Research On Map Construction And Positioning Of Navigation Robots

The research on autonomous navigation of navigation robot in unknown environment includes the construction of environmental map,autonomous positioning,path planning and dynamic obstacle avoidance,etc.Environmental map building and autonomous positioning are the most basic research aspects,which are the prerequisite for the realization of path planning and dynamic obstacle avoidance.So this article has carried on deep research and discussion from two aspects.The main contents of the work are as follows:The mechanical and control system of the navigation robot is designed,and the prototype is developed.In order to reduce the influence of noise points and smooth environmental data in two-dimensional lidar data,median filter and mean filter algorithm are studied,and the comparative experimental analysis of the filtering effect is carried out.In order to reduce the number of iterations of feature extraction and speed up the data processing,the segmentation of laser radar data is needed before the feature extraction.In this paper,data characteristics and the measurement principle of laser radar is studied.An adaptive lidar data segmentation algorithm based on environment feature is proposed.The threshold is determined by the intersection points of virtual environment line fitted by the adjacent point of the data and the adjacent laser scanning ray,and the adaptive threshold model is established.The judgment method of pseudo-breakpoint is proposed,and the algorithm is optimized.Data segmentation experiments under multiple environmental features are carried out,and the algorithm is compared and analyzed with segmented threshold segmentation algorithm and linear equation threshold segmentation algorithm.The main research of feature extraction in this topic is the extraction of linear features.First,a linear description model of laser radar data is set up,and the splitting-combining line feature extraction algorithm is studied.In order to reduce the influence of noise point,this paper optimize the splitting-combining method,improve the judging method of splitting point,and conduct a comparative experiment of linear feature extraction and analysis.The experiment conducted through the combination of filtering algorithm,adaptive segmentation algorithm based on environmental features and improved splitting-combining method to verify the effecti-veness and practicability of the above algorithm.The particle filter location algorithm based on Monte Carlo method is studied,and the implementation steps of particle filter location algorithm in the navigation robot system are established.Aiming at the problem of particle diversity attenuation caused by polynomial resampling in particle filter location algorithm,a polynomial combination resampling method is proposed,in which high-weight particles are sorted according to weights and then combined them.The experimental environment is set up in MATLAB to simulate the contrast experiment.Experiments show that the polynomial-combined resampling method can effectively increase the particle diversity and improve the positioning accuracy.The ROS operating system is built on the prototype of the navigation robot to realize the motion control function,the data acquisition function of the laser radar,the construction and positioning function of the map of the navigation robot prototype.The experiment of map building in the real environment is carried out,the problems appearing in the experiment are analyzed and the solutions are given.The location experiment is carried out,the mission point is sent to the navigation robot,and then measure the pose error of the navigation robot when it arrived at the task point.The experimental results show that the location accuracy of the improved particle filter algorithm can meet the requirements of indoor navigation robot.