Topic: Safety Decision-making, Planning and Control of Autopilot in Complex Scenes

Informant: Hu Chuan (Associate Professor, Shanghai Jiaotong University)

Reporting time: 14:00-17:00 pm on Saturday, May 16, 2026.

Address: Meeting Room 820, Automotive Technology Building, Wushan Campus, South China University of Technology

Inviter: Associate Professor Liu Qiaobin

Organizer: School of Mechanical and Automotive Engineering

Summary of report:

The complex traffic and road environment poses a serious security threat to the decision-making, planning and control of autonomous driving. This report focuses on the complex environment and driving conditions, fully explores the game and interaction between people and cars, and quantitatively considers the impact of environmental uncertainties, including the uncertainty of perception and prediction, on decision-making, improves the existing physical information-driven safety deep reinforcement learning method, and constructs safety decision-making methods to improve safety and generalization for different scenarios. At the same time, in the lower control module, aiming at the comprehensive requirements of control performance, driving safety and ride comfort, and considering the uncertainty of multi-modal trajectory prediction of surrounding vehicles, the control strategy combining adaptive dynamic programming and reinforcement learning-model predictive control is adopted to fully ensure the control safety and robustness. The above research can provide theoretical support with practical application value for automatic driving regulation and control in complex scenes and extreme working conditions.

Brief introduction of the speaker:

Hu Chuan, associate professor of School of Mechanical and Power Engineering, Shanghai Jiaotong University, doctoral supervisor, national overseas high-level young talents and leading talents in Shanghai. He has served as an assistant professor at the University of Alaska, and engaged in post-doctoral research at the University of Texas at Austin and the University of Waterloo in Canada. He received his doctor's degree in engineering from McMaster University and his bachelor's degree in engineering from Tsinghua University. Research areas include intelligent networked vehicles, unmanned vehicles and intelligent cockpits, including: perception/prediction/decision/planning, intelligent chassis control, large model/end-to-end driving, human-computer interaction and human-computer cooperative driving. He has presided over the research projects of NASA and the Ministry of Communications, and participated in and presided over many natural projects in China. Published more than 100 papers. He is the deputy editor-in-chief of many high-level journals in the industry, including TNNLS/TVT/TTE/TITS /TIV/SJ.