Assistive and Rehabilitation Robotics, Sensors and Actuators
Abstract: Individuals with impaired mobility often struggle to achieve smooth and efficient gait patterns due to limitations in knee flexion to create enough foot clearance. Wearable robots can assist individuals with impaired mobility by providing support and facilitating natural movement patterns during gait training for rehabilitation. This study addresses the need for effective assistance in knee joint flexion for individuals with impaired mobility by introducing and evaluating a soft robotic exosuit powered by a new Inflatable Flexion Actuator (IFA). The IFAs can generate torque and assist in flexion motion during the swing phase of the gait cycle. The design, characterization, and control of the IFAs are discussed. The soft exosuit is integrated with inertial measurement units (IMUs) and an instrumented treadmill to detect swing phases and control the actuators. To evaluate the effect of exosuit during flexion, surface electromyography (sEMG) sensors are placed to record the muscle activity of vastus lateralis, rectus femoris, and biceps femoris. A maximum average reduction of 41% in muscle activity in Rectus Femoris is observed, with an average reduction of 20.4%, 19.6%, and 25% in muscle activity in vastus lateralis, rectus femoris, and biceps femoris respectively, which indicates the potential of applying the soft inflatable exosuit to individuals with impaired mobility.
|
|
09:30-09:45, Paper TuAT5.2 | |
Robust Control of Exo-Abs, a Wearable Platform for Ubiquitous Respiratory Assistance |
|
Rezaei, Parham | University of Maryland, College Park |
Lee, Sang-Yoep | Massachusetts Institute of Technology |
Cho, KyuJin | Seoul National University |
Hahn, Jin-Oh | University of Maryland |
Keywords: Healthcare systems, Modelling and Control of Biomedical Systems, Control Applications
Abstract: Existing non-invasive breathing assist options compatible with out-of-hospital settings are limited and not appropriate to enable essential everyday activities, thereby deteriorating the quality of life. In our prior work, we developed the Exo-Abs, a novel wearable robotic platform for ubiquitous assistance of respiratory functions in patients with respiratory deficiency. This paper concerns the development of a model-based closed-loop control algorithm for the Exo-Abs to automate its breathing assistance. To facilitate model-based development of closed-loop control algorithms, we developed a control-oriented mathematical model of the Exo-Abs. Then, we developed a robust absolutely stabilizing gain-scheduled proportional-integral control algorithm for automating the breathing assistance with the Exo-Abs, by (i) solving a linear matrix inequality formulation of the Lyapunov stability condition against sector-bounded uncertainty and inter-individual variability in the mechanics of the abdomen and the lungs and (ii) augmenting it with a heuristic yet effective gain scheduling algorithm. Using in silico evaluation based on realistic and plausible virtual patients, we demonstrated the efficacy and robustness of the automated breathing assistance of the Exo-Abs under a wide range of variability in spontaneous breathing and Exo-Abs efficiency: the absolutely stabilizing gain-scheduled proportional-integral control resulted in small exhalation trajectory tracking error (<30ml) with smooth actuation, which was superior to (i) its proportional-integral control counterpart in tracking efficacy and to (ii) its proportional-integral-derivative control counterpart in chattering.
|
|
09:45-10:00, Paper TuAT5.3 | |
Development of a Virtual Patient Generator for Simulation of Vasopressor Resuscitation |
|
Kao, Yi-Ming | University of Maryland, College Park |
Chalumuri, Yekanth Ram | University of Maryland |
Sampson, Catherine | University of Texas Medical Branch |
Shah, Syed | University of Texas Medical Branch |
Salsbury, John | University of Texas Medical Branch |
Tivay, Ali | University of Maryland |
Kinsky, Michael | University of Texas Medical Branch |
Kramer, George | University of Texas Medical Branch |
Hahn, Jin-Oh | University of Maryland |
Keywords: Modeling and Control of Biotechnological Systems, Modelling and Control of Biomedical Systems, Modelling, Identification and Signal Processing
Abstract: This paper presents a virtual patient generator (VPG) intended to be used for pre-clinical in silico evaluation of autonomous vasopressor administration algorithms in the setting of experimentally induced vasoplegia. Our VPG consists of two main components: (i) a mathematical model that replicates physiological responses to experimental vasoplegia (induced by sodium nitroprusside (SNP)) and vasopressor resuscitation via phenylephrine (PHP) and (ii) a parameter vector sample generator in the form of a multi-dimensional probability density function (PDF) using which the parameters characterizing the mathematical model can be sampled. We developed and validated a mathematical model capable of predicting physiological responses to the administration of SNP and PHP. Then, we developed a parameter vector sample generator using a collective variational inference method. In a blind testing, the VPG developed by combining the two could generate a large number of realistic virtual patients (VPs) which could simulate physiological responses observed in all the experiments: on the average, 98.1% and 74.3% of the randomly generated VPs were physiologically legitimate and adequately replicated the test subjects, respectively, and 92.4% of the experimentally observed responses could be covered by the envelope formed by the subject-replicating VPs. In sum, the VPG developed in this paper may be useful for pre-clinical in silico evaluation of autonomous vasopressor administration algorithms.
|
|
10:00-10:15, Paper TuAT5.4 | |
Hypothermic Peritoneal Perfusion of Cold Oxygenated Perfluorocarbon May Improve the Efficacy of Extracorporeal Oxygenation: A Mathematical Model-Based Analysis |
|
Rezaei, Parham | University of Maryland, College Park |
Leibowitz, Joshua | University of Maryland |
Kadkhodaeielyaderani, Behzad | University of Maryland, College Park |
Moon, Yejin | University of Maryland |
Awad, Morcos | University of Maryland School of Medicine |
Stachnik, Stephen | University of Maryland |
Sarkar, Grace | University of Maryland |
Shaw, Anna | University of Maryland, College Park |
Naselsky, Warren | University of Maryland, School of Medicine |
Enofe, Nosayaba | Temple University, Temple University Hospital |
Stewart, Shelby | University of Maryland |
Culligan, Melissa | Temple University |
Friedberg, Joseph | Temple University |
Yu, Miao | University of Maryland |
Hahn, Jin-Oh | University of Maryland |
Keywords: Healthcare systems, Modelling and Control of Biomedical Systems
Abstract: Circulation of perfluorocarbon (PFC) through corporeal cavities (such as peritoneum and intestinal lumen) has received interest by virtue of its potential to supplement oxygenation through the lungs via mechanical ventilation. However, the technology is not yet mature enough for clinical application, due to the knowledge gaps regarding the limiting factors hampering oxygen transport from PFC to blood: diffusion, interfacial surface area, and peritoneal blood perfusion to list a few. How important these factors are relative to each other, and how these factors can be overcome remain unknown. In this paper, we investigate a novel hypothesis that hypothermic peritoneal perfusion of cold oxygenated PFC may improve oxygenation of blood by facilitating the diffusion of oxygen from PFC to blood. Our hypothesis originates from physics-inspired insights that both hypothermia and PFC cooling may facilitate the oxygen delivery from PFC to blood by increasing PFC-to-blood oxygen tension gradient: (i) hypothermia may decrease venous oxygen tension while (ii) cooling PFC may increase oxygen tension therein by increasing its oxygen solubility. By developing and analyzing a physics-based mathematical model capable of simulating oxygen tension responses to mechanical ventilation and peritoneal PFC perfusion under normothermic and hypothermic conditions, we analyzed the effect of hypothermic peritoneal cold PFC perfusion on blood oxygenation. The results predicted that peripheral oxygen saturation may be improved by 5%-10% by peritoneal perfusion of oxygenated 15°C PFC at 32°C body temperature compared with peritoneal perfusion of oxygenated 37.5°C PFC at 37.5°C body temperature. The results also predicted that cooling PFC may play a more meaningful role than hypothermia in facilitating the diffusion of oxygen. Pending the investigation of adverse impact of hypothermia and cold PFC on homeostasis, hypothermic cold PFC perfusion may improve peritoneal oxygenation by facilitating diffusion.
|
|
10:15-10:30, Paper TuAT5.5 | |
LSTM-Based Estimation of Time-Varying Parameters in a Spatiotemporal PDE Model for Prediction of Epidemic Spread |
|
David, Deepak Antony | University of Cincinanti |
Street, Logan | University of Cincinnati |
Ramakrishnan, Subramanian | University of Dayton |
Kumar, Manish | University of Cincinnati |
Keywords: Machine Learning in modeling, estimation, and control, Stochastic Systems, Modeling and Validation
Abstract: In this paper we apply a Long-Short-Term-Memory (LSTM) deep learning method for forecasting time-varying parameters of a Partial Differential Equation (PDE)-based compartmental dynamic model of epidemic spread. The predictive efficacy of such models depends on accurately estimating and updating time-varying model parameters, based on empirical infection data updated daily by public health systems during an epidemic. Investigating the role of deep learning methods is important in this context and motivates our work. We first note that numerical data used in this work correspond to empirical COVID-19 infection data for the state of Ohio, USA. The LSTM is subject to an iterative training process for a total period of 30 days such that model parameters are generated for each day. Each iteration generates parameter values corresponding to a specific day and therefore yields a comprehensive representation of the temporal dynamics of the infection progression. Moreover, the training process is designed to ensure both the model’s accuracy and predictive reliability. Using the day-to-day infection parameters yielded by the LSTM as input to the PDE model, a forecast of infection spread is obtained from the latter and validated against empirical COVID-19 spread data. In summary, by combining advanced deep learning techniques with epidemiological modeling, this study advances both our understanding of the complex, time-varying dynamics of epidemics and our ability to accurately forecast the dynamics from available empirical data.
|
|
10:30-10:45, Paper TuAT5.6 | |
Stochastic Predictive Control with Time-Joint State Constraint Violation Probability Objective for Drone Delivery of Medical Products |
|
Tran, Trung | The University of Michigan |
Kolmanovsky, Ilya V. | University of Michigan |
Keywords: Control Applications, Transportation Systems, Unmanned Ground and Aerial Vehicles
Abstract: In this paper, we apply stochastic predictive control to the problem of routing drone to deliver medical products. Our vehicle routing problem (VRP) considers a dynamically evolving demand, multiple depots, and the vehicle's battery and delivery capacity. The objective function for minimization is the time-joint state constraint violation probability (TJSCVP). In the context of medical drone delivery, constraint violation refers to the events where the drone runs out of battery or when the customers (local healthcare facilities) run out of medical stock. Thus our approach aims at maximizing the viability of the drone delivery operations. An illustrative example of medical drone delivery in Rwanda is considered.
|
|
TuAT6 |
Streeterville W |
Recent Advances in Control and Estimation Theory (2) |
Regular Session |
Chair: Radisavljevic-Gajic, Verica | Ajman Univeristy |
Co-Chair: Mohammadpour Velni, Javad | Clemson University |
|
09:15-09:30, Paper TuAT6.1 | |
Variable Impedance Control Using Deep Geometric Potential Fields |
|
Potu Surya Prakash, Nikhil | University of California Berkeley |
Seo, Joohwan | University of California, Berkeley |
Sreenath, Koushil | University of California, Berkeley |
Choi, Jongeun | Yonsei University |
Horowitz, Roberto | Univ. of California at Berkeley |
Keywords: Machine Learning in modeling, estimation, and control, Adaptive and Learning Systems, Robotics
Abstract: In this paper, we present a novel strategy for developing control laws for fully actuated mechanical systems evolving on smooth manifolds utilizing potential functions parameterized by neural networks. This method builds invariant conservative and dissipative potential functions using neural networks that generate a stabilizing nonlinear elastic and damping wrench pair (stable potential field). These elastic and damping forces can be used to replace the proportional and derivative control terms in impedance control to have more representative controllers that can be made to mimic expert demonstrations for kinesthetic teaching or to improve performance using an LQR style formulation. Furthermore, the principle of invariance is instrumental in enhancing the transferability of learning across different scenarios through the invariant potential functions.
|
|
09:30-09:45, Paper TuAT6.2 | |
Adaptive Uncertainty Quantification for Scenario-Based Control Using Meta-Learning of Bayesian Neural Networks |
|
Bao, Yajie | Intelligent Fusion Technology, Inc |
Mohammadpour Velni, Javad | Clemson University |
Keywords: Uncertain Systems and Robust Control, Machine Learning in modeling, estimation, and control, Control Design
Abstract: Scenario-based optimization and control has proven to be an efficient approach to account for system uncertainty. In particular, the performance of scenario-based model predictive control (MPC) schemes depends on the accuracy of uncertainty quantification. However, current learning- and scenario-based MPC (sMPC) approaches employ a single time-invariant probabilistic model (learned offline), which may not accurately describe time-varying uncertainties. Instead, this paper presents a model-agnostic meta-learning (MAML) of Bayesian neural networks (BNN) for adaptive uncertainty quantification that would be subsequently used for adaptive scenario-tree model predictive control design of nonlinear systems with unknown dynamics to enhance control performance. In particular, the proposed approach learns both a global BNN model and an updating law to refine the BNN model. At each time step, the updating law transforms the global BNN model into more precise local BNN models in real time. The adapted local model is then used to generate scenarios for sMPC design at each time step. A probabilistic safety certificate is incorporated in the scenario generation to ensure that the trajectories of the generated scenarios contain the real trajectory of the system and that all the scenarios adhere to the constraints with a high probability. Experiments using closed-loop simulations of a numerical example demonstrate that the proposed approach can improve the performance of scenario-based MPC compared to using only one BNN model learned offline for all time steps.
|
|
09:45-10:00, Paper TuAT6.3 | |
Relational Weight Optimization for Enhancing Team Performance in Multi-Agent Multi-Armed Bandits |
|
Kotturu, Monish Reddy | University of Massachusetts Lowell |
Vahedian Movahed, Saniya | University of Texas San Antonio |
Robinette, Paul | MIT |
Jerath, Kshitij | University of Massachusetts Lowell |
Redlich, Amanda | University of Massachusetts Lowell |
Azadeh, Reza | University of Massachusetts Lowell |
Keywords: Multi-agent and Networked Systems, Optimal Control, Stochastic Systems
Abstract: We introduce an approach to improve team performance in a Multi-Agent Multi-Armed Bandit (MAMAB) framework using Fastest Mixing Markov Chain (FMMC) and Fastest Distributed Linear Averaging (FDLA) optimization algorithms. The multi-agent team is represented using a fixed relational network and simulated using the Coop-UCB2 algorithm. The edge weights of the communication network directly impact the time taken to reach distributed consensus. Our goal is to shrink the timescale on which the convergence of the consensus occurs to achieve optimal team performance and maximize reward. Through our experiments, we show that the convergence to team consensus occurs slightly faster in large constrained networks.
|
|
10:00-10:15, Paper TuAT6.4 | |
Closed-Form Robust Safe Output-Feedback Control Design for Nonlinear Systems |
|
Dini, Navid | The University of Memphis |
Batmani, Yazdan | University of Kurdistan |
Davoodi, Mohammadreza | The University of Memphis |
Keywords: Nonlinear Control Systems, Optimal Control, Control Design
Abstract: Safety is a critical consideration in designing autonomous systems operating in complex, dynamic environments. Control barrier functions have emerged as a powerful framework for ensuring safety and performance in these systems. This paper presents a novel closed-form solution for designing safe controllers in output feedback control systems, eliminating the need for real-time quadratic programming optimization. This simplification facilitates the implementation of safety-critical controllers in control systems. The proposed method ensures safety and stability in the presence of bounded external disturbances and measurement noise. Simulation results demonstrate the effectiveness of this methodology.
|
|
10:15-10:30, Paper TuAT6.5 | |
Nash or Stackelberg? - a Comparative Study for Game-Theoretic Autonomous Vehicle Decision-Making |
|
Bateman, Brady | University of Missouri |
Xin, Ming | University of Missouri |
Tseng, Eric | Ford Motor Company |
Liu, Mushuang | University of Missouri |
Keywords: Optimal Control, Nonlinear Control Systems, Control Design
Abstract: This paper studies game-theoretic decision-making for autonomous vehicles (AVs). A receding horizon multi-player game is formulated to model the AV decision-making problem. Three classes of games, including Nash equilibrium games, strong Stackelberg equilibrium games, and weak Stackelberg equilibrium games are developed respectively. For each class of games, two solution settings, including pairwise games and multi-player games, are introduced, respectively. Comparative studies are conducted via statistical simulations to gain understandings of the performance of the three classes of games and of the two solution settings, respectively. The simulations are conducted in intersection-crossing scenarios, and the game performance is quantified by three metrics: safety, travel efficiency, and computational time.
|
|
10:30-10:45, Paper TuAT6.6 | |
Discrete-Time Linear-Quadratic Optimal Controller Driven by a Reduced-Order Observer Steady State Performance Loss |
|
Radisavljevic-Gajic, Verica | Ajman Univeristy |
Keywords: Estimation, Linear Control Systems, Optimal Control
Abstract: This paper formulates and solves a problem encountered in engineering practice when a discrete-time linear-quadratic optimal feedback controller uses state estimates obtained via a discrete-time reduced-order observer. Due to the use of state estimates instead of the actual state variables, the optimal quadratic performance is degraded in a pretty complex manner. The paper shows how to find the exact expression for the optimal performance degradation for the steady state case in terms of a solution of a reduced-order discrete-time algebraic Lyapunov equation. The quantities that impact the performance criterion loss are identified. Simulation results show that the optimal performance degradation can be considerably reduced by using the least square method to set up the reduced-order observer initial condition.
|
|
TuAT7 |
Prime 3 |
Past, Current, and Future of Robotaxi |
Tutorial Session |
Chair: Chen, Yan | Arizona State University |
Organizer: Chen, Yan | Arizona State University |
|
, Paper TuAT7.0 | |
Past, Current, and Future of Robotaxi (I) |
|
Chen, Yan | Arizona State University |
Li, Bin | Cummins |
|
, Paper TuAT7.0 | |
Past, Current, and Future of Robotaxi (I) |
|
Chen, Yan | Arizona State University |
Jiang, Yu | ClearMotion, Inc. |
|
09:15-09:30, Paper TuAT7.1 | |
The History of Robotaxi (I) |
|
Bin-Nun, Amitai | Cruise |
|
09:30-09:45, Paper TuAT7.2 | |
The Current State of Robotaxis (I) |
|
Li, Bin | Cummins |
|
09:45-10:00, Paper TuAT7.3 | |
The Evolving Landscape of Robo (I) |
|
Jiang, Yu | ClearMotion, Inc. |
|
TuBT1 |
Avenue Ballroom E |
Late-Breaking Research Results (2) |
Regular Session |
Chair: Landers, Robert G. | University of Notre Dame |
Co-Chair: Vermillion, Christopher | University of Michigan |
|
15:00-15:03, Paper TuBT1.1 | |
A Part-Scale Thermal Model for Laser Powder Bed Fusion Using U-Net with Attention |
|
Wang, Yanwen | Pennsylvania State University |
Wang, Qian | Penn State University |
|
15:03-15:06, Paper TuBT1.2 | |
Control Barrier Proximal Dynamics for Conservation-Based Energy Systems |
|
Marvi, Zahra | University of Minnesota |
Bullo, Francesco | Univ of California, Santa Barbara |
Alleyne, Andrew G. | University of Minnesota |
|
15:06-15:09, Paper TuBT1.3 | |
A Data-Driven Reduced Model of Lithium-Sulfur Battery Discharge |
|
Haddad, Noushin | University of Maryland, College Park |
Fathy, Hosam K. | University of Maryland |
|
15:09-15:12, Paper TuBT1.4 | |
Thermal and Morphology Control in Digital Glass Forming Processes |
|
Tiwari, Balark | University of Notre Dame |
Khadka, Nishan | University of Notre Dame |
Bos, Andre | Los Alamos National Laboratory |
Meredith, Doug | Los Alamos National Laboratory |
Bernardkin, John | Los Alamos National Laboratory |
Kinzel, Edward | University of Notre Dame |
Landers, Robert G. | University of Notre Dame |
|
15:12-15:15, Paper TuBT1.5 | |
Digital Metal Forming of Bending Processes |
|
Wang, Yi | University of Notre Dame |
Tiwari, Balark | University of Notre Dame |
Li, Zongze | University of Notre Dame |
Landers, Robert G. | University of Notre Dame |
|
15:15-15:18, Paper TuBT1.6 | |
Towards Battery Formation Protocol Optimization Via Pressure, Temperature, and Current Control: New Experimental and Modeling Insights |
|
Weng, Andrew | University of Michigan |
Siegel, Jason | University of Michigan |
Stefanopoulou, Anna G. | Univ of Michigan |
|
15:18-15:21, Paper TuBT1.7 | |
Estimating the State of Charge of Parallel-Connected Lithium-Ion Battery Cells Using Inverse-Causality Dynamic Models |
|
Lee, Hannah | Princeton University |
Casten, Casey | University of Maryland, College Park |
Fathy, Hosam K. | University of Maryland |
|
15:21-15:24, Paper TuBT1.8 | |
Dynamic Thermal Modeling of Air-Cooled Li-Ion Battery Pack Systems |
|
Hemmat, Mahsa | University of Minnesota |
Alleyne, Andrew G. | University of Minnesota |
|
15:24-15:27, Paper TuBT1.9 | |
Characterization of Human Driving Behaviors in Shared Vehicle Control Based on Level-K Game Theory |
|
Dudek, Aleksandra Anna | University of Michigan |
James, Scott | Applied Dynamics International |
Castanier, Matthew | US Army DEVCOM GVSC |
Vermillion, Christopher | University of Michigan |
Barton, Kira | University of Michigan |
|
15:27-15:30, Paper TuBT1.10 | |
Grey-Box Modeling of Human Cognitive State Dynamics in Automated Driving Contexts |
|
Jeevanandam, Sibibalan | Purdue University |
Wang, Xipeng | Purdue University |
Hsieh, Tyler | Purdue University |
Jain, Neera | Purdue University |
|
15:30-15:33, Paper TuBT1.11 | |
Throughput Gained versus Days Lost (TvD) in V2G Services Considering Multiple Battery Degradation Mechanisms |
|
Movahedi, Hamidreza | University of Michigan |
Pannala, Sravan | University of Michigan |
Siegel, Jason | University of Michigan |
Stefanopoulou, Anna G. | Univ of Michigan |
|
15:33-15:36, Paper TuBT1.12 | |
Barrier Function-Based Safety Control of Fluid Resuscitation |
|
Yaculak, Stacey | University |
Hahn, Jin-Oh | University of Maryland |
Fathy, Hosam K. | University of Maryland |
|
15:36-15:39, Paper TuBT1.13 | |
Nonlinear Model Predictive Control of a Latent Thermal Energy Storage Device for Electronics Cooling Applications |
|
Jain, Neera | Purdue University |
Gulewicz, Demetrius | Purdue University |
|
15:39-15:42, Paper TuBT1.14 | |
Online Sensing and Control Strategies for Precision Application Processes |
|
Hawa, Angelo | University of Michigan |
Barton, Kira | University of Michigan |
|
15:42-15:45, Paper TuBT1.15 | |
Battery System Identification with Coupled Nonlinear Electro-Thermal Dynamics Via Bayesian Optimization |
|
Tu, Hao | University of Kansas |
Fang, Huazhen | University of Kansas |
|
15:45-15:48, Paper TuBT1.16 | |
Online, Nonparametric, Risk-Aware Optimization of Combustion Efficiency in Compression-Ignition Engines Using Gaussian Kernel Density Estimation |
|
Ahmed, Omar | University of Michigan |
Middleton, Robert | University of Michigan |
Stefanopoulou, Anna G. | Univ of Michigan |
Kim, Kenneth | DEVCOM Army Research Laboratory |
Kweon, Chol-Bum | DEVCOM Army Research Laboratory |
|
15:48-15:51, Paper TuBT1.17 | |
Experimental Implementation of Differentially Flat Reference Models for Bipedal Walking Robots |
|
Kumar, Akshay | Indian Institute of Technology Bombay |
Gumalapuram, Manideep | Indian Institute of Technology Bombay |
Sangwan, Vivek | Indian Institute of Technology Bombay |
|
TuBT2 |
Avenue Ballroom W |
Advanced Control Systems and Robotics |
Regular Session |
Co-Chair: Abaid, Nicole | Virginia Polytechnic Institute and State University |
|
15:00-15:03, Paper TuBT2.1 | |
Nonlinear Disturbance Observer with Sliding Mode Control for a Fabric Soft Robotic Arm |
|
Qiao, Zhi | Arizona State University |
Tao, Weijia | Arizona State University |
Zhang, Wenlong | Arizona State University |
Keywords: Soft Robotics, Robotics, Control Applications
Abstract: Soft robotics, renowned for their flexibility, have become increasingly prevalent in interaction tasks due to their inherent compliance. For practical applications, simplified low-dimensional models are conventionally employed to represent soft robot dynamics; however, these models suffer from large uncertainties which present significant challenges for control design. This paper introduces a robust control strategy that combines a linear parameter-varying (LPV) model with a nonlinear disturbance observer-based sliding mode control (NDOSMC). The NDOSMC consists of a first-order LPV dynamic model of the soft actuator, a fast nonlinear disturbance observer for precise and rapid disturbance estimation, and a sliding mode control for improved robustness against uncertainties. The closed-loop tracking control system is proven to be asymptotically stable. The proposed algorithm is applied in a trajectory tracking task for a two-segment inflatable soft robot arm. The effectiveness of the controller is validated through both simulation and experimental results. Notably, reductions in the root mean square errors, averaged over seven experimental trials, are observed across all state variables. This demonstrates the controller's high trajectory tracking accuracy, even in the presence of underlying model uncertainties.
|
|
15:03-15:06, Paper TuBT2.2 | |
On Fusing Active and Passive Acoustic Sensing for Simultaneous Localization and Mapping |
|
Bradley, Aidan | Virginia Tech |
Abaid, Nicole | Virginia Polytechnic Institute and State University |
Keywords: Estimation, Multi-agent and Networked Systems, Sensors and Actuators
Abstract: Evidence suggests that bats are able to take advantage of both their own echolocation signals (active sensing) and the signals of conspecifics in their environment (passive sensing). This work follows a bioinspired approach to investigate whether we can enable robots to do the same. We have simulated a pair of vehicles that acoustically sense their environment both actively and passively. Our results show that, while the ability to fuse acoustic sensing techniques may not provide a significant improvement over active sensing alone, it is rarely worse and often allows for more information about the environment to be observed.
|
|
15:06-15:09, Paper TuBT2.3 | |
Enhancing Prosthetic Control with Ultrasound Imaging: A Convolutional Neural Network Approach for Hand Gesture Recognition |
|
Chen, Yun | The University of Alabama |
Bao, Xuefeng | University of Pittsburgh |
He, Hongsheng | The University of Alabama |
Zhang, Qiang | The University of Alabama |
Keywords: Biomechanical Systems, Modelling and Control of Biomedical Systems, Machine Learning in modeling, estimation, and control
Abstract: Human hand gesture recognition using biological signals from the forearm is an increasingly significant area of research, with implications across various fields such as prosthetic development, rehabilitation, and human-machine interaction. However, traditional hand gesture recognition with surface electromyography (sEMG) technique has some challenges, including cross-talk from neighboring muscles, low signal-to-noise ratio, and inability to measure deep muscles. In the current study, we proposed to use brightness mode (B-mode) ultrasound images from the muscles of the forearm anterior side as an alternative neuromuscular interface to recognize hand gestures. We designed a convolutional neural network (CNN) classifier to build the personalized mapping from static ultrasound images to eight different hand gestures. To evaluate the performance of the proposed CNN classifier, an ultrasound images dataset and labeled gestures from four young healthy participants were collected and analyzed. Results from offline intra-subject (personalization) validation, quasi-real-time validation, and real-time validation showed high classification accuracy of 99.65%, 97.47%, and 90.83%, respectively. In addition, real-time hand gesture recognition could be executed within 50 ms per image frame by using the proposed CNN classifier. Our findings demonstrated promising real-time hand gesture recognition with high accuracy by using B-mode ultrasound images and the proposed CNN classifier for prosthetic hand control.
|
|
15:09-15:12, Paper TuBT2.4 | |
Uncertainty Rejection in Multirotor Motor-Propeller Actuators Using RPM Feedback |
|
Charla, Sesha | Purdue University |
Yao, Bin | Purdue University |
Voyles, Richard | Purdue Univesity |
Keywords: Control Design, Control Applications, Unmanned Ground and Aerial Vehicles
Abstract: In this study, we introduce a control strategy utilizing Desired Compensation Adaptive Robust Control (DCARC) to improve input to propeller velocity tracking in motor-propeller actuators for reducing the uncertainties in the thrust generation process. The DCARC approach minimizes the influence of sensor noise on model compensation input by using states from desired trajectory with estimated parameters for its calculation. Additionally, the design of the reference governor accounts for input saturation effects on the system states for optimal performance within the operating limits. The effectiveness of the proposed control strategy is demonstrated through experimental validation on a motor-propeller test setup.
|
|
15:12-15:15, Paper TuBT2.5 | |
Data-Driven Safe Control of Stochastic Nonlinear Systems |
|
Esmaeili, Babak | Michigan State University |
Modares, Hamidreza | Michigan State University |
Keywords: Nonlinear Control Systems, Stochastic Systems, Automotive Systems
Abstract: This paper introduces a data-based safe control design for stochastic nonlinear systems. The controller consists of two parts: a linear component ensuring lambda-contractivity for set invariance, and a nonlinear component minimizing the impact of nonlinearities. A closed-form presentation of both dynamics is provided, linking control gains directly to data and decision variables. The design is formulated as a semidefinite programming problem (SDP) to ensure robust set invariance, incorporating extra constraints for unmeasured noise and residual nonlinear effects. The effectiveness of this approach is validated through simulations, demonstrating its potential to enhance safety and performance in controlling nonlinear dynamics.
|
|
15:15-15:18, Paper TuBT2.6 | |
A Koopman-Based Approach for Torque Vectoring in Electric Vehicles |
|
Sassella, Andrea | Politecnico Di MIlano |
Lenzo, Basilio | University of Padua |
Keywords: Modelling, Identification and Signal Processing, Automotive Systems, Control Applications
Abstract: Torque Vectoring plays a pivotal role in enhancing vehicle dynamics and performance. In electric vehicles with multiple motors, the torque at each wheel may be controlled independently, offering significant opportunity to enhance safety and stability. This study explores the application of Koopman-based Model Predictive Control (MPC) in torque vectoring systems. By considering an electric vehicle equipped with four in-wheel motors, a linear Koopman model is identified from data. The model is then used to formulate a linear MPC. The performance of the scheme is assessed through numerical simulations.
|
|
15:18-15:21, Paper TuBT2.7 | |
Impact-Free Gaits for Planar Bipeds: Changing Walking Speed and Gait |
|
Khandelwal, Aakash | Michigan State University |
Kant, Nilay | Mainspring Energy |
Mukherjee, Ranjan | Michigan State Univ |
Keywords: Path Planning and Motion Control, Robotics
Abstract: The problems of changing the walking speed and stride length of impact-free gaits for point-foot planar bipeds are addressed. The impact-free gaits are designed using an approach developed in prior work. It is shown that the impulse controlled Poincare map (ICPM) approach can be modified to transition between orbits defining gaits with different walking speeds, and the continuous controller can be changed during the swing phase to transition between gaits that have distinct stride lengths. The effectiveness of the approaches is demonstrated using simulations carried out on a five-link biped.
|
|
15:21-15:24, Paper TuBT2.8 | |
Coordinated Admittance-Impedance Control with Force Excitation for Compliant, Underactuated Aerial Manipulation |
|
McArthur, Daniel R. | Sandia National Laboratories |
Slightam, Jonathon | Sandia National Laboratories |
Spencer, Steven | Sandia National Laboratories |
Buerger, Stephen P. | Sandia National Laboratories |
Keywords: Unmanned Ground and Aerial Vehicles, Robotics, Control Design
Abstract: This work presents a coordinated admittance-impedance control architecture for underactutated unmanned aerial manipulators (UAM) that maintains inherent stability during physical contact with the environment. Utilizing dynamically reconfigurable in-flight system parameters, the architecture’s intuitive control interface allows aerial manipulation (AM) task execution via human-guided teleoperation or automated maneuvers/trajectories. An admittance-controlled unmanned aerial vehicle (UAV) and impedance-controlled manipulator arm are coordinated by a shared force excitation. The effectiveness of the proposed control architecture is validated in hardware experiments demonstrating smooth transitions between free-flight and environmental contact, along with the ability to maintain constant contact while sliding the UAM's end-effector vertically and horizontally across a surface.
|
|
15:24-15:27, Paper TuBT2.9 | |
Multi-Agent Navigation Using Convex Lifting on Dynamic Environment |
|
Konyalioglu, Turan | Centrale-Supélec |
Olaru, Sorin | CentraleSupelec |
Niculescu, Silviu-Iulian | Laboratory of Signals and Systems (L2S) |
Ballesteros-Tolosana, Iris | Renault SAS, CentraleSupelec |
Flores, Carlos | Ampere Software Technologies |
Keywords: Path Planning and Motion Control, Multi-agent and Networked Systems, Optimal Control
Abstract: This paper revisits the conventional convex lifting method for space partition in static topologies by bringing obstacle avoidance guarantees for multi-agent navigation strategies. Further, the present work proposes a new methodology that adapts the convex lifting method in a dynamic environment. To complete the developments towards navigation, the paper introduces a Model Predictive Control (MPC) method for trajectory planning and navigation for an agent and presents examples of a multi-agent system in a cluttered environment prioritizing the safety and control objectives of each agent.
|
|
15:27-15:30, Paper TuBT2.10 | |
Precision Plowing: An Approach for Clods Size Estimation Via ECOC Classifier |
|
Gambarotto, Luca | Dipartimento Di Elettronica Informazione E Bioingegneria, Polite |
Corno, Matteo | Politecnico Di Milano |
Savaresi, Sergio | Politecnico Di Milano |
Benvenuti, Davide | SDF Group |
Portanti, Samuele | Dipartimento Di Elettronica Informazione E Bioingegneria, Polite |
Conconi, Andrea | SDF Group |
Keywords: Agricultural Systems, Estimation, Machine Learning in modeling, estimation, and control
Abstract: Automatic agricultural guidance requires an accurate soil quality assessment during tillage operations. This is particularly true for assisted plowing, where the tractor speed and the plow-specific settings (pitch, sinking, and aperture) should be adjusted to guarantee an optimal soil crumbling. In this paper, we present a novel method for estimating soil clod size after the plowing activity, using a mono-camera mounted on the cabin of a tractor. Our technique relies on the Bird's Eye View reconstruction of a soil patch and its classification using an Error-Correcting Output Codes (ECOC) classifier, trained on features extracted from grayscale images. The main novelty of our approach lies in the real-time implementation on a moving vehicle. Furthermore the use of the ECOC classifies yields precise and efficient assessment of the clod size. Experimental results on real in-field collected data demonstrate that our estimator is sufficiently robust and accurate to provide a solid basis for automatic adjustment of plow settings during agricultural operations with a moving vehicle.
|
|
15:30-15:33, Paper TuBT2.11 | |
Lifted Explicit Interpolating Control for Low-End Embedded Microcontrollers: An Active Vibration Control Case Study |
|
Gulan, Martin | Slovak University of Technology in Bratislava |
Takács, Gergely | Automationshield.com |
Olaru, Sorin | CentraleSupelec |
Keywords: Control Design, Linear Control Systems, Control Applications
Abstract: Motivated by contributing to the use of advanced, model-based control in embedded applications using extremely limited computing hardware, in this paper we present a constrained control alternative to linear model predictive control (MPC). In the proposed two-layer scheme, we revisit the interpolation-based control framework, and show that the structure of its explicit solution can be conveniently exploited by the convex lifting concept to generate low-complexity controllers. Their performance, fast evaluation time and low memory footprint are demonstrated in an active vibration control case study, where deployed on an 8-bit ATtiny85 microcontroller.
|
|
15:33-15:36, Paper TuBT2.12 | |
A Feed-Forward Featured Cascade PI Control Strategy for Tram Active Steering |
|
Ren, Xiaotao | Traila AG |
Brack, Tobias | Traila AG |
Morris, Tom | Traila AG |
Keywords: Control Design, Control Applications, Linear Control Systems
Abstract: Active steering is a technique that allows a railway vehicle to steer itself by adjusting the angle of a wheelset in order to avoid contact between flange and rail. This can improve wear, ride comfort, safety, and energy efficiency. This article examines the control requirements for a tram vehicle navigating curves with radii as small as 20 meters. An analysis of the angle that an actively steered axle has to maintain reveals that entering and leaving a curve represent a quadratic disturbance to the angle. To deal with this disturbance, a cascade PI controller is proposed. The controller consists of two loops: an inner loop that controls the wheel angle and an outer loop that controls the lateral deviation of the tram. The controller parameters are designed using a linearized model of the tram and are validated by simulations. The results show that the cascade PI controller can effectively reduce the quadratic disturbance and improve the lateral performance of the tram in curves.
|
|
15:36-15:39, Paper TuBT2.13 | |
State Estimation for a Tethered Underwater Kite |
|
Bhattacharjee, Debapriya | University of Maryland, College Park |
Fathy, Hosam K. | University of Maryland |
Keywords: Power and Energy Systems, Estimation, Underwater Vehicles
Abstract: This paper presents a state and fluid velocity estimator for a tethered underwater kite. The proposed approach does not require fluid dynamics modeling or expensive instrumentation. The paper formulates a state and disturbance estimation problem, where estimating the kite’s state enables the ultimate goal of free stream fluid velocity estimation. The formulation assumes that one can measure the kite’s tether length, the tether’s azimuth and elevation angles, the kite’s Euler angles, and the relative fluid velocity in the body frame. We use unscented Kalman filtering to estimate the kite’s states and free stream fluid velocity. The estimator provides good accuracy if kite attitude measurements are available: a conclusion examined both in simulation and using Fisher analysis.
|
|
15:39-15:42, Paper TuBT2.14 | |
Sampling-Based Risk-Aware Path Planning Around Dynamic Engagement Zones |
|
Wolek, Artur | University of North Carolina at Charlotte |
Weintraub, Isaac | Air Force Research Laboratory |
Von Moll, Alexander | Air Force Research Laboratory |
Casbeer, David | Air Force Research Laboratory |
Manyam, Satyanarayana Gupta | Infoscitex Corp. (AFRL) |
Keywords: Path Planning and Motion Control, Unmanned Ground and Aerial Vehicles, Optimal Control
Abstract: Existing methods for avoiding dynamic engagement zones (EZs) and minimizing risk leverage the calculus of variations to obtain optimal paths. While such methods are deterministic, they scale poorly as the number of engagement zones increases. Furthermore, optimal-control based strategies are sensitive to initial guesses and often converge to local, rather than global, minima. This paper presents a sampling-based approach to obtain a feasible flight plan for a Dubins vehicle to reach a desired location in a bounded operating region in the presence of a large number of engagement zones. The dynamic EZs are coupled to the vehicle dynamics through its heading angle. Thus, the dynamic two-dimensional obstacles in the (x,y) plane can be transformed into three-dimensional static obstacles in a lifted (x,y,heading) space. This insight is leveraged in the formulation of a Rapidly-exploring Random Tree (RRT*) algorithm. The algorithm is evaluated with a Monte Carlo experiment that randomizes EZ locations to characterize the success rate and average path length as a function of the number of EZs and as the computation time made available to the planner is increased.
|
|
15:42-15:45, Paper TuBT2.15 | |
Video-Rate AFM Imaging Using Signal Transformation Technique for Raster Scanning |
|
Khodabakhshi, Erfan | The University of Texas at Dallas |
Moheimani, S.O. Reza | University of Texas at Dallas |
Keywords: Mechatronic Systems, Motion and Vibration Control, Control Design
Abstract: Achieving high-speed, high-resolution positioning is crucial yet challenging due to the increased sensitivity to measurement noise at higher bandwidths. This study explores a signal-transformation-based control technique for enhancing raster scanning in atomic force microscopy. An integral resonant controller is employed to augment the closed-loop bandwidth by damping the dominant mode of the scanner's fast axis. The nanopositioner's first resonance frequency exceeds 15 kHz with the open-loop scan range exceeding 5.8 μm. A double integrator is wrapped around the damping loop to improve the tracking performance. Subsequently, the signal transformation approach (STA) is implemented within the closed-loop system, and its effectiveness is benchmarked against a signal-preshaping method. Tracking experiments were conducted across scanning frequencies ranging from 10 Hz to 300 Hz, encompassing a 2 μm × 2 μm scan area. The root mean square (RMS) tracking error was maintained below 71 nm. Experimental results underscore the technique's effectiveness, notably achieving rapid time-lapse AFM imaging at rates up to 10 frames per second.
|
|
15:45-15:48, Paper TuBT2.16 | |
An Analytical Method for Finding All Dynamically Admissible Paths Around an Obstacle That Maximize Friction Utilization |
|
Pagan, Michael | Hendrick Motorsports |
Harnett, Stephen Joseph | The Pennsylvania State University |
Pentzer, Jesse | The Applied Research Laboratory, the Pennsylvania State Universi |
Reichard, Karl | The Applied Research Laboratory, the Pennsylvania State Universi |
Brennan, Sean | Pennsylvania State University |
Keywords: Path Planning and Motion Control, Robotics, Optimal Control
Abstract: Path planning through obstacle fields is critically important for advancing the safety and utility of autonomous off-road vehicles. This paper describes the development of an analytical method for modifying point-to-point paths into high-speed paths. The method starts by using, as an input, a point-to-point path plan through an obstacle field such as one generated from an A* type algorithm. This point-to-point method represents the shortest path solution as a series of waypoints that the vehicle must reach in sequence, with each waypoint typically constrained by a nearby obstacle. Because the waypoints produce a C0 continuous but non-differentiable path, the implementation of a waypoint-following method at high speed often requires a vehicle to slow almost completely to a stop, turn, and then speed up at every waypoint. To generate a high-speed path, the point-to-point path must be modified into line segments connected by C1+ smooth curves to maximize vehicle speed while keeping the constraint of avoiding all obstacle collisions. This paper develops such a high-speed path-planning algorithm. The algorithm designs a path consisting of straight line and constant-radius arcs that meet acceleration and speed limits. These line segments and arcs optimize the utilization of available surface friction or, via straightforward transformations, user-defined limits on lateral or longitudinal accelerations including powertrain limits, rollover limits, etc.
|
|
15:48-15:51, Paper TuBT2.17 | |
Automatic Flight Control for a Quadrotor Drone |
|
Wi, Yejin | University of Houston |
Cescon, Marzia | University of Houston |
Keywords: Unmanned Ground and Aerial Vehicles, Control Design, Control Applications
Abstract: In this paper, we detail the design of the control system, which exploits the identified models aimed at achieving trajectory tracking. The identified quadrotor’s in-plane and vertical dynamics through the Optimized Predictor-Based Subspace Identification (PBSIDopt) technique were used for control design. The classical cascade architecture, which is comprised of a PID compensator in the inner loop and a proportional controller in the outer loop, was designed through frequency analysis. The designed control system was validated both in simulation and with actual flight experiments.
|
|
15:51-15:54, Paper TuBT2.18 | |
Super-Twisting Impedance Control for Robust and Compliant Interaction Using a Redundant Robot Manipulator |
|
Dyrhaug, Jan Inge | Norwegian University of Science and Technology (NTNU) |
Foseid, Eirik Lothe | Norwegian University of Science and Technology |
Schmidt-Didlaukies, Henrik M. | Norwegian University of Science and Technology |
Basso, Erlend A. | Norwegian University of Science and Technology |
Pettersen, Kristin Y. | Norwegian Univ. of Science and Tech |
Gravdahl, Jan Tommy | Norwegian University of Science and Technology (NTNU) |
Keywords: Robotics, Uncertain Systems and Robust Control, Nonlinear Control Systems
Abstract: To perform robotic intervention tasks reliably with high accuracy under uncertainty and unknown disturbances, robust control methods are required. A problem using standard robust control techniques, however, is that the contact forces cannot be considered as disturbances in intervention operations and compliance to the unknown contact geometry and forces is crucial. In a recent paper, we proposed a control method based on the generalized super-twisting algorithm, which was proved to achieve the desired impedance with respect to the contact forces even in the presence of external disturbances and modeling errors, i.e., combined robustness and compliance is obtained. In this paper, we validate and demonstrate the performance of this method through experiments where the proposed generalized super-twisting impedance control law is implemented on a Franka Emika robot manipulator. Three different scenarios, considering contact with different types of objects, under a constant external disturbance, are tested.
|
|
15:54-15:57, Paper TuBT2.19 | |
A Method to Detect the Sudden Stopping in an Assistive Robot for the Visually Impaired People |
|
Li, Jinyang | Politecnico Di Milano |
Corno, Matteo | Politecnico Di Milano |
Farina, Marcello | Politecnico Di Milano |
Keywords: Assistive and Rehabilitation Robotics, Estimation, Control Applications
Abstract: To enhance the autonomy of assistive robots for people with visual impairments, the Blind-assistive aUtonomous Droid Device (BUDD-e) program has been launched to emulate the role of a guide dog. BUDD-e consists of two main components: YAPE, an autonomous vehicle for navigation, and a novel smart tether system for safe user guidance. The study focuses on a method to detect when the user suddenly stops in order to avoid excessive tugging. The proposed method uses a Time-Domain Signal-Threshold-Based method, which involves selecting appropriate signals and threshold values to differentiate between normal movement and stopping events. Experiments show the tether force signal plays the most crucial role in accurately identifying stops and we study the trade-off involved in selecting this force threshold.
|
|
15:57-16:00, Paper TuBT2.20 | |
Enabling Tactical Pursuit-Evasion Game Strategies Via Shaping Task Regulation with Coverage Control |
|
Hicks, Gregory | JHUAPL |
Xu, Xiaotian | University of Maryland, College Park |
Diaz-Mercado, Yancy | University of Maryland |
Keywords: Multi-agent and Networked Systems, Path Planning and Motion Control, Robotics
Abstract: Multi-agent pursuit-evasion games are complex. A practical approach to managing this complexity is through policy that seeks to achieve higher-level game-state objectives via localized tactical tasking of teams of agents working collaboratively. Common tactical tasks are field shaping tasks through which the pursuers exercise influence over the evader in order to alter the state of the game. In this paper we develop the mathematics of field shaping, demonstrate how shaping tasks are defined, and indicate how task performance may be robustly regulated using coverage control. A formal general framework for the conceptual description of shaping tasks is provided. The results are validated in simulation of an example tactical task which regulates pursuers to get in between the evader and its target for different pursuer team sizes.
|
|
TuBT3 |
Prime 1 |
Modeling, Estimation, and Control of Energy Systems |
Invited Session |
Chair: Dey, Satadru | The Pennsylvania State University |
Co-Chair: de Castro, Ricardo | University of California, Merced |
|
15:00-15:15, Paper TuBT3.1 | |
A Study on Control Co-Design for Optimizing Microgrid Sustainability (I) |
|
Jahan, Tania Rifat | Texas Tech University |
Ouedraogo, Asmaou | Texas Tech University |
Docimo, Donald | Texas Tech University |
Keywords: Power and Energy Systems, Control of Smart Buildings and Microgrids, Control Applications
Abstract: This paper studies the optimization of microgrid plant and controller features to reduce environmental impacts. The configuration and control of grid technology is critical for storing and supplying energy. While these systems are traditionally designed for maximizing efficiency and frequency regulation, there is a shift towards minimizing grid environmental footprints. This work presents a framework to enable sustainability-centric microgrid design, with two main features. The first is the inclusion of control co-design (CCD), which expands the design space and potential capabilities of the microgrid. The second is the introduction of sustainability-centric objective functions, categorized into environmental impact from microgrid component manufacturing, operation, and disposal. After introducing the candidate microgrid’s model, controller, and CCD framework, the system is optimized to support a data center during a blackout. The relationship between the sustainability objective functions and the plant and controller design variables are explored. Pareto fronts are identified and studied, providing a comparison of the influence of each sustainability category on environmental impact.
|
|
15:15-15:30, Paper TuBT3.2 | |
Dynamical Modeling of Battery Life-Cycle Ecosystem (I) |
|
Bhaskar, Kiran | The Pennsylvania State University |
Vyas, Shashank Dhananjay | The Pennsylvania State University |
Padisala, Shanthan Kumar | The Pennsylvania State University |
Rahn, Christopher D. | Penn State Univ |
Hu, Xianbiao | Penn State University |
Dey, Satadru | The Pennsylvania State University |
Keywords: Power and Energy Systems, Modeling and Validation, Control Applications
Abstract: The growing popularity of electric vehicles (EVs) results to an enormous number of Li-ion battery packs in the transportation market. However, the EV battery packs retire typically around 80% of their nominal capacity owing to the high performance and EV range standards. These retired cells still have high potential in them before really receding the energy storage market. Therefore, instead of recycling or discarding these battery packs, their usage in the second-life market, for less demanding applications such as stationary energy storage applications, promotes sustainability and reduces the battery cost for such applications. In order to ensure the efficiency of such sustainable approach, it is crucial to understand the battery ecosystem that supports this operation. However, existing literature significantly lacks a systematic understanding of such an ecosystem. In this work, we explore a dynamical modeling approach for battery life-cycle ecosystem which encompasses the coupled nature of first-life and second-life usages along with recycling. Multiple representative scenarios are simulated to show the model performance and its applicability in optimizing the control decision to meet the second-life demand and minimize wastage.
|
|
15:30-15:45, Paper TuBT3.3 | |
MPC-Based Real-Time Energy Management of Freight Hybrid Locomotives (I) |
|
Ahuja, Nitisha | The Pennsylvania State University |
Bhaskar, Kiran | The Pennsylvania State University |
Martin, Jay | Pennsylvania State University |
Rahn, Christopher D. | Penn State Univ |
Pangborn, Herschel | The Pennsylvania State University |
Keywords: Power and Energy Systems, Automotive Systems, Control Applications
Abstract: To reduce greenhouse gas (GHG) emissions, freight railways are integrating battery-electric locomotives (BELs) in conventional trains, which are primarily powered by diesel-electric locomotives (DELs). The energy management of these hybrid trains involves optimizing power allocation between BELs and DELs based on route power requirements. Since the average fleet age of conventional freight trains is over 20 years, energy management strategies should include a trade-off between battery degradation and fuel savings. In addition, since DELs operate at finite notch point settings, which cannot practically be changed at high frequency during operation, a supervisory controller should be able to account for rate limitations on this discrete control input. Incorporating the above objectives and constraints, we propose using mixed-integer model predictive control (MPC) for real-time energy management. Using a realistic freight train route, the proposed controller showed 33% fuel savings as compared to conventional DEL-only trains, and a battery capacity loss of 23% was noted for 10 years of operation. In contrast, the idealized fuel savings for the given hybrid train configuration is 40% as found by non-causal dynamic programming, and capacity loss is approximately the same.
|
|
15:45-16:00, Paper TuBT3.4 | |
Novel Tour Construction Heuristic for Pick-Up and Delivery Routing Problems (I) |
|
Goutham, Mithun | The Ohio State University |
Stockar, Stephanie | The Ohio State University |
Keywords: Estimation, Transportation Systems, Path Planning and Motion Control
Abstract: In logistic applications that require the pickup and delivery of items, route optimization problems can be modeled as precedence constrained traveling salesperson problems. The combinatorial nature of this problem restricts the application of exact algorithms to small instances, and heuristics are generally preferred due to their tractability. However, due to precedence constraints that restrict the order in which locations can be visited, heuristics outside of the nearest neighbor algorithm have been neglected in literature. This paper presents an adapted convex hull cheapest insertion heuristic that accounts for precedence constraints and compares its solutions with the nearest neighbor heuristic using the TSPLIB benchmark dataset. The performance of the proposed heuristic is observed to be strongly dependent on the spatial characteristics of the precedence constraints, and while it is particularly suited to cases where pickups are located in the periphery and deliveries are centrally located, it does not perform as well in the case where delivery nodes are located in the periphery.
|
|
16:00-16:15, Paper TuBT3.5 | |
Safe Control of Reconfigurable Batteries (I) |
|
Ebrahimi, Iman | University of California, Merced |
de Castro, Ricardo | University of California, Merced |
Keywords: Control Applications, Optimal Control
Abstract: In this paper, we introduce a novel control algorithm for managing modular reconfigurable batteries. Our approach combines Control Allocation (CA) and Control Barrier Functions (CBF) to optimize operational efficiency while enforcing electro-thermal safety constraints. More specifically, this approach is designed to regulate the output voltage and equalize State-of-Charge (SoC) across battery modules as a means to alleviate the weakest-module problem. It also modulates the power load applied to each module to minimize thermal runaway risks. Numerical results reveal that our approach outperforms conventional Model Predictive Control (MPC) methods, reducing the average computational time by up to 81%, while providing similar safety guarantees.
|
|
16:15-16:30, Paper TuBT3.6 | |
Graceful Safety Control: Motivation and Application to Battery Thermal Runaway |
|
Moon, Yejin | University of Maryland |
Fathy, Hosam K. | University of Maryland |
Keywords: Control Applications, Control Design, Power and Energy Systems
Abstract: This paper surveys the literature on barrier function-based safety control and identifies a critical research gap, namely, the need for safety control algorithms that degrade gracefully in the presence of system failures and anomalies. The paper proposes a novel control design paradigm that embeds the notion of graceful degradation within control barrier function theory. Intuitively, the idea is to switch from one safety control mode to another in response to changes in the system's environment and/or damage/hazard state. Mathematically, we achieve this by constructing a non-monotonic barrier function, then utilizing slack variables to "soften" the resulting safety constraints. We illustrate this approach for a simple model of the thermal runaway dynamics of two cells in a battery pack. When one battery cell experiences thermal runaway, a benchmark safety controller focuses its cooling efforts on that hot cell, even when doing so is futile. The proposed controller, in contrast, focuses on preventing thermal runaway propagation instead. The end result is quite profound: both battery cells burn down in the benchmark case, but only one cell burns down with the proposed graceful safety controller.
|
|
TuBT4 |
Prime 2 |
Advanced Driver Assistance Systems and Vehicle Automation |
Invited Session |
Chair: Peters, Diane | Kettering University |
Co-Chair: Zhao, Junfeng | Arizona State University |
|
15:00-15:15, Paper TuBT4.1 | |
A Transitional Intelligent Driver Model Enabling Vehicle Longitudinal Motion Prediction in Lane-Change Maneuvers (I) |
|
Demir, Sude Ela | Univeristy of Texas at Austin |
Zhou, Xingyu | University of Texas at Austin |
Zhang, Yanze | The University of North Carolina at Charlotte |
Luo, Wenhao | University of North Carolina at Charlotte |
Wang, Junmin | University of Texas at Austin |
Keywords: Automotive Systems, Modeling and Validation, Path Planning and Motion Control
Abstract: This paper proposes an extension of the Intelligent Driver Model (IDM) for predicting the vehicle longitudinal motion intentions during lane-changing maneuvers. The extension systematically creates a single, modified IDM model that can predict the ego vehicle’s longitudinal motion with respect to both leading vehicles in the current and target lanes before, during, and after the lane change. A dynamic weighting function, determined by the ego vehicle’s lateral displacement throughout the lane-changing maneuver, assigns relative importance to each of the two leading vehicles during integration. Several candidates for the dynamic weighting function are suggested in this context. Using a high-fidelity, moving-base driving simulation system, a human-in-the-loop pilot study was carried out, specifically recording ego vehicle motion data during lane changes. This data was utilized to calibrate the transitional IDM, demonstrating its efficacy in predicting the ego vehicle’s longitudinal motion during lane changes. Furthermore, we compared the performance of various transitional functions and identified the hyperbolic tangent function as the most effective choice.
|
|
15:15-15:30, Paper TuBT4.2 | |
Development and Control of an Autonomous RC Racing Car (I) |
|
Mekky, Ahmed | MathWorks |
Jain, Ronak | Kettering University |
Bandreddi, Shivani | Kettering University |
Peters, Diane | Kettering University |
Keywords: Automotive Systems, Control Applications, Control Design
Abstract: The primary objective of this project is to develop, and design control systems for, an autonomous RC vehicle such that it completes a marked track as fast as possible. Upon the development of the hardware and the necessary software, a mathematical model for the vehicle was identified and validated via real-time experimentation. Utilizing the identified model, three control algorithms were developed for the control of the lateral motion of the vehicle, namely, PID, MPC, and LQR, and are evaluated for their performance and the time taken to complete one lap on a marked track. All the necessary software is running on an onboard NVIDIA Jetson Nano, except for the commanded longitudinal velocity; which is received from a host computer through an established ROS network
|
|
15:30-15:45, Paper TuBT4.3 | |
Simultaneous Localization and Mapping with Road Markings Identified from LiDAR Intensity (I) |
|
Wang, Shun-Yen | Arizona State University |
Meng, Jingxiong | Arizona State University |
Wishart, Jeffrey | Science Foundation Arizona/Arizona Commerce Authority |
Zhao, Junfeng | Arizona State University |
Keywords: Intelligent Autonomous Vehicles, Robotics, Sensors and Actuators
Abstract: Simultaneous Localization and Mapping (SLAM) algorithms play a crucial role in automated vehicles. These vehicles utilize Light Detection and Ranging (LiDAR), Radio Detection and Ranging (RADAR) and camera sensors to perceive their surroundings and use the onboard sensor data to determine their location. High Definition (HD) maps enable automated vehicles to precisely pinpoint their location and navigate with lane-level accuracy. Additionally, HD maps provide information such as traffic light locations, lane placement, crosswalks, and more. However, manually labeling objects is a time-consuming process. Existing LiDAR SLAM algorithms struggle to visualize lane markings and road boundaries, making it difficult to accurately label lanes. Cameras have been widely used for lane marking detection, but they are sensitive to weather conditions and environmental lighting. To address these challenges, we enhance the existing Hdl-Graph-Slam by incorporating LiDAR intensity data into the point cloud map. To achieve this, the RANSAC method has been introduced to execute ground plane extraction. Next, a road marking detector based on the Otsu thresholding method is employed, which separates LIDAR point clouds into two segments: the road surface and road markings. The Otsu thresholding method has been modified for better accuracy, enabling us to determine the vehicle's pose and enhancing the efficiency of the algorithm. The results were compared to the google satellite map and the gps data. As a result of using the modified SLAM algorithm, the road markings are visualized and achieve centimeter-level accuracy. These enhancements contribute to the robustness and accuracy of automated vehicles, particularly in scenarios involving lane detection and road boundary recognition.
|
|
15:45-16:00, Paper TuBT4.4 | |
Starlink for Localization: A Low Earth Orbit Satellites Based Approach for Vehicle Localization (I) |
|
Meng, Jingxiong | Arizona State University |
Chen, Yan | Arizona State University |
Zhao, Junfeng | Arizona State University |
Keywords: Intelligent Autonomous Vehicles, Automotive Systems, Estimation
Abstract: As the Low Earth Orbit (LEO) satellite constellation becomes more mature, LEO for vehicle localization could be a supplement or a replacement in challenging situations of traditional Global navigation satellite systems (GNSS). This paper designs a dedicated LEO satellite simulator for vehicle localization, which also integrates with IMU for more precise localization. The simulator includes satellite trajectory generation, observable satellite identification, vehicle localization solver, and an Extended Kalman Filter (EKF) sensor fusion strategy for more precise localization. The LEO simulator was also seamlessly integrated with the Carla simulator for vehicle localization assessments. Comprehensive tests were conducted with the simulation tools to evaluate LEO satellite-based vehicle localization performance across different satellite counts. Furthermore, the capability and reliability of the developed LEO simulator are verified by the experimental data. Both simulation and experiment results indicate the potential of LEO satellites for precise and reliable vehicle localization for automated driving. Additionally, the designed simulator offers a versatile platform that is flexible to integrate with other satellite simulators or other sensors’ measurements, like LiDAR and camera, for precise localization or other usages.
|
|
16:00-16:15, Paper TuBT4.5 | |
Is It Necessary to Calibrate All Parameters for Each Driver? (I) |
|
Wang, Yanbing | Argonne National Lab |
de Souza, Felipe | Argonne National Lab |
Han, Jihun | Argonne National Laboratory |
Karbowski, Dominik | Argonne National Laboratory |
Keywords: Estimation, Modeling and Validation
Abstract: Learning the variability of driver behavior can help with understanding the driver heterogeneity and traffic patterns, designing driver-specific vehicle automation and control, and improving the accuracy of micro-simulation tools. However, even understanding the driving behavior of a small population can be challenging, due to the large number of total parameters that need to be calibrated. This study investigates whether individually calibrating driver models provides a more accurate description of the population or if it is inherently "overparameterized". We propose an approach to analyze calibration performance using various reduced-order versions of a car-following model based on the Optimal Velocity model (OVM). We explore which parameters can be eliminated from calibration without sacrificing overall accuracy and how reducing the number of calibrated parameters impacts the representation of driver heterogeneity. Our preliminary results indicate that while reduced-parameter models compromise accuracy, the extent varies depending on which parameters are fixed first. Furthermore, fixing one parameter alters the distribution of other parameters, suggesting possible dependencies among these model parameters.
|
|
16:15-16:30, Paper TuBT4.6 | |
Validation of Energy Saving from Cooperative Driving Automation Via Vehicle-In-The-Loop Tests (I) |
|
Hyeon, Eunjeong | Argonne National Laboratory |
Russo, Miriam | Argonne National Laboratory |
Zhan, Lu | Argonne National Laboratory |
Jeong, Jongryeol | Argonne National Laboratory |
Kim, Namdoo | Argonne National Laboratory |
Han, Jihun | Argonne National Laboratory |
Misra, Priyash | Argonne National Laboratory |
Stutenberg, Kevin | Argonne National Laboratory |
Karbowski, Dominik | Argonne National Laboratory |
Keywords: Path Planning and Motion Control, Optimal Control, Transportation Systems
Abstract: This paper presents an experimental validation of energy savings achieved through cooperative driving automation (CDA) via vehicle-in-the-loop (VIL) testing in car-following scenarios. The impacts of different CDA classes—from status-sharing to prescriptive—on vehicle energy efficiency are explored. In the experiments, a plug-in hybrid electric vehicle runs on a chassis dynamometer, integrated with simulation software rendering a virtual environment. Results indicate that when agreement-seeking cooperation operates even with a minimal number of vehicles, energy can be saved by up to 5% over human driving. Our findings highlight the considerable promise of CDA technologies for enhancing energy efficiency, especially fostering research on agreement-seeking cooperation.
|
|
TuBT7 |
Prime 3 |
A Tutorial on Easy Learning, Adaptation in Feedback Systems |
Tutorial Session |
Chair: Abramovitch, Daniel Y. | Agilent Technologies |
Co-Chair: Leang, Kam K. | University of Utah |
Organizer: Abramovitch, Daniel Y. | Agilent Technologies |
Organizer: Leang, Kam K. | University of Utah |
|
15:00-15:15, Paper TuBT7.1 | |
A Tutorial on Easy Learning and Adaptation in Feedback Systems (I) |
|
Abramovitch, Daniel Y. | Agilent Technologies |
Widrow, Bernard | Stanford University |
|
15:15-15:30, Paper TuBT7.2 | |
Estimation and Control: Practical Application of Machine Learning for Feedback Systems (I) |
|
M Anderson, Jacob | University of Utah |
McKee, Sasha M. | University of Utah |
Leang, Kam K. | University of Utah |
| |