Task planning of Robotic Grasping for Deep-Container Environment Based on Dynamic Space Constraints RRTC Algorithm

Multi-axis collaborative manipulators encounter challenges such as slow planning speed, low efficiency, and poor path quality while conducting grasping operations. This paper proposes the Dynamic Space Constraints RRTC motion planning algorithm that considers the grasping pose in the deep-container picking problem and introduces a novel method to optimize the collision-free grasping pose in compliance with environmental constraints. The paper also proposes a prior graph path generation algorithm based on the end-effector’s sphere space constraint that compresses the body’s search space to improve efficiency. Avoiding complicated non-convex constraint modeling required for obstacle avoidance can be achieved by modeling nonconvex sets with an obstacle-free state as the union of a finite number of convex regions. Based on the prior path, the algorithm generates a dynamic constraint manifold and uses a bidirectional search tree that employs a sliding constraint manifold for traction. The algorithm includes an adaptive step length mechanism dynamically based on the obstacle distance for higher efficiency and accuracy. Experimental results demonstrate that the Dynamic Space Constraints RRTC algorithm plans an optimized path that satisfies the constraint condition compared to traditional algorithms in terms of computation time and efficiency.