Three-Dimensional Point Cloud Map Construction Based On Multi-Sensor Coupling

The development of robot technology has provided new ideas for unmanned operations in the environment of biomedical cold storage. Focusing on the sample transportation requirements in the complex environment of biological laboratories, this study centers on the biological sample transportation robot, delving into the construction of a three-dimensional point cloud map based on LiDAR, and initially builds a three-dimensional point cloud map simulating the cold storage environment, providing a solid foundation for the subsequent autonomous navigation of the transportation robot. Firstly, the principle of data acquisition by LiDAR is analyzed, the generated point cloud map is preprocessed, the NDT registration algorithm is adopted for the construction of the point cloud map, and finally, the information coupling of LiDAR, IMU and wheel odometry is carried out. This algorithm has a faster front-end registration speed, with a stable loop closure positioning error within 5 cm for every kilometer, and higher accuracy and robustness.