Zhao, Xinshi (Technical University of Munich), Liu, Di (Technical University of Munich), Yang, Kang (Southeast University), Baldi, Simone (Southeast University)

Lyapunov-Based Inverse Reinforcement Learning for Vehicle-Following Traffic Scenarios

Scheduled for presentation during the Invited Session "Learning-powered and Knowledge-driven Autonomous Driving I" (ThAT1), Thursday, September 26, 2024, 12:10−12:30, Salon 1

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on December 26, 2024

Abstract

This work presents an approach for capturing vehicle-following behavior on highways, based on Inverse Reinforcement Learning (IRL) with control Lyapunov function. The idea is to describe the vehicle-following behavior as an optimal control problem with its underlying cost and constraints expressed in terms of the data. Using the highD dataset as case study, we identify vehicle-following dynamics and frame them in an IRL context. Using kernel regression, we show that such IRL boils down to a Quadratic Programming (QP), solvable using standard optimization routines.