Unmanned trackless rubber wheeler based on LD improved Cartographer mapping algorithm underground SLAM autonomous navigation method and test

As one of the important forms of equipment for the auxiliary transportation system of mines, the autonomous trackless rubber-tyred vehicles will be an inevitable direction as a result of the intelligent development of mines. However, the current research in this field is still in its infancy, especially the lack of SLAM autonomous navigation method for the special harsh environment of underground roadways. In order to solve this problem, it is necessary to research and develop efficient and accurate mapping algorithms to support the safety and reliability of unmanned underground operation. Firstly, a two-dimensional raster map of the underground roadway environment is constructed based on the Cartographer algorithm, and the Lazy Decision (LD) algorithm is introduced to optimize it, which solves the overlapping and fuzzy phenomena during Cartographer mapping and improves the mapping accuracy. Secondly, the adaptive Monte Carlo positioning (AMCL) algorithm is selected to solve the positioning problem of unmanned trackless rubber wheeler, and the experimental results show that the AMCL algorithm can quickly achieve particle convergence, which can be completed within 12 seconds at the fastest, and shows high accuracy in the whole positioning process. Thirdly, the A* algorithm is applied in the global path planning, and the second-order Bezier curve is used to realize the path smoothing process, so as to improve the accuracy and efficiency of path planning, and solve the problem that the path planned by the traditional A* algorithm has multiple inflection points and large curvature. In addition, the local path planning is carried out by TEB algorithm to realize the obstacle avoidance function in the real-time environment of unmanned trackless rubber wheeler. Through the joint simulation experiment to test two different path planning algorithms, the results show that the global path after the second-level Bezier curve optimization is smoother at the corner and has smaller curvature, which helps to improve the stability and driving efficiency of the path planning, and the TEB algorithm can quickly plan the obstacle avoidance path, so that the unmanned vehicle can avoid obstacles smoothly. Finally, the underground roadway simulation test scene is built in the laboratory to carry out unmanned experiments, and the results show that the AMCL algorithm can achieve efficient convergence of particles in a short time, and only needs a positioning distance of less than 3 meters to complete the positioning task. The smoothed A* and TEB algorithms can be used to plan smoother and more passable paths, which have no obvious turning at the inflection point, and can quickly avoid obstacles, while completing obstacle avoidance actions to achieve safe driving in the process, which can meet the requirements of underground unmanned driving.