Trajectory Generation for Multi-phase Proximity Operations of a Space Robot

Space mission objectives have evolved from launching of satellites, to on-orbit operations including space manipulation, in-space assembly and debris mitigation. Each of these missions involves a sequence of proximity operations including orbit transfers, close-inspection, docking/berthing/capturing, servicing etc. As the number of satellites has increased drastically over the past decade, defining individual tasks/operations for each satellite is not economical. Hence there is a need for partial/full autonomy to these specialized spacecrafts or space robots. The paper explains the implementation of a multi-phase autonomous guidance strategy for proximity operations of a space robot. A constrained optimization-based trajectory generation coupled with learning-based situation awareness enables smooth transition between various operation phases as well as achieving the defined mission objectives. The algorithms are simulated in a high-fidelity physics framework integrated with sensors and actuators. The scenario and simulation results are presented for docking, however can be extended to other on-orbit operations with necessary modifications. Suggestions are provided to enhance the capability of proposed framework for end-to-end validation of mission objectives.