Adopting Gradient Vector and Particle Filter in APT for Space Optical Communication

Acquisition and tracking for Light beacon in space optical communication is an important and challenging work. The facula of the received light beacon is modeled as a two-dimensional rigid circle, inside which gradient vectors are computed through the idea of key-point descriptor. By way of the Bhattacharyya distance between the circle object distribution and reference template distribution, observation probability model of system is built. Combining with observation probability and tow-order recursive motion model, an improved particle filter of optimization and combination is adopted to forecast the state with updating the reference template. Simulation results indicate that the method is effective even when the facula is moving rapidly and briefly occluded.