Research On Lidar Navigation Algorithm Of Indoor Mobile Robot

The rapid development of robot technology will bring human beings into a new era.In the future,all kinds of service robots will be distributed in thousands of households.Taking indoor mobile robot as the research background,this paper studies the algorithm of each key sub-module when the robot performs indoor navigation tasks.Since indoor mobile robot has strict requirements on positioning and mapping accuracy and real-time performance of the system,this paper starts from the original data collected by sensors to analyze and improve the mainstream algorithm of each module.In order to remove the motion distortion of lidar caused by the rotation of the carrier,the accuracy of location map is improved.Firstly,the mathematical model of lidar is established,and two main methods of distortion removal are analyzed.After establishing the kinematics model of the robot,the accuracy of its position estimation is greatly improved by calibrating the wheel odometer.On this basis,the method of linear interpolation and nonlinear interpolation is used to remove the motion distortion caused by lidar based on the measurement information of odometer.Mainstream laser laser radar is analyzed at the same time,location and building the mechanism of the graph algorithm,and in different size of indoor environment simulated experiment,choose a kind of built figure higher precision and computational complexity is moderate,laser radar graph algorithm positioning and built at the same time,on this basis puts forward an improved positioning and built figure algorithm at the same time,it to the original algorithm to simplify a complex distribution,under the premise that ensure undistorted sampling with normal distribution to replace,the normal distribution given by and mean and variance based on the summation of the simple algorithm.To solve the problem that Montecarlo algorithm sometimes fails in global positioning,random particles are introduced,and the enhanced Montecarlo algorithm based on random particles can accurately identify the location of mobile robot in the environment map in the simulation test.In terms of path planning,the traditional particle swarm optimization algorithm is prone to fall into the local optimal solution,and the computational amount is too large to meet the high real-time requirements.The method of setting pheromone threshold and limiting range is adopted to improve it.Through code transplantation,the performance of each module is verified.Due to the effect of noise in the process of building figure was unobstructed area may appear noise,become impassable in scale map,we introduce the map editing functions to solve,namely can erase originally there is no noise,to expect the robot reach the virtual wall painting line said,edit the map based on the original map,and the modified will become the ultimate navigation map.