Study On Algorithms And Hardware-In-The-Loop Simulation Of Integrated Vision/Inertial Relative Navigation

Relative navigation is the critical technology of proximity operations for spacecraft, such as formation flying of micro satellites, space attack-defense and RVD. According to the characteristics of proximity operations with close range relative navigation, an algorithm of multi-rate Kalman data fusion filter for the integrated navigation with vision and inertial is researched mainly, supported by the Major State Basic Research Development Program of China (973 Program)—"Study of new concept, new mechanism of small spacecraft". Based on the MicroSim, a scheme of relative navigation phycical simulation system is designed and validated with the relative navigation algorithms. The main achievements obtained in this dissertation are as follows:Firstly, according to the close range relative navigation and generalized by employing the development of final approach phase for RVD, the monocular vision navigation algorithm based on the P4P coplanar is conducted, with the original errors analyzed, and the mathematical simulation is verified.After analyzed the monocular vision navigation algorithm, for the advantages and disadvantages of INS and vision-based navigation (VisNav), an algorithm of integrated INS/VisNav is studied. According to the multi-rate sampling problem, a multi-rate data fusion algorithm is provided then. The whole filtering processing is divided into two independent processes which include measurement update and time update. The fast sampling rate for INS is selected as the filter interval, and given both measurement update and time update when the vision sampling data is coming, on the contrary given only time update. Both theory analysis and mathematical simulation indicate that this algorithm can increase the data update rate, so that enhance the performance of the filter and augment the robust of the navigation system.Secondly, spacecraft relative navigation hardware-in-the-loop simulation system is designed and established based on MicroSim. The scheme of this system is given, the software and hardware of image sampling for monocular vision-based navigation is developed, and some critical teniques including vision camera calibration and image processing algorithm are also studied. Thus this system can have an ability to validate several kinds of relative navigation system by hardware-in-the-loop simulation.Finally, both of the monocular vision navigation algorithm and multi-rate Kalman data fusion filter algorithm are validated by this physical simulation system. The simulation results compared with mathematical simulation indicate that these relative algorithms are validated and the relative navigation hardware-in-the-loop simulation system is available also.