Multi-source information fusion SLAM front-end strategy based on EKF

The invention provides a multi-source information fusion algorithm based on extended Kalman filtering for solving the problem that scanning matching depends on an initial value in construction of a laser radar front-end sub-graph. The EKF is fused with a wheel encoder and IMU pre-integration, based on the obtained mileage pose (position + pose), the fused pose and a covariance matrix are obtained, and quaternion spherical interpolation is carried out on the fused pose to remove laser radar motion distortion; in order to solve the problem that laser radar inter-frame matching needs more feature points, a matching strategy between frames and sub-graph matching is improved by using a current laser radar data frame and local map related scanning matching strategy; and meanwhile, the fusion pose is used as an initial value of laser radar iteration, so that the iteration time is optimized, and the positioning precision is improved. Finally, simulation is carried out through a gazebo platform, and the effectiveness