Research On Autonomous Navigation Path Planning Algorithm Based On Multi-sensor Information Fusion

The accuracy of autonomous navigation is the key index to measure the level of robot intelligence,so improving the real-time and accuracy of related technologies has become the research focus for intelligent robots to realize autonomous navigation.Aiming at the limitations of constructing environment maps by a single sensor and the problems of low efficiency of path planning and high time and space complexity of traditional A~* algorithm,this thesis mainly improves the following aspects:(1)Aiming at the problems of lack of laser point cloud information and motion distortion in the construction of map by single 2D lidar,low accuracy and prone to positioning deviation in the construction of global environment map by single depth camera,a low-cost and high-precision multi-sensor information fusion system is constructed in this thesis.First,make use of the advantage that the depth camera can obtain three-dimensional depth information to make up for the lack of lidar point cloud;Second,the use of linear interpolation to solve the problem of missing information generated by 2D lidar due to low resolution and robot’s own occlusion.Thirdly,the error factor algorithm based on density peak clustering is constructed,and the data of the two sensors are fused to solve the problem of lidar motion distortion,and the depth camera can only scan a fixed angle,so the accuracy of constructing the global map is low.Experimental simulation results show that based on the above three improvements,the completeness of the map model constructed by using the fusion of the two sensor data and the accuracy of the data have been improved to a large extent.(2)The traditional A~* algorithm has the problem of low efficiency of path planning and cannot guarantee to get the optimal path in the environment with large and complex search range.Based on the multi-sensor information fusion to construct the environment map,the traditional A~* algorithm is improved by introducing the cost scaling factor and improving the heuristic function.The simulation results show that the scheme can effectively improve the efficiency of path planning,while ensuring the possibility of obtaining the optimal path.(3)On the basis of using multi-sensor information fusion to construct the environment map,route planning using the improved A~* algorithm of(2)still has the problem of increasing data volume.Improving the efficiency of route planning and speed up the selection of optimal nodes,the binomial heap is introduced as the basic idea of complete sorting,and the open list is optimized,with the top of the heap being the minimum F-value point,which can ensure the orderliness of the data structure while reducing the time and space complexity.Finally,experimental validation is conducted based on the above three parts of the improvement,and the experimental results of the traditional A~* algorithm for singlesensor constructed maps,the traditional A~* algorithm for multi-sensor information fusion constructed maps and the improved A~* algorithm for multi-sensor information fusion constructed maps are compared under the same environment,and the results of multiple experiments are statistically analyzed.The experimental results show that the introduction of the bifurcated heap improved A~* algorithm on the basis of multi-sensor information fusion to construct the environment map can effectively improve the efficiency of path planning while ensuring the real-time and reliability of obtaining the optimal path.