Space-based multi-sensor multi-target tracking for geosynchronous space objects

This study develops a space-based multi-sensor multi-target tracking (MSMTT) strategy for geosynchronous space object (GEO) catalog maintenance. Compared with traditional ground-based approaches, the space-based sensors have better detection capability because they are rarely restricted by weather conditions and the day-night gap. Two space-based sensor networks based on geosynchronous transfer orbit and Sun synchronous orbit are investigated, and two hybrid sensor networks using both space and ground based sensors are proposed as a complementary approach. The expression of the Rényi divergence for δ Generalized Labeled Multi-Bernoulli densities is derived and utilized as the reward function for sensor management. The proposed MSMTT method is formulated in the framework of the Labeled Multi-Bernoulli filter in conjunction with the information theoretic sensor management which assigns sensor tasks by maximizing the information gain. The proposed method is validated by two case studies, involving large scale GEO objects tracking and dim GEO objects tracking. Results indicate the better performance of the spaced-based and hybrid sensor network compared with the ground-based approach in terms of state and cardinality estimation, and the Rényi divergence achieves similar accuracy compared to the commonly used Cauchy-Schwarz divergence. Further, the GTO sensor network is more effective for dim GEO object tracking.