Design of GNSS/SLAM/IMU Multi-Source Information Fusion Navigation System

The traditional vision/inertial/satellite integrated navigation and positioning system often faces the challenge of time reference mismatch in the initial stage, which is manifested in the timing deviation between satellite positioning data and visual perception information, which affects the overall performance of the system. In order to solve the problem of accuracy error, a new design scheme of GNSS/SLAM/IMU multi-source information fusion navigation system is proposed. This system can not only maintain high precision positioning when GNSS has no signal or weak signal, but also reduce the accumulated drift caused by inertial sensor, and solve the shortcomings of absolute position drift of vision sensor and easy to be disturbed by light changes and dynamic objects. The technical realization path can be divided into three progressive levels: the first stage is to solve the transformation relationship between the carrier coordinate system and the geodetic coordinate system through kinematic modeling, establish the space mapping relationship of the double coordinate system based on the motion characteristics of the rigid body, and complete the joint initialization of the multi-source sensor. In the second stage, the time synchronization error between the satellite observation data and the visual odometer is accurately identified by the kinematic integral characteristic of the inertial sensor. In the third stage, the dynamic compensation mechanism is introduced into the graph optimization framework, and the parameter joint optimization is realized through the graph optimization model, and the multi-source sensor data is fused by Bayesian estimation theory. In order to verify the effectiveness of the algorithm, this paper carries out validation on the public data set and the collection of actual data. Through empirical analysis, GNSS/SLAM/IMU multi-source information fusion navigation system has shown significant improvement in positioning accuracy. Comparative experimental data show that after optimization, the root-mean-square deviation of the trajectory of the system is reduced by 40.77 %, the standard deviation data is reduced by 40.97 %, and the stability is improved by 41.73 %. This innovative improvement of GNSS/SLAM/IMU multi-source information fusion navigation system effectively improves the positioning accuracy of traditional multi-source fusion navigation algorithm.