Deep Reinforcement Learning for Time Optimal Velocity Control using\n Prior Knowledge

2018/11/28 by Gabriel Hartmann, Hartmann, Gabriel, Zvi Shiller +3 · 1 citation
Engineering · #Autonomous Vehicle Technology and Safety #FOS: Computer and information sciences #Machine Learning (cs.LG) #Robotics (cs.RO) #Traffic control and management #Vehicle Dynamics and Control Systems

paper · pdf · doi:10.48550/arxiv.1811.11615

openalex publication_date 2018/11/28 · openalex created_date 2022/08/01 · openalex updated_date 2026/07/28

Abstract

Autonomous navigation has recently gained great interest in the field of\nreinforcement learning. However, little attention was given to the time optimal\nvelocity control problem, i.e. controlling a vehicle such that it travels at\nthe maximal speed without becoming dynamically unstable (roll-over or sliding).\n Time optimal velocity control can be solved numerically using existing\nmethods that are based on optimal control and vehicle dynamics. In this paper,\nwe use deep reinforcement learning to generate the time optimal velocity\ncontrol. Furthermore, we use the numerical solution to further improve the\nperformance of the reinforcement learner. It is shown that the reinforcement\nlearner outperforms the numerically derived solution, and that the hybrid\napproach (combining learning with the numerical solution) speeds up the\ntraining process.\n

Cited by

• A Visual Tool for Interactive Model Explanation using Sensitivity Analysis

Related

• Combining Benefits from Trajectory Optimization and Deep Reinforcement Learning
• Optimal control of point-to-point navigation in turbulent time-dependent flows using Reinforcement Learning
• Automating Vehicles by Deep Reinforcement Learning using Task Separation with Hill Climbing
• Towards Optimal Head-to-head Autonomous Racing with Curriculum Reinforcement Learning
• High-speed Autonomous Racing using Trajectory-aided Deep Reinforcement Learning