Research On SLAM Algorithms Of LiDAR For Micro Multi-Rotor UAV

With the development of autonomous flight in the MUAV without GNSS signal environment,based on the overview of the key technology of the laser radar SLAM and the research results of MUAV,the influence of the SLAM algorithm complexity and the MUAV platform attitude on the SLAM algorithm and the low cost two-dimensional scanning laser radar are given on the basis of the overview of the key technology of the laser radar SLAM and the research results of the unknown indoor environment.The key problems,such as the SLAM implementation algorithm,are carried out.For the problem of implementing SLAM by MUAV equipped with a low-cost 2D laser radar,the geometric feature correlation matching-SLAM algorithm with lower computational complexity is studied.The method of combining the extreme value with the best value for regional segmentation improves the accuracy of region segmentation and lays a foundation for further feature extraction.Increase the corner type to detect corner features,improve the correctness of detection corner features and increase the number of corner points,so that feature matching is more accurate.The low-cost two-dimensional laser radar data characteristics are the key to the accuracy of the SLAM algorithm.This paper analyzes the error of obtaining environmental data.The experimental results show that the use of low-cost 2D laser radar using geometric feature-SLAM algorithm can achieve the positioning of the structured indoor environment and the construction of the environment contour map.In order to solve the problem of feature extraction error and feature matching failure in the SLAM algorithm caused by the attitude change of the MUAV platform,a SLAM combination algorithm based on ICP for the micro inertial system is presented.Among them,the geometric feature-SLAM algorithm is used to provide the initial pose for the ICP algorithm and reduce the number of iterations of the ICP algorithm.Using the flip angle and pitch angle provided by the micro-inertia system to compensate for posture changes results in non-horizontal environmental features and reduces feature extraction errors.The heading angle of the SLAM is compensated by the heading angle provided by the micro-inertial system to reduce the positioning error.Using signature database retrieval mechanism to guide and match the feature number to solve the problem of feature matching failure.In order to verify the effectiveness of the optimized ICP-SLAM algorithm,an experimental test platform equipped with RPLIDAR-A1 and AHRS on a dual-axis electric turntable was set up to acquire environmental information and attitude information.The above-mentioned combined algorithm was tested dynamically and globally.Build the experiment.The experiment verifies the feasibility and effectiveness of the above-mentioned ICP-based SLAM combinatorial algorithm.