Intelligent Control for a Non-holonomic Constrained Mobile Robot with Proximal Policy Optimization
- Resource Type
- Conference
- Authors
- Xie, Junran; Wang, Qingling
- Source
- 2022 34th Chinese Control and Decision Conference (CCDC) Control and Decision Conference (CCDC), 2022 34th Chinese. :2913-2918 Aug, 2022
- Subject
- General Topics for Engineers
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Training
Simulation
Robot control
Reinforcement learning
Markov processes
Robustness
Real-time systems
Proximal Policy Optimization
Non-holonomic Constrained
Mobile Robot
- Language
- ISSN
- 1948-9447
The dominant control algorithm using in the non-holonomic constrained mobile robot usually depends on accurate modeling and complex calculations. This paper proposes a new deep reinforcement learning-based method to solve the control problem of the non-holonomic constrained mobile robot. An end-to-end control of the mobile robot is realized by proximal policy optimization learning algorithm. The control policy based on reinforcement learning can more efficiently calculate the actions according to the current state of the robot. This rapid autonomous response is expected to open up a path for the real-time controllable robot. Moreover, this paper establishes an obstacle environment based on the kinematics model of the non-holonomic constrained mobile robot to verify the effectiveness of the algorithm. Finally, this simulation results illustrate the control method with favorable robustness and generalization for the non-holonomic constrained mobile robot control.