LAUSR

During the operation of space manipulators for debris removal, the obstacles moving in the workspace must be avoided. We propose a unified modelling framework for multiple moving obstacles and a collision-free trajectory planning method for a redundant space manipulator. The complete properties of an obstacle, including its shape, dimension, pose (position and orientation), and velocity (linear and angular), are defined in the model. The obstacle surface is represented by a super quadratic function whose parameters are adjusted to describe different shapes and dimensions. Pseudo-distance is defined to evaluate the proximity extent between the manipulator and an obstacle. Considering multiple different obstacles, we present an approach to normalize the pseudo-distances. The self-motion of the redundant manipulator was used to optimize the normalized pseudo-distance by adaptive redundancy resolution. By ensuring that the pseudo-distance was always larger than the safety threshold value, collisions with the obstacles were avoided. The proposed method solved the problem for which the Euclidean distance was difficult, or even impossible, to calculate for 3-D cases. When handling multiple different obstacles, the proposed method was much easier and had higher computational efficiency than previous methods. The proposed method was verified by the simulation of typical missions.