Constrained attitude path generation of spacecraft base on rapidly exploring random tree and quadratic programming

During the lifetime of a spacecraft, a lot of attitude maneuver operations are necessary for mission requirement. If a special sensor, which cannot point to Sun or other bright objects, is installed on the spacecraft, a least separation angle between the direction of the sensor and the bright object must be kept. Therefore, the attitude free space is restricted and the attitude should not slew arbitrarily. The analogous case exists for communication with earth by antenna and power supply by solar panel. The ability to autonomously control attitude maneuver under geometric constraints is one of the key technologies in some future space missions. This paper proposed an attitude slew maneuver path generation algorithm, which is based on Rapidly Exploring Random Tree (RRT) and Quadratic programming(QP). Due to the mature technology of uniform sample in attitude space, the RRT, as a global planner, freely and incrementally expands to some chosen uniform sample point in a greedy pattern. By this means, the geometric constraints can be well satisfied and the tree can expand fast and terminate within finite iterations. Meanwhile, the local planner transforms the dynamic equation to linear constraints of quadratic programming problem. In each step, only a trivial numerical procedure is needed although it assumes the main computational cost. Consequently, the geometric and dynamic constraints are met respectively, and the maneuver path of kinodynamics is generated. Simulation results and analysis validate the advantages of the proposed algorithm.