Research On The Terminal Attitude Determination Method Of Robots Based On LPMS-B

The laser tracker was a mature solution for motion detection, especially for the attitude determination of robots with high precision. But the high cost of laser tracker made its extensive use unrealistic. In the light of this situation, attitude sensor based attitude determination with lower cost was proposed. Lots of examples proved its feasibility, such as UAV navigation, vehicle system and robot control system. It was believed that the attitude determination system based on sensors had huge application prospects. Therefore, this thesis proposed a robot attitude determination system named LPMS-B, which was an attitude determination algorithm based on gyroscope, accelerometer and magnetometer. Also, the novel determination system was verified on the OTC robot platform. The main contributions of this research could be listed as follows.(1)The research status of attitude determination was introduced, including the development of attitude sensors and detecting technology, algorithms for attitude solution and data-driven calculation methods.(2)To improve the precision of terminal attitude determination of robots, a determination method based on quaternion attitude calculation and improved gauss-newton (IGN) algorithm was proposed, which was called LPMS-B. And this method was combined with Kalman filter and complementary filter for better performance. In addition, gyroscope based rotational chair test and accelerometer based elevator test was conducted to verify the feasibility of data from gyroscope and accelerometer.(3)To test the performance of the proposed attitude determination method, single axle rotation and multi-axis motion experiment were conducted on the OTC robot platform. Based on these experiments, multiple attitude determination algorithms were implemented for comparison, including IGN algorithm, Kalman filter based quaternion calculation algorithm and complementary filter based quaternion calculation algorithm. It was found that complementary filter based quaternion calculation algorithm showed the best performance. In addition, the error impact during the attitude determination process was considered.(4)There were several other opportunities to extend the work in the future, especially the field of accuracy error of the attitude determination.