Yixu He
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yixu He has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Yang Liu
School of Vehicle and Mobility, Tsinghua University, Beijing, China
3 shared publications - Hongyi Lin
School of Vehicle and Mobility, Tsinghua University, Beijing, China
2 shared publications Xiaobo Qu
2 shared publications- Xiaobo Qu
Tsinghua University
2 shared publications Yang Liu
2 shared publications- Cheng Lyu
Chair of Transportation Systems Engineering, Technical University of Munich, Munich, Germany
1 shared publication
Latest Publications
- 14 January 2026
- Published by Elsevier in Transportation Research Part D: Transport and Environment
- Vol. 153
A transformer-enhanced causal analysis framework for inter-stop energy consumption of electric bus
- 14 January 2026
- Published by Elsevier in Transportation Research Part D: Transport and Environment
- Vol. 153
A transformer-enhanced causal analysis framework for inter-stop energy consumption of electric bus
- Open Access
- 1 December 2025
- Published by KEAI Publishing in Journal of Traffic and Transportation Engineering (English Edition)
- Vol. 12 (6) , 1784-1803
Generative models for the evolution of transportation systems
- 1 December 2025
- Published by KEAI Publishing in Journal of Traffic and Transportation Engineering (English Edition)
- Vol. 12 (6) , 1784-1803
Open AccessGenerative models for the evolution of transportation systems
- Open Access
- 25 July 2025
- Published by Elsevier in Green Energy and Intelligent Transportation
- Vol. 4 (6)
LiDAR-IMU SLAM framework in autonomous modular bus docking systems
- 25 July 2025
- Published by Elsevier in Green Energy and Intelligent Transportation
- Vol. 4 (6)
Open AccessLiDAR-IMU SLAM framework in autonomous modular bus docking systems
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (8) , 12110-12114
Enhancing State Representation in Multi-Agent Reinforcement Learning for Platoon-Following Models
With the growing prevalence of autonomous vehicles and the integration of intelligent and connected technologies, the demand for effective and reliable vehicle speed control algorithms has become increasingly critical. Traditional car-following models, which primarily focus on individual vehicle pairs, exhibit limitations in complex traffic environments. To this end, this paper proposes an enhanced state representation for the application of multi-agent reinforcement learning (MARL) in platoon-following scenarios. Specifically, the proposed representation, influenced by feature engineering techniques in time series prediction tasks, thoroughly accounts for the intricate relative relationships between different vehicles within a platoon and can offer a distinctive perspective on traffic conditions to help improve the performance of MARL models. Experimental results show that the proposed method demonstrates superior performance in platoon-following scenarios across key metrics such as the time gap, distance gap, and speed, even reducing the time gap by 63%, compared with traditional state representation methods. These enhancements represent a significant step forward in ensuring the safety, efficiency, and reliability of platoon-following models within the context of autonomous vehicles.
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (8) , 12110-12114
Enhancing State Representation in Multi-Agent Reinforcement Learning for Platoon-Following Models
With the growing prevalence of autonomous vehicles and the integration of intelligent and connected technologies, the demand for effective and reliable vehicle speed control algorithms has become increasingly critical. Traditional car-following models, which primarily focus on individual vehicle pairs, exhibit limitations in complex traffic environments. To this end, this paper proposes an enhanced state representation for the application of multi-agent reinforcement learning (MARL) in platoon-following scenarios. Specifically, the proposed representation, influenced by feature engineering techniques in time series prediction tasks, thoroughly accounts for the intricate relative relationships between different vehicles within a platoon and can offer a distinctive perspective on traffic conditions to help improve the performance of MARL models. Experimental results show that the proposed method demonstrates superior performance in platoon-following scenarios across key metrics such as the time gap, distance gap, and speed, even reducing the time gap by 63%, compared with traditional state representation methods. These enhancements represent a significant step forward in ensuring the safety, efficiency, and reliability of platoon-following models within the context of autonomous vehicles.
- 12 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Transportation Systems Magazine
- Vol. 16 (3) , 111-124
Deep Demand Prediction: An Enhanced Conformer Model With Cold-Start Adaptation for Origin–Destination Ride-Hailing Demand Prediction
In intelligent transportation systems, one key challenge for managing ride-hailing services is the balancing of traffic supply and demand while meeting passenger needs within vehicle availability constraints. Accurate origin–destination (OD) demand predictions can empower platforms to execute timely reallocation of cruising vehicles and improve ride-sharing services. Nonetheless, the complexity of OD-based demand prediction arises from intricate spatiotemporal dependencies and a higher need for precision compared to zone-based predictions, which leads to many unprecedented OD pairs. To tackle this issue, we design a comprehensive set of 102 features, including travel demand, passenger count, travel volume, liveliness, weather, and cross features. We also introduce an enhanced conformer model, which is composed of a single conformer block that integrates feedforward layers, multihead self-attention mechanisms, and depth-wise separable convolution layers. To address the cold-start problem and manage large values, we design a specific algorithm for OD pairs lacking training data and apply a technique to handle larger values. Our approach demonstrates a marked improvement in prediction performance, with an 18% decrease in the total travel demand error and up to a 47% reduction for certain larger values in some cases. Through extensive experiments on a dataset collected from a city, provided by a ride-hailing platform, our proposed methods significantly outperform the most advanced models.
- 12 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Transportation Systems Magazine
- Vol. 16 (3) , 111-124
Deep Demand Prediction: An Enhanced Conformer Model With Cold-Start Adaptation for Origin–Destination Ride-Hailing Demand Prediction
In intelligent transportation systems, one key challenge for managing ride-hailing services is the balancing of traffic supply and demand while meeting passenger needs within vehicle availability constraints. Accurate origin–destination (OD) demand predictions can empower platforms to execute timely reallocation of cruising vehicles and improve ride-sharing services. Nonetheless, the complexity of OD-based demand prediction arises from intricate spatiotemporal dependencies and a higher need for precision compared to zone-based predictions, which leads to many unprecedented OD pairs. To tackle this issue, we design a comprehensive set of 102 features, including travel demand, passenger count, travel volume, liveliness, weather, and cross features. We also introduce an enhanced conformer model, which is composed of a single conformer block that integrates feedforward layers, multihead self-attention mechanisms, and depth-wise separable convolution layers. To address the cold-start problem and manage large values, we design a specific algorithm for OD pairs lacking training data and apply a technique to handle larger values. Our approach demonstrates a marked improvement in prediction performance, with an 18% decrease in the total travel demand error and up to a 47% reduction for certain larger values in some cases. Through extensive experiments on a dataset collected from a city, provided by a ride-hailing platform, our proposed methods significantly outperform the most advanced models.
- Top Cited Papers
- 8 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (1) , 1654-1666
Deep Adaptive Control: Deep Reinforcement Learning-Based Adaptive Vehicle Trajectory Control Algorithms for Different Risk Levels
In this study, we explore the problem of adaptive vehicle trajectory control for different risk levels. Firstly, we introduce a sliding window-based car-following scenario extraction method, propose a new alternative traffic conflict assessment metric, and build a comprehensive traffic scenario library. Secondly, based on deep reinforcement learning (RL), we design an adaptive car-following trajectory control algorithm, which is called Deep Adaptive Control , to cope with different traffic risk levels. Thirdly, we design five metrics in terms of safety, comfort, and energy consumption, and experimentally compare Deep Adaptive Control with human drivers and RL benchmarks. The experimental results show the superiority of Deep Adaptive Control compared to human drivers and existing RL methods, which can follow the preceding vehicle closely in low-risk situations to improve traffic efficiency, keep distance from the preceding vehicle in high-risk situations to improve safety and be optimal in terms of comfort and fuel consumption metrics.
- 8 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (1) , 1654-1666
Top Cited PapersDeep Adaptive Control: Deep Reinforcement Learning-Based Adaptive Vehicle Trajectory Control Algorithms for Different Risk Levels
In this study, we explore the problem of adaptive vehicle trajectory control for different risk levels. Firstly, we introduce a sliding window-based car-following scenario extraction method, propose a new alternative traffic conflict assessment metric, and build a comprehensive traffic scenario library. Secondly, based on deep reinforcement learning (RL), we design an adaptive car-following trajectory control algorithm, which is called Deep Adaptive Control , to cope with different traffic risk levels. Thirdly, we design five metrics in terms of safety, comfort, and energy consumption, and experimentally compare Deep Adaptive Control with human drivers and RL benchmarks. The experimental results show the superiority of Deep Adaptive Control compared to human drivers and existing RL methods, which can follow the preceding vehicle closely in low-risk situations to improve traffic efficiency, keep distance from the preceding vehicle in high-risk situations to improve safety and be optimal in terms of comfort and fuel consumption metrics.
- 1 August 2023
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12747, 34-397
Backlit traffic scene image enhancement based on foreground and background characteristics
Due to the low foreground brightness and distorted background brightness of traffic backlight images, a technique of fusing LIME and improved histogram equalization is proposed to eliminate the backlight phenomenon in images. Firstly, the input image is segmented into two parts, foreground and background, using the maximum interclass variance method. Then LIME method is used globally for the backlit image to enhance the foreground brightness while maintaining the color distortion, and only the foreground part of the processed image is retained. Then the pixel value distribution of the background part is calculated separately, and the global histogram equalization results on the three RGB channels are mapped one by one to the corresponding limited interval, which improves the contrast of the background. Finally, the Canny operator is used to detect the black edges at the front background stitching, and three adaptive filtering templates are generated based on the black edges to perform step-by-step mean filtering on the black edges, eliminating the black edges and improving the visual quality of the image. The average metrics of the proposed method on the laboratory selfconstructed CHD_B dataset are, respectively, NIQE of 5.37, BRISQUE of 47.74, average gradient of 0.79, information entropy of 6.61, and average running time of 11.38s, which are synthetically better than the current backlight image enhancement methods and have better visual quality and visibility
- 1 August 2023
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12747, 34-397
Backlit traffic scene image enhancement based on foreground and background characteristics
Due to the low foreground brightness and distorted background brightness of traffic backlight images, a technique of fusing LIME and improved histogram equalization is proposed to eliminate the backlight phenomenon in images. Firstly, the input image is segmented into two parts, foreground and background, using the maximum interclass variance method. Then LIME method is used globally for the backlit image to enhance the foreground brightness while maintaining the color distortion, and only the foreground part of the processed image is retained. Then the pixel value distribution of the background part is calculated separately, and the global histogram equalization results on the three RGB channels are mapped one by one to the corresponding limited interval, which improves the contrast of the background. Finally, the Canny operator is used to detect the black edges at the front background stitching, and three adaptive filtering templates are generated based on the black edges to perform step-by-step mean filtering on the black edges, eliminating the black edges and improving the visual quality of the image. The average metrics of the proposed method on the laboratory selfconstructed CHD_B dataset are, respectively, NIQE of 5.37, BRISQUE of 47.74, average gradient of 0.79, information entropy of 6.61, and average running time of 11.38s, which are synthetically better than the current backlight image enhancement methods and have better visual quality and visibility