Abstract: To endow robotic arms with enhanced autonomous obstacle avoidance and path planning capabilities in complex dynamic environments, deep reinforcement learning (DRL) methods have been widely ...