Journal
COMPUTERS
Volume 12, Issue 4, Pages -Publisher
MDPI
DOI: 10.3390/computers12040084
Keywords
autonomous vehicles; cruise control; multi-agent deep reinforcement learning; path following control; artificial intelligence
Ask authors/readers for more resources
This paper proposes a self-adaptive real-time cruise control system design for autonomous ground vehicles to achieve path-following control. It uses a multi-agent deep reinforcement learning technique to control acceleration and steering. The efficacy of the proposed control system is evaluated through simulation-based analysis and compared with two state-of-the-art controllers.
Significant progress has recently been made in transportation automation to alleviate human faults in traffic flow. Recent breakthroughs in artificial intelligence have provided justification for replacing human drivers with digital control systems. This paper proposes the design of a self-adaptive real-time cruise control system to enable path-following control of autonomous ground vehicles so that a self-driving car can drive along a road while following a lead vehicle. To achieve the cooperative objectives, we use a multi-agent deep reinforcement learning (MADRL) technique, including one agent to control the acceleration and another agent to operate the steering control. Since the steering of an autonomous automobile could be adjusted by a stepper motor, a well-known DQN agent is considered to provide the discrete angle values for the closed-loop lateral control. We performed a simulation-based analysis to evaluate the efficacy of the proposed MADRL path following control for autonomous vehicles (AVs). Moreover, we carried out a thorough comparison with two state-of-the-art controllers to examine the accuracy and effectiveness of our proposed control system.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available