Adaptive Control Strategy for Space Robot Target Manipulation Based on Decoupled Dynamic Modeling

The manipulation or grasping of unknown objects represents an important task for free-floating space robots. The dynamic coupling between the target object and the robot often introduces complexity into the control design. To address this, we employ an extended Rosenberg embedding method to dynamically decouple the center-body, unknown object, and dual arm. Based on the decoupled model, an adaptive control consisting of three components is introduced. The first two parts, under the known system model, are designed for the task trajectory following. The third part utilizes a boundary function to enhance the system's robustness in the face of the worst effects arising from uncertain parameters of the target. Combining these three terms allows the unknown target to track the desired trajectory while maintaining the attitude of the center body for flight safety. The stability of the robot system under the proposed method is theoretically guaranteed and the control performance is numerically verified by the simulation in 3-D.