Research on the positioning method of In-pipe detector based on strapdown inertial navigation and global optimization algorithm

To avoid the jamming risk, positioning error, and application limitation caused by In-pipe detector odometers, this paper introduces a global optimization algorithm to estimate the speed information, which can be an alternative to odometers. In the paper, a small In-pipe detector positioning system that incorporates MEMS IMU, motion state recognition, and global optimization algorithm is proposed. In this system, a machine learning algorithm identifies the motion state of the In-pipe detector. The global optimization algorithm obtains the optimal estimates of velocity and attitude angle for each segment of the trajectory. The extended Kalman filter fuses the estimates with the inertial navigation solving results. Finally, the target point is accurately located. Vehicle simulation experiments show that the positioning accuracy of the system without odometry reaches 1.3%, and the positioning accuracy is improved by 73.73% when it is used for the integrated navigation of inertial navigation plus odometr. The system provides a new method of obtaining velocity information, which can effectively improve positioning accuracy.