Jingwei Lu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Jingwei Lu has a total of 44 co-authors. Here are the top co-authors by number of shared publications:
- Fei-Yue Wang
Institute of Automation, Chinese Academy of Sciences
7 shared publications - Qinglai Wei
Chinese Academy of Sciences
6 shared publications - Lefei Li
Department of Industrial Engineering, Tsinghua University, Beijing, China
4 shared publications - Xiang Cheng
Peking University
3 shared publications - Xingxia Wang
State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China
3 shared publications Shuliang Liu
Macau Institute of System Engineering, Macau University of Science and Technology, Macau, China
2 shared publications
Latest Publications
- 17 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (3) , 2941-2954
GLIDE: Generative LLM-Driven Data Augmentation for Meta-Services With Applications to E-Commerce Scenarios
In this article, a novel data augmentation method is developed for Meta-Services based on large language models (LLMs) and requirement-functional-logical-physical (RFLP) models, referred to as generative LLM-driven data augmentation (GLIDE), and an application of GLIDE to e-commerce recommendation systems is presented. First, the framework of GLIDE is designed under Meta-Services, including a generalized encoder and decoder architecture, as well as designers, which captures the idea of “designers in the loop” in Meta-Services. Furthermore, the generalized encoder and decoder architecture is established based on RFLP models in systems engineering, aiming to translate requirements and knowledge into prompts and obtain ideal LLM outputs under physical constraints. Subsequently, the application of the GLIDE method to e-commerce recommendation systems is given in detail, which is committed to enriching users’ behavioral data and providing an effective way to decouple data augmentation and recommendation modules. Finally, recommendation experiments are performed on real-world and synthetic behavior data, and the empirical results demonstrate the effectiveness of our GLIDE method.
- 17 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (3) , 2941-2954
GLIDE: Generative LLM-Driven Data Augmentation for Meta-Services With Applications to E-Commerce Scenarios
In this article, a novel data augmentation method is developed for Meta-Services based on large language models (LLMs) and requirement-functional-logical-physical (RFLP) models, referred to as generative LLM-driven data augmentation (GLIDE), and an application of GLIDE to e-commerce recommendation systems is presented. First, the framework of GLIDE is designed under Meta-Services, including a generalized encoder and decoder architecture, as well as designers, which captures the idea of “designers in the loop” in Meta-Services. Furthermore, the generalized encoder and decoder architecture is established based on RFLP models in systems engineering, aiming to translate requirements and knowledge into prompts and obtain ideal LLM outputs under physical constraints. Subsequently, the application of the GLIDE method to e-commerce recommendation systems is given in detail, which is committed to enriching users’ behavioral data and providing an effective way to decouple data augmentation and recommendation modules. Finally, recommendation experiments are performed on real-world and synthetic behavior data, and the empirical results demonstrate the effectiveness of our GLIDE method.
- 9 August 2025
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 179
Enhancing DRL-based lane change with rule-integrated sampling
- 9 August 2025
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 179
Enhancing DRL-based lane change with rule-integrated sampling
- 30 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automation Science and Engineering
- Vol. 22 (15455955) , 19317-19327
Learning-Based Parallel Control for Unknown Nonaffine Nonzero-Sum Games
In this paper, a novel nonzero-sum game (NSG) method is developed for completely unknown nonaffine nonlinear discrete-time (DT) systems, which is referred to as model-free NSG (MNSG). First, novel dynamic control laws are developed for NSGs using parallel control, namely introducing controls into feedback. Subsequently, an augmented N-player NSG is formulated according to the original N-player NSG to derive the dynamic control laws. Furthermore, we show that the control stabilities of the original and augmented N-player NSGs are equivalent. In the meantime, we prove that optimal control of the augmented N-player NSG is equivalent to near-optimal control of the original N-player NSG, and the Nash equilibrium of the original N-player NSG can be achieved. Then, a model-free learning scheme is developed to obtain the solution of the augmented N-player NSG using online policy iteration, and neither using a model network to predict unknown dynamics nor off-policy reinforcement learning (RL) is needed in the scheme. Lastly, numerical analysis, including the NSG of a DT system with unknown control-nonaffine dynamics and coupled controls, confirms the correctness of our MNSG method. The associated code is available at: https://github.com/lujingweihh/Adaptive-dynamic-programming-algorithms/tree/main/model_free_nonzero_sum_games_discrete_time.
- 30 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Automation Science and Engineering
- Vol. 22 (15455955) , 19317-19327
Learning-Based Parallel Control for Unknown Nonaffine Nonzero-Sum Games
In this paper, a novel nonzero-sum game (NSG) method is developed for completely unknown nonaffine nonlinear discrete-time (DT) systems, which is referred to as model-free NSG (MNSG). First, novel dynamic control laws are developed for NSGs using parallel control, namely introducing controls into feedback. Subsequently, an augmented N-player NSG is formulated according to the original N-player NSG to derive the dynamic control laws. Furthermore, we show that the control stabilities of the original and augmented N-player NSGs are equivalent. In the meantime, we prove that optimal control of the augmented N-player NSG is equivalent to near-optimal control of the original N-player NSG, and the Nash equilibrium of the original N-player NSG can be achieved. Then, a model-free learning scheme is developed to obtain the solution of the augmented N-player NSG using online policy iteration, and neither using a model network to predict unknown dynamics nor off-policy reinforcement learning (RL) is needed in the scheme. Lastly, numerical analysis, including the NSG of a DT system with unknown control-nonaffine dynamics and coupled controls, confirms the correctness of our MNSG method. The associated code is available at: https://github.com/lujingweihh/Adaptive-dynamic-programming-algorithms/tree/main/model_free_nonzero_sum_games_discrete_time.
- Open Access
Output Feedback Integrated Guidance and Control Design for Autonomous Underwater Vehicles Against Maneuvering Targets
Traditional guidance and control systems often treat guidance and control systems separately, leading to reduced interception accuracy and responsiveness, especially during high-speed terminal trajectories. These limitations are further exacerbated in autonomous underwater vehicles (AUVs) due to unknown wave/current disturbances, harsh underwater acoustic conditions, and limited sensor capabilities. To address these challenges, this paper studies an integrated guidance and control (IGC) design for AUVs intercepting maneuvering targets with unknown disturbances and unmeasurable system states. The IGC model is derived based on the relative motion equations between the AUV and the target, incorporating the lateral dynamics of the AUV. A model transformation is introduced to synthesize external disturbances with unmeasurable states, extending the resultant disturbance to a new system state. A finite-time convergent extended state observer (ESO) is thus designed for the transformed system to estimate the unknown signals. Using these estimates from the observer, a finite-time event-triggered sliding mode controller is developed, ensuring finite-time convergence of system errors to an adjustable residual set, as rigorously proven through Lyapunov stability analysis. Simulation results demonstrate the superiority of the proposed method in achieving higher interception accuracy and faster response compared to traditional guidance and control approaches with unknown disturbances and unmeasurable states.
Open AccessOutput Feedback Integrated Guidance and Control Design for Autonomous Underwater Vehicles Against Maneuvering Targets
Traditional guidance and control systems often treat guidance and control systems separately, leading to reduced interception accuracy and responsiveness, especially during high-speed terminal trajectories. These limitations are further exacerbated in autonomous underwater vehicles (AUVs) due to unknown wave/current disturbances, harsh underwater acoustic conditions, and limited sensor capabilities. To address these challenges, this paper studies an integrated guidance and control (IGC) design for AUVs intercepting maneuvering targets with unknown disturbances and unmeasurable system states. The IGC model is derived based on the relative motion equations between the AUV and the target, incorporating the lateral dynamics of the AUV. A model transformation is introduced to synthesize external disturbances with unmeasurable states, extending the resultant disturbance to a new system state. A finite-time convergent extended state observer (ESO) is thus designed for the transformed system to estimate the unknown signals. Using these estimates from the observer, a finite-time event-triggered sliding mode controller is developed, ensuring finite-time convergence of system errors to an adjustable residual set, as rigorously proven through Lyapunov stability analysis. Simulation results demonstrate the superiority of the proposed method in achieving higher interception accuracy and faster response compared to traditional guidance and control approaches with unknown disturbances and unmeasurable states.
- 22 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 55 (4) , 2884-2896
Parallel Control for Nonzero-Sum Games With Completely Unknown Nonlinear Dynamics via Reinforcement Learning
This article utilizes parallel control to investigate the problem of continuous-time (CT) nonzero-sum games (NZSGs) for completely unknown nonlinear systems via reinforcement learning (RL), and a parallel control-based NZSG (PNZSG) method is developed without reconstructing unknown dynamics or employing off-policy integral RL (IRL). First, novel dynamic control policies (DCPs) are developed for NZSGs by introducing controls into feedback, and an augmented system with augmented performance indices is constructed to derive the DCPs. Then, we theoretically analyze the effect of the DCPs on the control stability and performance indices, and the optimality of PNZSG is proven to be equivalent to the optimality of the original NZSGs. Subsequently, an IRL technique is employed to achieve the developed PNZSG method, and we show that no prior knowledge of the dynamics of NZSGs is needed to deploy the developed PNZSG method because of the augmented system and performance indices. Finally, numerical examples, including cooperative adaptive cruise control (CACC) of a vehicular platoon, demonstrate the correctness of the developed PNZSG method. The associated code is available at: https://github.com/lujingweihh/Adaptive-dynamic-programming-algorithms/tree/main/model_free_nonzero_sum_games.
- 22 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 55 (4) , 2884-2896
Parallel Control for Nonzero-Sum Games With Completely Unknown Nonlinear Dynamics via Reinforcement Learning
This article utilizes parallel control to investigate the problem of continuous-time (CT) nonzero-sum games (NZSGs) for completely unknown nonlinear systems via reinforcement learning (RL), and a parallel control-based NZSG (PNZSG) method is developed without reconstructing unknown dynamics or employing off-policy integral RL (IRL). First, novel dynamic control policies (DCPs) are developed for NZSGs by introducing controls into feedback, and an augmented system with augmented performance indices is constructed to derive the DCPs. Then, we theoretically analyze the effect of the DCPs on the control stability and performance indices, and the optimality of PNZSG is proven to be equivalent to the optimality of the original NZSGs. Subsequently, an IRL technique is employed to achieve the developed PNZSG method, and we show that no prior knowledge of the dynamics of NZSGs is needed to deploy the developed PNZSG method because of the augmented system and performance indices. Finally, numerical examples, including cooperative adaptive cruise control (CACC) of a vehicular platoon, demonstrate the correctness of the developed PNZSG method. The associated code is available at: https://github.com/lujingweihh/Adaptive-dynamic-programming-algorithms/tree/main/model_free_nonzero_sum_games.
- 12 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 11 (7) , 1702-1704
Deep Reinforcement Learning or Lyapunov Analysis? A Preliminary Comparative Study on Event-Triggered Optimal Control
Dear Editor, This letter develops a novel method to implement event-triggered optimal control (ETOC) for discrete-time nonlinear systems using parallel control and deep reinforcement learning (DRL), referred to as Deep-ETOC. The developed Deep-ETOC method introduces the communication cost into the performance index through parallel control, so that the developed method enables control systems to learn ETOC policies directly without triggering conditions. Then, dueling double deep Q-network (D3QN) is utilized to achieve our method. In simulations, we present a preliminary comparative study of DRL and Lyapunov analysis for ETOC.
- 12 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 11 (7) , 1702-1704
Deep Reinforcement Learning or Lyapunov Analysis? A Preliminary Comparative Study on Event-Triggered Optimal Control
Dear Editor, This letter develops a novel method to implement event-triggered optimal control (ETOC) for discrete-time nonlinear systems using parallel control and deep reinforcement learning (DRL), referred to as Deep-ETOC. The developed Deep-ETOC method introduces the communication cost into the performance index through parallel control, so that the developed method enables control systems to learn ETOC policies directly without triggering conditions. Then, dueling double deep Q-network (D3QN) is utilized to achieve our method. In simulations, we present a preliminary comparative study of DRL and Lyapunov analysis for ETOC.
- 4 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (3) , 4470-4479
Event-Triggered Parallel Control Using Deep Reinforcement Learning With Application to Comfortable Autonomous Driving
A novel event-triggered control (ETC) method, called deep event-triggered parallel control (deep-ETPC), is presented to achieve path tracking for comfortable autonomous driving (CAD) using parallel control and deep deterministic policy gradient (DDPG). Based on parallel control, the developed deep-ETPC method constructs a dynamic control policy by introducing variation rates of controls. By employing variation rates of controls, the developed deep-ETPC method is capable of indicating communication loss and comfortable driving indices in the reward, and then enables reinforcement learning (RL) agents to learn comfortable ETC driving policies directly. Moreover, the communication loss, which reflects ETC, is integrated into the reward, so there is no need to additionally design/train triggering conditions, which can be considered a type of multi-tasking learning. Furthermore, an EPTC-oriented DDPG algorithm is developed to achieve the developed deep-ETPC method, making DDPG applicable to ETC. Empirical results, including tracking a simple straight line trajectory and a complicated sinusoidal trajectory, demonstrate the effectiveness of the developed deepETPC method.
- 4 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (3) , 4470-4479
Event-Triggered Parallel Control Using Deep Reinforcement Learning With Application to Comfortable Autonomous Driving
A novel event-triggered control (ETC) method, called deep event-triggered parallel control (deep-ETPC), is presented to achieve path tracking for comfortable autonomous driving (CAD) using parallel control and deep deterministic policy gradient (DDPG). Based on parallel control, the developed deep-ETPC method constructs a dynamic control policy by introducing variation rates of controls. By employing variation rates of controls, the developed deep-ETPC method is capable of indicating communication loss and comfortable driving indices in the reward, and then enables reinforcement learning (RL) agents to learn comfortable ETC driving policies directly. Moreover, the communication loss, which reflects ETC, is integrated into the reward, so there is no need to additionally design/train triggering conditions, which can be considered a type of multi-tasking learning. Furthermore, an EPTC-oriented DDPG algorithm is developed to achieve the developed deep-ETPC method, making DDPG applicable to ETC. Empirical results, including tracking a simple straight line trajectory and a complicated sinusoidal trajectory, demonstrate the effectiveness of the developed deepETPC method.
- 20 February 2024
- Published by Elsevier in Neurocomputing
- Vol. 578
Nearly optimal stabilization of unknown continuous-time nonlinear systems: A new parallel control approach
- 20 February 2024
- Published by Elsevier in Neurocomputing
- Vol. 578
Nearly optimal stabilization of unknown continuous-time nonlinear systems: A new parallel control approach
- 30 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 53 (10) , 6375-6387
Continuous-Time Stochastic Policy Iteration of Adaptive Dynamic Programming
In this article, we study the optimal control problem of continuous-time (CT) time-invariant nonlinear systems with stochastic nonlinear disturbances. A new stochastic adaptive dynamic programming (ADP) method is developed to solve the Hamilton–Jacobi–Bellman equation (HJBE). Under the conditional expectation, the value function and the control law are successively approximated simultaneously. The asymptotic stability of the closed-loop stochastic system in probability is analyzed by the stochastic Lyapunov direct method, and the convergence of the developed ADP method is given. Finally, four simulations illustrate the effectiveness of the developed method.
- 30 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics: Systems
- Vol. 53 (10) , 6375-6387
Continuous-Time Stochastic Policy Iteration of Adaptive Dynamic Programming
In this article, we study the optimal control problem of continuous-time (CT) time-invariant nonlinear systems with stochastic nonlinear disturbances. A new stochastic adaptive dynamic programming (ADP) method is developed to solve the Hamilton–Jacobi–Bellman equation (HJBE). Under the conditional expectation, the value function and the control law are successively approximated simultaneously. The asymptotic stability of the closed-loop stochastic system in probability is analyzed by the stochastic Lyapunov direct method, and the convergence of the developed ADP method is given. Finally, four simulations illustrate the effectiveness of the developed method.
- Top Cited Papers
- 27 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 8 (4) , 2821-2831
Event-Triggered Deep Reinforcement Learning Using Parallel Control: A Case Study in Autonomous Driving
This paper utilizes parallel control to investigate the problem of event-triggered deep reinforcement learning and develops an event-triggered deep Q-network (ETDQN) for decision-making of autonomous driving, without training an explicit triggering condition . Based on the framework of parallel control, the developed ETDQN incorporates information of actions into the feedback and constructs a dynamic control policy. First, in the realization of the dynamic control policy, we integrate the current state and the previous action to construct the augmented state as well as the augmented Markov decision process. Meanwhile, it is shown theoretically that the goal of the developed dynamic control policy is to learn the variation rate of the action. The augmented state contains information of the current state and the previous action, which enables the developed ETDQN to directly design the immediate reward considering communication loss. Then, based on dueling double deep Q-network (dueling DDQN), we establish the augmented action-value, value, and advantage functions to directly learn the optimal event-triggered decision-making policy of autonomous driving without an explicit triggering condition. It is worth noticing that the developed ETDQN applies to various deep Q-networks (DQNs). Empirical results demonstrate that, in event-triggered control, the developed ETDQN outperforms dueling DDQN and reduces communication loss effectively. Source code of this paper is available at: https://github.com/lujingweihh/Event-triggered-deep-Q-network https://github.com/lujingweihh/Event-triggered-deep-Q-network.
- 27 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 8 (4) , 2821-2831
Top Cited PapersEvent-Triggered Deep Reinforcement Learning Using Parallel Control: A Case Study in Autonomous Driving
This paper utilizes parallel control to investigate the problem of event-triggered deep reinforcement learning and develops an event-triggered deep Q-network (ETDQN) for decision-making of autonomous driving, without training an explicit triggering condition . Based on the framework of parallel control, the developed ETDQN incorporates information of actions into the feedback and constructs a dynamic control policy. First, in the realization of the dynamic control policy, we integrate the current state and the previous action to construct the augmented state as well as the augmented Markov decision process. Meanwhile, it is shown theoretically that the goal of the developed dynamic control policy is to learn the variation rate of the action. The augmented state contains information of the current state and the previous action, which enables the developed ETDQN to directly design the immediate reward considering communication loss. Then, based on dueling double deep Q-network (dueling DDQN), we establish the augmented action-value, value, and advantage functions to directly learn the optimal event-triggered decision-making policy of autonomous driving without an explicit triggering condition. It is worth noticing that the developed ETDQN applies to various deep Q-networks (DQNs). Empirical results demonstrate that, in event-triggered control, the developed ETDQN outperforms dueling DDQN and reduces communication loss effectively. Source code of this paper is available at: https://github.com/lujingweihh/Event-triggered-deep-Q-network https://github.com/lujingweihh/Event-triggered-deep-Q-network.