This work builds upon predictive control strategies for optimal charging of lithium batteries under thermal constraints by extending the problem to data-enabled predictive control (DeePC), which uses data-driven Hankel matrices generated offline for real-time prediction in lieu of an explicit model. Moreover, the robust framework for DeePC is considered for the case of noisy Hankel matrix data, and the nonlinear robustness constraint is tightened to a linear constraint, ensuring compatibility with standard quadratic programming (QP) solvers. Simulation results demonstrate that the proposed DeePC framework successfully navigates the trade-off between fast charging and thermal safety, achieving convergence within 1% of the reference SOC after 500 seconds under specific tuning while maintaining core temperatures in a safe vicinity. Under conservative tunings, the controller exhibits a charging rate of 0.816 C, remaining within the recommended safe range to mitigate early-life plating. Hence, the DeePC approach as a viable control framework for generating charging strategies is validated.

This paper presents an approach to model driver lane change behavior using maximum entropy inverse reinforcement learning (MaxEnt IRL). The proposed method aims to generate personalized lane change trajectories that reflect individual driving preferences while ensuring user privacy. To achieve this, driving data are collected from individual driver and used to train the model, while scale-based transformation is utilized to encrypt the data during cloud-based training. Bezier curves are employed to parameterize the lane change trajectories due to their ability to produce smooth, continuous paths. The MaxEnt IRL algorithm is then used to infer a reward function that represents each driver's preferences by learning optimal weights for a cost function that describes the lane change behavior. The proposed approach is tested over several real-world data to demonstrate its effectiveness in capturing personal driving styles under various conditions. The proposed trajectory encryption approach is compared to a benchmark differential privacy mechanism, and the results clearly show that the proposed method outperforms. Experimental results demonstrate that the proposed method can generate lane change paths that closely align with the behavior exhibited by individual driver, offering an approach for personalization in autonomous driving.

We investigate deterministic and nonblocking supervisory control of discrete event systems under cyber-attacks using the ALTER (Attack Language for Transition-basEd Replacement) model. While prior works consider supervisory control that achieves either the large (upper bound) language or small (lower bound) language separately, deterministic supervisory control achieves both large language and small language at the same time to ensure that the language generated by the supervised system is unique and deterministic. We introduce two new concepts of CA-D-controllability and CA-D-observability and prove that they are necessary and sufficient for the existence of a deterministic supervisor. For nonblocking supervisory control, the objective is to ensure that the supervised system can always reach marked states under any attack scenario. We prove that relative closure, CA-D-controllability, and CA-D-observability together are necessary and sufficient for the existence of a nonblocking supervisor. We further develop methods to verify CA-D-controllability and CA-D-observability. We also illustrate our results using a robotic system example.

This work investigates the use of dense extended Kalman filter (DEKF) to simultaneously estimate both the state-of-charge (SOC) and equivalent-circuit parameters of serial-connected, lithium-ion battery cells under limited voltage measurement. Stability analyses are conducted to certify the reliability of the DEKF for simultaneous state-parameter estimation under certain mild assumptions. To begin, the local asymptotic stability of the expected dense error in the presence of nonlinearity is verified by selection of an appropriate Lyapunov function. Then the expected sparse error is shown to be marginally stable in the absence of nonlinearities by proving that the spectral radius of the sparse expected error dynamics is exactly equal to 1. Finally, a sufficient condition for sparse stability in the presence of nonlinearity is obtained via selection of an appropriate Lyapunov function. Simulation trials under several different scenarios are executed, confirming that the associated error dynamics are indeed stable for the duration of the simulation time.

This paper proposes a reinforcement learning (RL)-based control framework for real-time state-of-charge (SOC) balancing in reconfigurable lithium-ion battery packs. Traditional battery management strategies, such as passive and active balancing, suffer from energy inefficiency and limited adaptability. Reconfigurable battery systems offer enhanced flexibility through dynamic switchable topologies. However, to fully exploit the potential of the battery pack, an intelligent control policy to determine the optimal battery configuration in real-time is needed, which is a difficult task due to the large optimization space. To address this challenge, the reconfiguration problem is formulated as a Markov Decision Process, and a policy optimization framework is proposed to guide switch decisions based on observed cell states, such as the SOC of the individual cells in addition to the current pack configuration. The proposed RL agent operates on discrete action space and is guided by a multi-objective reward function tailored for battery management application. To avoid the high cost associated with physical battery training, we construct a linear surrogate model for efficient simulation-based learning. The proposed RL controller is evaluated against rule-based and greedy-based baselines, achieving superior SOC balancing performance with reduced switching frequency and improved runtime. In addition, comparison to true optimal obtained by dynamic programming demonstrates that the proposed RL controller achieves competitive results with significantly lower computational cost.

The key to realize the accurate modeling of photovoltaic (PV) systems depends on extracting high-precise unknown parameters of equivalent circuit models for PV cells. However, high-precise values for these unknown parameters are hard to achieve because of PV cells’ highly nonlinear dynamic characteristics. Whale optimization algorithm (WOA) gets regarded as one promising approach for PV cell models’ parameter extraction because of its good search performance. Nevertheless, WOA provides strong local exploitation but exhibits low exploration, which leads to premature convergence. To address this issue, this work puts forward an enhanced WOA named RLEWOA. On the one hand, an improved ranking-based differential mutation operator is designed for enhancing the exploration capability. This operator can select random yet superior individuals to steer the movement of target individuals in the promising direction. On the other hand, reinforcement learning is integrated to help the target individuals select more suitable position update strategies at different stages for better balancing exploitation and the exploration. The resultant approach is implemented for nine cases and analyzed in comparison with other algorithms. Simulation results show that RLEWOA significantly outperforms other algorithms and demonstrates very high accuracy for extracting PV cell models’ parameters.

Autonomous parking remains a challenging task due to the need for accurate trajectory tracking, smooth steering, and stable heading control under diverse manoeuvring conditions. Conventional model predictive control (MPC) can handle system constraints effectively, but its performance depends heavily on manually tuned cost weights. This paper proposes a reinforcement learning-assisted MPC (RL-assisted MPC) framework to improve autonomous vehicle parking performance. A deep Q-network (DQN) agent is trained to dynamically select the cost function weights of an MPC controller, enabling real-time adaptation based on the vehicle's current state. The hybrid approach leverages the predictive optimisation capability of MPC together with the adaptive decision-making of RL, enabling the controller to adjust trade-offs in real time without manual re-tuning. The framework is evaluated across five different parking scenarios and compared against static-weight MPC baselines. Experimental evaluations demonstrate that the proposed RL-assisted MPC framework achieves comparable or better lateral tracking accuracy, while consistently providing smoother steering behaviour and improved heading stability compared with baseline controllers using static MPC weights. The proposed framework is further evaluated under randomly selected and previously unseen initial vehicle positions, demonstrating its robustness and generalisation across diverse parking configurations. The results demonstrate that RL-assisted MPC improves robustness and generalisation in automated parking systems, highlighting the potential of combining model-based predictive control with RL for autonomous driving.

Abstract This paper presents an event-triggered model predictive control (ET-MPC) strategy for a DC–DC buck converter to reduce computational burden while maintaining high dynamic performance. A four-mode discrete-time model is established to accurately capture the converter switching behavior, and a Kalman filter is integrated into the control framework to estimate load disturbances and improve control accuracy. Unlike conventional time-triggered MPC (TT-MPC), which solves an optimization problem at every time-step, the proposed ET-MPC evaluates the optimal switching sequence only when a voltage deviation exceeds a predefined threshold. This mechanism significantly reduces the number of online optimizations while preserving the regulation of output voltage. Simulation results demonstrate that ET-MPC achieves up to 94% reduction in computational effort with comparable transient response, low steady-state error, and acceptable switching frequency. The proposed controller enables an efficient real-time implementation of model predictive control (MPC) for power converters under varying operating conditions.

Iterative closest point (ICP) and its variants rep resent a fundamental technique for rigid registration between two point sets, with wide applications ranging from robotics to 3D reconstruction. The primary limitations of ICP include slow convergence due to its linear convergence rate, along with the per-iteration nearest-neighbor search, which scales linearly with source cloud size and dominates runtime on dense datasets. To address these issues, this paper proposes a modified ICP method that achieves fast convergence while maintaining robust alignment. Specifically, we introduce an adaptive threshold-based downsampling technique that selectively reduces the number of source points processed during early iterations, decreasing the computational cost for each iteration. This is followed by a full cloud refinement stage designed to recover alignment accuracy and mitigate any degradation introduced by the downsampling phase. Simulation results demonstrate that the proposed method achieves notable reductions in computational time across the majority of benchmark datasets, including Stanford mesh models and LiDAR point clouds, while maintaining comparable registration accuracy relative to the baseline method.

This article presents a novel graph neural network (GNN)-based surrogate model (GNN-SM) for predicting state evolution in reconfigurable battery packs. By leveraging graph-based representations of battery cell interconnections, the proposed approach addresses the unique challenge of estimating the imbalance in state-of-charge (SOC) and temperature of cells of a battery pack in dynamic battery configurations. Unlike conventional methods that focus on instantaneous state estimation, our GNN-SM predicts future SOC and temperature distributions by considering both current system state and switch configuration. The model architecture combines graph attention networks with pooling operations to effectively capture cell-to-cell interactions and battery pack-level dynamics. Numerical results under constant current and constant power discharge conditions demonstrate that GNN-SMsignificantly outperforms baseline feedforward neural network (FNN) and FNN-attention models, achieving up to 73.7% reduction in root-mean-square error for temperature imbalance prediction and 46% reduction for SOC imbalance prediction. Furthermore, the model provides a 1629-fold speedup over high fidelity physics-based simulator while maintaining mean absolute percentage errors below 2% for temperature and 8% for SOC predictions. The scalability of GNN-SM is further validated on a 100-cell reconfigurable battery pack, where the proposed approach achieves high accuracy despite being trained on an extremely small fraction of all possible configurations. Finally, robustness analysis under sensor noise conditions demonstrates that GNN-SM maintains reliable predictions even under high noise levels.

This paper presents a control framework based on reinforcement learning (RL) for achieving real-time state-ofcharge (SOC) balancing within reconfigurable lithium-ion battery systems. Reconfigurable battery architectures provide enhanced flexibility via dynamically adjustable circuit topologies. Realizing their full capabilities requires an intelligent controller that can determine optimal topologies in real-time. This work addresses the problem by modeling the reconfiguration task as a Markov Decision Process and developing a policy optimization approach that directs switching actions according to cell states. The proposed RL agent operates on discrete action space and is guided by a multi-objective reward function. In particular, the controller is designed to balance the tradeoff among SOC equalization performance, switching effort, and runtime. To avoid the high cost associated with physical battery training, a surrogate model is proposed for efficient simulation-based learning. The surrogate model provides an accurate approximation of the battery pack dynamics while significantly reducing the computational burden during training. The proposed RL controller is evaluated against rule-based and greedy-based baselines, achieving superior SOC balancing performance with improved runtime. Specifically, the proposed controller achieves a final SOC spread of 0.0163 V, corresponding to more than 17% improvement over the greedybased controller and 58% improvement over the rule-based controller.

This paper investigates the problem of information control in networked multi-user systems, where agents such as robots, sensors, and software entities interact via a communication network to achieve individual or shared goals. Information control involves deciding which state estimates to share or broadcast, balancing cooperation among friends and privacy from adversaries. Since each user has only partial knowledge of the system, efficient protocols for sharing relevant data to balance privacy, security, and transparency is needed. This study models multi-user systems as discrete-event systems where agents need to distinguish certain state pairs in order to perform their tasks. We systematically study and solve critical problems to address the key aspects of information control: determining the necessity of shared information, minimizing communication for security, and maximizing public information release when required. A framework that addresses private communications, public broadcasting, and adversarial dynamics, offering strategies to meet both security and transparency requirements is introduced. Solutions and algorithms are proposed to solve these problems.

This paper presents a personalized lane change control framework that combines the maximum entropy inverse reinforcement learning (MaxEnt IRL) with model predictive control (MPC). Instead of manually tuning the MPC cost weight, the proposed method learns a cost function from expert driving demonstrations using interpretable trajectory features. The learned weights are incorporated into the MPC formulation to generate personalized lane change trajectories that mimic the expert driving style. Simulation results in CARLA show that the IRL-trained MPC controller can reproduce distinct driving styles and matches expert behaviors in heading, lateral motion, and steering. Furthermore, real world on-road testing on an ISUZU truck demonstrated that the proposed controller generalizes well to unseen initial conditions. Overall, the proposed approach reduces manual calibration efforts, eliminates the need for explicit path planning, improves controller interpretability, and enables personalized driver behavior replication.

Autonomous driving systems deployed in the real world suffer from excessive hardware redundancy and prohibitive deployment barriers. To address these challenges, LightAD is proposed as a lightweight, vision-based autonomous driving system framework tailored for simple, low-speed environments such as confined work zones and low-speed logistics. Diverging from industrial-grade solutions that rely on costly proprietary sensors and high-performance computing, LightAD implements system-level simplification. By leveraging consumer-grade hardware and lightweight algorithms for perception, planning, and control, LightAD significantly reduces costs and integration complexity while satisfying necessary autonomous driving functionality. Field trials at the Intelligent Ground Vehicle Competition on a physical vehicle confirmed the feasibility of LightAD and revealed practical implementation lessons. These results indicate that by combining low-cost general-purpose hardware with lightweight algorithms, LightAD effectively supports fundamental autonomous driving functionality in simple, well-defined, and low-speed environments, offering a feasible engineering pathway for cost-effective deployment of autonomous driving.

This paper presents a Gain-Scheduled Data-Enabled Predictive Control (DeePC) framework for a cell-level battery thermal management system. The system explicitly captures the electro-thermal coupling of individual cells and their heat exchange with the coolant, and divides the operation into multiple zones. To deal with the nonlinearity that is commonly present in battery systems, zone-specific Hankel matrices are constructed from simulation-generated data, allowing the DeePC controllers to adaptively schedule control gains according to the current thermal condition. Comprehensive simulation results demonstrate that the proposed approaches can effectively track target temperature and improves temperature uniformity across cells without an explicit system model.

Model predictive control (MPC) is widely used for autonomous vehicle path tracking due to its ability to handle system constraints and optimize performance over a prediction horizon. However, frequent online optimization imposes high computational demands, making the real-time implementation of MPC challenging. Event-triggered MPC aims to solve this issue by updating control actions only when a predefined condition is met, but it executes precomputed control sequences in an open-loop fashion between events, potentially allowing errors to accumulate. This letter proposes an event-triggered MPC framework integrated with a linear inter-event control mechanism to address this limitation. The proposed inter-event controller applies a least-squares-based linear model to generate control inputs in real-time during inter-event periods, enabling continuous feedback corrections. Experimental results on a Quanser QCar2 platform demonstrate that the proposed approach improves tracking accuracy by 10% while significantly reducing the number of MPC optimizations compared to standard event-triggered MPC, offering an efficient solution for real-time path tracking problem.

For large interconnected systems, centralized model predictive control (MPC) may be computationally intractable while distributed MPC may not achieve global optimum. To fill this gap, a new MPC framework, named asynchronous MPC, has been proposed in literature, where only a subset of subsystems is selected for optimization at each time step. For the remaining subsystems, the previously optimized control sequence is shifted and reused. Such an approach can balance computational load and control performance, and can be viewed as a framework in between centralized MPC and distributed MPC. However, prior work assumes each subsystem to be deterministic and does not consider the impacts of external disturbance. Moreover, feasibility and stability analysis for asynchronous MPC has not been studied. This paper addresses these limitations by (i) incorporating the prediction error introduced by external disturbance into reconfiguration policy, (ii) providing a stability condition for the closed-loop system, and (iii) establishing the recursive feasibility. Numerical results using battery cell-to-cell balancing control confirms the stability of the closed-loop system and demonstrates over 86% computation reduction.

This paper presents a practical experiment for estimating the state-of-charge (SOC) of individual cells in a series-connected heterogeneous lithium-ion battery pack, where only the terminal voltage of the battery pack is measured. To deal with real-time computation constraints, the dense extended Kalman filter (DEKF) algorithm has been proposed in the literature, which has a significantly lower computational complexity compared to the regular extended Kalman filter for this specific estimation problem. This work supplements the existing work by conducting a real-world experiment to validate the performance of the DEKF. Specifically, experiments involving a battery pack of three cells connected in series were conducted, where the battery pack was discharged under a constant current load. A genetic algorithm was applied to identify missing model parameters, as well as tuning the DEKF for optimal convergence and accurate SOC estimation. Our experimental results confirm that the proposed DEKF accurately estimates the SOC of each cell regardless of the hardware limitations and uncertainty, making it suitable for low-cost, real-time battery management systems. In particular, the SOC estimation error can be kept well under 1% even if the initial estimate is far from the true SOC.

This paper proposes a reinforcement learning-assisted model predictive control (RL-assisted MPC) framework to improve autonomous vehicle parking performance. A Deep Q-Network (DQN) agent is trained to dynamically select the cost function weights of an MPC controller, enabling real-time adaptation based on the vehicle’s current state. The hybrid framework leverages the predictive optimization capabilities of MPC and the adaptive decision-making strength of RL. Experimental evaluations demonstrate that the proposed RL-assisted MPC framework achieves comparable lateral tracking accuracy, while consistently providing smoother steering behavior and improved heading stability compared to baseline controllers using static MPC weights. The results highlight the potential of integrating RL with model-based control for enhancing robustness and adaptability in automated parking systems.