Path planning and dynamic collision avoidance algorithm under COLREGs via deep reinforcement learning

Keyphrases

• Collision Avoidance Algorithm 100%
• Path Planning 100%
• Pathway Dynamics 100%
• COLREGs 100%
• Dynamic Collision Avoidance 100%
• Deep Reinforcement Learning (deep RL) 100%
• Unmanned Surface Vehicle 66%
• Network Structure 16%
• Risk Estimation 16%
• Reward Function 16%
• Automatic Control System 16%
• Generalization Ability 16%
• Training Effectiveness 16%
• Real-time Performance 16%
• State Space 16%
• Real-time Security 16%
• Collision Avoidance 16%
• Experience Replay 16%
• Priority Sampling 16%
• Rudder Angle 16%
• Switching Time 16%
• Encounter Situation 16%
• Continuous Output 16%
• High-dimensional Input 16%
• Autonomous Obstacle Avoidance 16%
• Sampling Mechanism 16%
• Relative Parameters 16%
• Navigation Safety 16%

Engineering

• Collision Avoidance 100%
• Path Planning 100%
• Deep Reinforcement Learning 100%
• Unmanned surface vehicles 50%
• Core Technology 12%
• Automatic Control System 12%
• Safe Navigation 12%
• Rudder Angle 12%
• Thrust Angle 12%
• Paper Path 12%
• Input State 12%
• Dimensional Input 12%
• Switching Time 12%

Computer Science

• path-planning 100%
• Deep Reinforcement Learning 100%
• Network Structure 16%
• Assessment Model 16%
• State Space 16%
• Generalization Ability 16%
• Autonomous Vehicle 16%
• Core Technology 16%
• Real Time Performance 16%