The invention belongs to the technical field of intelligent
robot navigation and positioning, and discloses an intelligent
robot autonomous mapping method in a GNSS rejection environment, which comprises the following steps of: monitoring a GNSS
signal state in real time, and switching to pure
laser SLAM mapping when detecting that the number of available satellites, PDOP or pseudo-range residual exceeds a threshold; the method comprises the following steps: acquiring three-dimensional
point cloud data, filtering and denoising, obtaining a filtered and denoised current frame
point cloud, performing
laser radar mapping and positioning registration, selecting a
key frame point cloud, constructing lightweight neural implicit scene representation, and training an MLP network;
loopback detection is carried out, a sparse
factor graph is constructed,
key frame poses are corrected, then the
key frame poses are applied to original three-dimensional point
cloud data, and a two-dimensional grid map is generated. According to the invention, under the condition that the GNSS
signal is blocked or fails, the environment map with high precision and low drift is continuously generated, and the real-time performance and robustness of autonomous mapping are remarkably improved.