Research On Simultaneous Localization And Indoor Mapping Of Robot Based On Laser Sensor

Nowadays,with the development of science and technology,studies in artificial intelligence mobile robot has become a new great upsurge.The main purpose of this kind of robot is to make it effectively locate itself in an unknown environment Then Simultaneous Localization and Mapping(SLAM)is the basis for achieving the above targets.Meanwhile,it is also the popular research direction these days.As a matter of fact,SLAM technology has important theoretical significance and application value.In this paper,360° two-dimensional planar laser sensor,odometer and IMU are used as the main sensors to obtain data information,and the four-wheel mobile platform built by ourselves is used as an experimental platform.The research,given an indoor environment,aims to study SLAM issue(RBPF-SLAM),which a mobile robot based on Rao-Blackwellized particle filter has to be faced with.The main studies are as follows:(1)Firstly,the paper introduces the kinematic configuration under the relevant coordinate system.Besides,It's researches show the construction of track derivation and odometer motion model.And the paper studies the data transmission format of the laser sensor,reads data and converts it through MATLAB as well.In this paper,the Occupancy Grid Map algorithm is selected as the mapping model.Meanwhile the paper completes the theoretical derivation of Occupancy Grid Map,and the effect of the theory is verified by MATLAB.(2)As for the estimation error increases,which is caused by lack of the sample particle's diversity and can't be avoided throughout the SLAM,the paper will deduce and find out the reasons for it basing on the theory of Bayesian filtering and the particle filter localization algorithm step by step.Furthermore,the paper applies an Adaptive Deterministic System Resampling algorithm to alleviate the problem and reduce the estimation error.MATLAB simulation results the improved algorithm's Effectiveness.(3)Based on this,it also studies the traditional RBPF-SLAM algorithm and improves it by applying the Butterfly group heuristic algorithm to adjust and optimize its probability proposal distribution.Putting forward an improved SLAM algorithm based on Adaptive Deterministic System Resampling method And the MATLAB simulation results demonstrate the effectiveness of the improved algorithm.(4)The mobile platform being the slave and the PC the host,this paper,in reality,applies the Robot Operate System(ROS)framework and operates the original RBPF-SLAM method together with the improved one.The purpose of it is to complete the environment map construction.The result suggests that the improved SLAM algorithm has a better map construction effect than the traditional one.