FPGA Implementation Of Relative Position And Attitude Calculation

Position and attitude measurement is an important subject in these filed about modern control, navigation, track, et al. The position and attitude measurement system should have advantage on these aspects: precision, real-time and facilitated installation .The position and attitude measurement system based on computer vision has got enough regard in the process of close docking of spacecraft and installation of large work-piece. But the existing position and attitude measurement system based on computer vision focus on measurement part. For the calculating part, there is not too much mention and research, and the final data processing and calculation are operated by computer or DSP. So, further researching and optimizing the calculation part circuit has important significance for position and attitude measurement system's speed upgrade and transformation into chip.In this paper, the principle of relative position and attitude determination which include position and attitude measurement based on monocular vision and the transformation between different spatial coordinates is described. In connection with the calculating part, an optimized VLSI implementation scheme is designed. The relative position and attitude calculating circuit is mapped to the following five basic components: vector normalizing, vector cross product, matrix equations solution, relative attitude angle calculation and relative position calculation. Each component is designed based on CORDIC algorithm, which only has simple shift-add computation units and are very suitable to be implemented in FPGA chip. The simulation result by ISE and ModelSim software shows that the scheme is feasible. Finally, it introduces how to hang the designed ISE project in the EDK OPB bus as the custom IP core, so as to implement the embedded design scheme.