Orbit-Attitude Coupled Control for Multi-Target Tracking Based on Partition Pattern Search

The growing amount of small space debris poses a significant threat to spacecrafts, with the potential to cause structural damage or even catastrophic failures. Space-based close proximity observation offers an effective alternative to conventional ground-based methods, enabling more precise debris detection through active orbital maneuvers and sensor attitude alignment. However, this approach faces two main challenges. The first is ensuring safe orbit and attitude control to achieve accurate target observation, especially when information about multiple debris is limited; The second is addressing the orbit-attitude coupling problem, which involves a high-dimensional optimization task requiring significant computational resources. This paper presents a robust orbit-attitude coupled control method leveraging Outer Probability Measures to address uncertainties caused by incomplete knowledge of multiple debris. To mitigate the computational complexity, a partitioned pattern search algorithm is developed, which incorporates the concept of coordinate descent to find the optimal control solution iteratively. This method decomposes the high-dimensional optimization problem into more manageable sub-problems, thereby reducing the computational burden. A convergence analysis is conducted, and an initial guess generation strategy is introduced to further accelerate the optimization process. Simulation results validate its robustness and accuracy in close proximity.