Vision-based autonomous docking for self-reconfigurable CubeSats

To meet the performance requirements of real-time and high accuracy for autonomous docking of modular self-reconfigurable CubeSats, a vision-based method is proposed by setting a cooperative target on a CubeSat and combining with Fractional-Order Darwinian Particle Swarm Optimization algorithm. Firstly, a monocular camera is utilized to grab images of the cooperation target. Then an image intelligent segmentation method is proposed based on FO-DPSO algorithm to achieve the recognition of the feature points of the cooperative target. These feature points' coordinates of the cooperative target are acquired by extracting the regional center of ellipse areas from the grabbed images. Finally, the relative position and attitude of the tracking CubeSat are calculated by the EPnP algorithm. The experiments are carried out on the established prototype. The results show that the extraction accuracy of the cooperation target feature points is within 0.932 pixel, and the positioning deviation of the tracking CubeSat is within 0.025 mm. The proposed autonomous docking method has a good real-time performance, and can meet the requirements of practical application.