Jizheng Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Jizheng Liu has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Lei Zhang
Beijing Institute of Technology
6 shared publications - Zhenpo Wang
National Engineering Research Center for Electric Vehicles, Beijing Institute of Technology, Beijing, China
6 shared publications Shuhui Cheng
Beijing Institute of Technology
2 shared publications- Xiaolin Ding
Beijing Institute of Technology
2 shared publications - Cong Wang
School of Vehicle and Mobility, Tsinghua University, Beijing, China
1 shared publication - Dongpu Cao
School of Vehicle and Mobility, Tsinghua University, Beijing, China
1 shared publication
Latest Publications
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. PP, 1
AnchorSED: A Segmentation Enhanced Detection Framework Based on Sparse Anchor for Multi-View 3D Object Detection
Existing 3D detection methods typically exploit 2D image information through region-level feature extraction and auxiliary semantic supervision. However, these strategies do not explicitly establish a correspondence between 3D candidates and corresponding 2D instance-level representations, which limits the ability to handle objects with complex shapes or severe occlusions. To address this limitation, we propose an anchorbased framework that establishes an anchor-conditioned 2D-3D interaction for jointly exploiting 2D and 3D annotations. Specifically, each 3D anchor and its sampled points are projected onto multiple image planes to identify anchor-associated regions of interest (ROIs), from which view-specific instance segmentation masks are predicted. These 2D representations are subsequently mapped back to their corresponding 3D anchors and aggregated through geometry-aware cross-modal feature alignment. Our approach extends conventional sparse detectors by introducing three components: (1) multi-view ROIs prediction, which maps 3D anchors and sampled points onto multi-view image planes to predict regions of interest (ROIs); (2) multi-view instance segmentation, generating pixel-accurate masks corresponding to predicted ROIs; and (3) cross-modal feature alignment, which fuses geometrically consistent 2D segmentation features with their associated 3D anchors. Furthermore, to bridge the geometric discrepancy between 2D images and 3D objects, we distill multi-scale visual knowledge from a pretrained large-scale self-supervised vision model into our detector, enhancing geometric consistency. Evaluated on the nuScenes benchmark, our framework achieves state-of-the-art performance while maintaining computational efficiency. The code and auxiliary supervision data are available at https://github.com/HuiAuto/AnchorSED.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. PP, 1
AnchorSED: A Segmentation Enhanced Detection Framework Based on Sparse Anchor for Multi-View 3D Object Detection
Existing 3D detection methods typically exploit 2D image information through region-level feature extraction and auxiliary semantic supervision. However, these strategies do not explicitly establish a correspondence between 3D candidates and corresponding 2D instance-level representations, which limits the ability to handle objects with complex shapes or severe occlusions. To address this limitation, we propose an anchorbased framework that establishes an anchor-conditioned 2D-3D interaction for jointly exploiting 2D and 3D annotations. Specifically, each 3D anchor and its sampled points are projected onto multiple image planes to identify anchor-associated regions of interest (ROIs), from which view-specific instance segmentation masks are predicted. These 2D representations are subsequently mapped back to their corresponding 3D anchors and aggregated through geometry-aware cross-modal feature alignment. Our approach extends conventional sparse detectors by introducing three components: (1) multi-view ROIs prediction, which maps 3D anchors and sampled points onto multi-view image planes to predict regions of interest (ROIs); (2) multi-view instance segmentation, generating pixel-accurate masks corresponding to predicted ROIs; and (3) cross-modal feature alignment, which fuses geometrically consistent 2D segmentation features with their associated 3D anchors. Furthermore, to bridge the geometric discrepancy between 2D images and 3D objects, we distill multi-scale visual knowledge from a pretrained large-scale self-supervised vision model into our detector, enhancing geometric consistency. Evaluated on the nuScenes benchmark, our framework achieves state-of-the-art performance while maintaining computational efficiency. The code and auxiliary supervision data are available at https://github.com/HuiAuto/AnchorSED.
- 8 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 11 (2) , 5950-5960
Delay Compensation Control of Heterogeneous Vehicle Platoons: A Principal–Agent Structure
This paper proposes a Principal-Agent Cooperative Adaptive Cruise Control (PA-CACC) framework that integrates with the Smith predictor to simultaneously deal with actuator and communication delays for heterogeneous vehicle platoons. Theoretical analysis indicates that the proposed PA-CACC controller can substantially reduce the minimum time headway to the sum of actuator and communication delays while overcoming the impact of initial errors. Comprehensive comparison studies have been conducted to compare the performance of the proposed PA-CACC scheme with the state-of-the-art methods including a-CACC and Smith-predictor-based CACC. The results show that the proposed PA-CACC method outperforms the other two approaches and presents a robust solution to time-varying delays and parameter uncertainty. These underscore the effectiveness of the proposed method to enhance traffic efficiency and stability for heterogeneous vehicle platoons.
- 8 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 11 (2) , 5950-5960
Delay Compensation Control of Heterogeneous Vehicle Platoons: A Principal–Agent Structure
This paper proposes a Principal-Agent Cooperative Adaptive Cruise Control (PA-CACC) framework that integrates with the Smith predictor to simultaneously deal with actuator and communication delays for heterogeneous vehicle platoons. Theoretical analysis indicates that the proposed PA-CACC controller can substantially reduce the minimum time headway to the sum of actuator and communication delays while overcoming the impact of initial errors. Comprehensive comparison studies have been conducted to compare the performance of the proposed PA-CACC scheme with the state-of-the-art methods including a-CACC and Smith-predictor-based CACC. The results show that the proposed PA-CACC method outperforms the other two approaches and presents a robust solution to time-varying delays and parameter uncertainty. These underscore the effectiveness of the proposed method to enhance traffic efficiency and stability for heterogeneous vehicle platoons.
- 16 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (11) , 16375-16388
A Safety-Enhanced Reinforcement Learning-Based Decision-Making and Motion Planning Method for Left-Turning at Unsignalized Intersections for Automated Vehicles
Left-turning at unsignalized intersections poses significant challenges for automated vehicles. On this regard, Deep Reinforcement Learning (DRL) methods can achieve better traffic efficiency and success rate than rule-based methods, but they occasionally lead to collisions. This paper proposes a safety-enhanced method that integrates the DRL and the Dimensionality Reduction Monte Carlo Tree Search (DRMCTS) algorithm to achieve safety-enhanced trajectory planning at unsignalized intersections. First, DRMCTS is employed to address the partially observable Markov decision process problem. Through dimensionality reduction, it effectually enhances computational efficiency and problem-solving performance. Then a unified framework is introduced by simultaneously implementing DRL and the Gaussian Mixture Model Hidden Markov Model (GMM-HMM) in real-time. DRL determines actions in the current state while GMM-HMM identifies the turning intentions of surrounding vehicles (SVs). Under safe driving conditions, DRL makes decisions and outputs longitudinal acceleration with optimized ride comfort and traffic efficiency. When unsafe driving conditions are detected, DRMCTS would be activated to generate a collision-free trajectory to enhance the ego vehicle's driving safety. Through comprehensive simulations, the proposed scheme demonstrates superior traffic efficiency and reduced collision rates at unsignalized intersections with multiple SVs present.
- 16 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (11) , 16375-16388
A Safety-Enhanced Reinforcement Learning-Based Decision-Making and Motion Planning Method for Left-Turning at Unsignalized Intersections for Automated Vehicles
Left-turning at unsignalized intersections poses significant challenges for automated vehicles. On this regard, Deep Reinforcement Learning (DRL) methods can achieve better traffic efficiency and success rate than rule-based methods, but they occasionally lead to collisions. This paper proposes a safety-enhanced method that integrates the DRL and the Dimensionality Reduction Monte Carlo Tree Search (DRMCTS) algorithm to achieve safety-enhanced trajectory planning at unsignalized intersections. First, DRMCTS is employed to address the partially observable Markov decision process problem. Through dimensionality reduction, it effectually enhances computational efficiency and problem-solving performance. Then a unified framework is introduced by simultaneously implementing DRL and the Gaussian Mixture Model Hidden Markov Model (GMM-HMM) in real-time. DRL determines actions in the current state while GMM-HMM identifies the turning intentions of surrounding vehicles (SVs). Under safe driving conditions, DRL makes decisions and outputs longitudinal acceleration with optimized ride comfort and traffic efficiency. When unsafe driving conditions are detected, DRMCTS would be activated to generate a collision-free trajectory to enhance the ego vehicle's driving safety. Through comprehensive simulations, the proposed scheme demonstrates superior traffic efficiency and reduced collision rates at unsignalized intersections with multiple SVs present.
- 16 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 10 (2) , 2988-3001
Efficient Eco-Driving Control for EV Platoons in Mixed Urban Traffic Scenarios Considering Regenerative Braking
Connected and automated vehicles provide enormous opportunities for improving fuel economy, safety and capacity of the transportation system. In this paper, an Eco-driving speed advisory for electric vehicle platoons is proposed by taking regenerative braking and braking torque distribution into account. An efficient Poly-Eco speed planning method is then presented to improve computational efficiency. A control scheme comprising a splitting/merging decision-making and a modified intelligent driver model vehicle-following controller is established to verify the effectiveness of the proposed Poly-Eco speed advisory in mixed traffic scenarios. Comprehensive hardware-in-the-loop tests are conducted to examine the proposed Poly-Eco control scheme in terms of energy consumption, traffic efficiency and ride comfort. The rule-based and sequential programming (SP) methods are used for comparison. The comparison results show that the proposed Poly-Eco method can reduce the energy consumption by 4.71% while guaranteeing lower jerk and less arrival time compared with the commonly-used SP method. The lapse time per period for the proposed Poly-Eco is only 2.3% of that for the SP method. In addition, the proposed Poly-Eco method stages superior performance under typical mix traffic scenarios.
- 16 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 10 (2) , 2988-3001
Efficient Eco-Driving Control for EV Platoons in Mixed Urban Traffic Scenarios Considering Regenerative Braking
Connected and automated vehicles provide enormous opportunities for improving fuel economy, safety and capacity of the transportation system. In this paper, an Eco-driving speed advisory for electric vehicle platoons is proposed by taking regenerative braking and braking torque distribution into account. An efficient Poly-Eco speed planning method is then presented to improve computational efficiency. A control scheme comprising a splitting/merging decision-making and a modified intelligent driver model vehicle-following controller is established to verify the effectiveness of the proposed Poly-Eco speed advisory in mixed traffic scenarios. Comprehensive hardware-in-the-loop tests are conducted to examine the proposed Poly-Eco control scheme in terms of energy consumption, traffic efficiency and ride comfort. The rule-based and sequential programming (SP) methods are used for comparison. The comparison results show that the proposed Poly-Eco method can reduce the energy consumption by 4.71% while guaranteeing lower jerk and less arrival time compared with the commonly-used SP method. The lapse time per period for the proposed Poly-Eco is only 2.3% of that for the SP method. In addition, the proposed Poly-Eco method stages superior performance under typical mix traffic scenarios.
- 11 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 8 (3) , 2062-2076
Integrated Post-Impact Planning and Active Safety Control for Autonomous Vehicles
How to reduce traffic accidents and improve vehicle safety has always aroused widespread concern. Secondary and serial collision accidents can result in more serious hazards as the initial collision can easily destabilize a high-speed vehicle and the driver may fail to maintain effective control due to panic. With this in mind, a planning-integrated active safety control system is developed for post-impact vehicles to avoid subsequent accidents. The vehicle dynamics model is established considering the roll degree of freedom. The constraint equivalent methods based on the octagon and rhombus envelopes are proposed to linearize the road adhesion constraint and the obstacle avoidance constraint, respectively. A planning-integrated model predictive controller (MPC) is developed for post-impact vehicles to achieve the coordination of stability recovery and the avoidance of secondary collision with surrounding vehicles. In the meantime, an obstacle avoidance decision strategy based on the safe braking distance is designed to cope with different traffic scenes. Furthermore, a control allocator based on particle swarm optimization is developed to achieve the high-efficiency allocation of the resultant control output of MPC under post-impact extreme conditions. The proposed scheme is verified under comprehensive driving scenarios through hardware-in-loop tests.
- 11 January 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 8 (3) , 2062-2076
Integrated Post-Impact Planning and Active Safety Control for Autonomous Vehicles
How to reduce traffic accidents and improve vehicle safety has always aroused widespread concern. Secondary and serial collision accidents can result in more serious hazards as the initial collision can easily destabilize a high-speed vehicle and the driver may fail to maintain effective control due to panic. With this in mind, a planning-integrated active safety control system is developed for post-impact vehicles to avoid subsequent accidents. The vehicle dynamics model is established considering the roll degree of freedom. The constraint equivalent methods based on the octagon and rhombus envelopes are proposed to linearize the road adhesion constraint and the obstacle avoidance constraint, respectively. A planning-integrated model predictive controller (MPC) is developed for post-impact vehicles to achieve the coordination of stability recovery and the avoidance of secondary collision with surrounding vehicles. In the meantime, an obstacle avoidance decision strategy based on the safe braking distance is designed to cope with different traffic scenes. Furthermore, a control allocator based on particle swarm optimization is developed to achieve the high-efficiency allocation of the resultant control output of MPC under post-impact extreme conditions. The proposed scheme is verified under comprehensive driving scenarios through hardware-in-loop tests.
- Open Access
- 30 July 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-9
Low-Power Communication Signal Enhancement Method of Internet of Things Based on Nonlocal Mean Denoising
In order to improve the transmission effect of low-power communication signal of Internet of Things and compress the enhancement time of low-power communication signal, this paper designs a low-power communication signal enhancement method of Internet of Things based on nonlocal mean denoising. Firstly, the residual of one-dimensional communication layer is preprocessed by convolution core to obtain the residual of one-dimensional communication layer. Then, according to the two classification recognition methods, the noise reduction signal feature recognition of the low-power communication signal of the Internet of Things is realized, the nonlocal mean noise reduction algorithm is used to remove the low-power communication signal of the Internet of Things, and the weight value between similar blocks is calculated according to the European distance method. Finally, the low-power communication signal enhancement of the Internet of Things is realized by the nonlocal mean value denoising method. The experimental results show that the communication signal enhancement time overhead of this method is low, which is always less than 2.6s. The lowest bit error rate after signal enhancement is about 1, and the signal-to-noise ratio is up to 18dB, which shows that this method can achieve signal enhancement.
- 30 July 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-9
Open AccessLow-Power Communication Signal Enhancement Method of Internet of Things Based on Nonlocal Mean Denoising
In order to improve the transmission effect of low-power communication signal of Internet of Things and compress the enhancement time of low-power communication signal, this paper designs a low-power communication signal enhancement method of Internet of Things based on nonlocal mean denoising. Firstly, the residual of one-dimensional communication layer is preprocessed by convolution core to obtain the residual of one-dimensional communication layer. Then, according to the two classification recognition methods, the noise reduction signal feature recognition of the low-power communication signal of the Internet of Things is realized, the nonlocal mean noise reduction algorithm is used to remove the low-power communication signal of the Internet of Things, and the weight value between similar blocks is calculated according to the European distance method. Finally, the low-power communication signal enhancement of the Internet of Things is realized by the nonlocal mean value denoising method. The experimental results show that the communication signal enhancement time overhead of this method is low, which is always less than 2.6s. The lowest bit error rate after signal enhancement is about 1, and the signal-to-noise ratio is up to 18dB, which shows that this method can achieve signal enhancement.
- Open Access
- 22 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 71 (11) , 11571-11588
A Comprehensive Vehicle Stability Assessment System Based on Enabling Tire Force Estimation
This paper presents an enabling comprehensive vehicle stability assessment system covering vehicle longitudinal, yaw and roll stability. First, the longitudinal, lateral and vertical tire forces are separately estimated using the strong tracking unscented Kalman filter and the conventional Kalman filter based on low-cost on-board sensors. Then, a comprehensive vehicle longitudinal, yaw and roll stability space is established by utilizing the normalized tire friction ellipse and the Load Transfer Ratio value. Finally, the vehicle stability is determined and predicted based on current driver's inputs. The hardware-in-loop experimental results show that the proposed vehicle stability assessment system can accurately estimate the longitudinal, lateral and vertical tire forces with the normalized root mean square errors of 1.87$\%$ , 1.07$\%$ and 1.43$\%$ , and exhibits satisfying performance for vehicle longitudinal, yaw and roll stability evaluation and prediction. This bears significance for the efficient functioning of active control systems to improve vehicle safety under critical driving conditions.- 22 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 71 (11) , 11571-11588
Open AccessA Comprehensive Vehicle Stability Assessment System Based on Enabling Tire Force Estimation
This paper presents an enabling comprehensive vehicle stability assessment system covering vehicle longitudinal, yaw and roll stability. First, the longitudinal, lateral and vertical tire forces are separately estimated using the strong tracking unscented Kalman filter and the conventional Kalman filter based on low-cost on-board sensors. Then, a comprehensive vehicle longitudinal, yaw and roll stability space is established by utilizing the normalized tire friction ellipse and the Load Transfer Ratio value. Finally, the vehicle stability is determined and predicted based on current driver's inputs. The hardware-in-loop experimental results show that the proposed vehicle stability assessment system can accurately estimate the longitudinal, lateral and vertical tire forces with the normalized root mean square errors of 1.87$\%$ , 1.07$\%$ and 1.43$\%$ , and exhibits satisfying performance for vehicle longitudinal, yaw and roll stability evaluation and prediction. This bears significance for the efficient functioning of active control systems to improve vehicle safety under critical driving conditions. - Top Cited Papers
- 11 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 71 (4) , 3648-3659
Integrated Vehicle-Following Control for Four-Wheel-Independent-Drive Electric Vehicles Against Non-Ideal V2X Communication
Connected and automated vehicles (CAVs) have attracted tremendous interests worldwide. Four-Wheel-Independent-Drive Electric Vehicles (FWID EVs) have the potential of improving vehicle handling performance and energy consumption. In this paper, an integrated vehicle-following control scheme for FWID EVs with Vehicle-to-Everything (V2X) communication capability is proposed to account for nonideal communication such as time-varying delays and packet dropouts. A packet dropout compensator is put forward to compensate for V2X information loss. A longitudinal controller with a delay compensator is then synthesized and integrated with a lateral model predictive controller to enable vehicle-following control. The stability of the proposed controller is validated theoretically and experimentally under comprehensive driving scenarios through Hardware-In-the-Loop tests. The results demonstrate that the proposed controller has good vehicle-following performance against nonideal V2X communication. This attests to its competency for being used in vehicle platoon control.
- 11 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 71 (4) , 3648-3659
Top Cited PapersIntegrated Vehicle-Following Control for Four-Wheel-Independent-Drive Electric Vehicles Against Non-Ideal V2X Communication
Connected and automated vehicles (CAVs) have attracted tremendous interests worldwide. Four-Wheel-Independent-Drive Electric Vehicles (FWID EVs) have the potential of improving vehicle handling performance and energy consumption. In this paper, an integrated vehicle-following control scheme for FWID EVs with Vehicle-to-Everything (V2X) communication capability is proposed to account for nonideal communication such as time-varying delays and packet dropouts. A packet dropout compensator is put forward to compensate for V2X information loss. A longitudinal controller with a delay compensator is then synthesized and integrated with a lateral model predictive controller to enable vehicle-following control. The stability of the proposed controller is validated theoretically and experimentally under comprehensive driving scenarios through Hardware-In-the-Loop tests. The results demonstrate that the proposed controller has good vehicle-following performance against nonideal V2X communication. This attests to its competency for being used in vehicle platoon control.