SLAM Application Of Mobile Robot And Its Algorithm Research Based On Multi-sensor Information Fusion

With the unceasing improvement of the intelligence level of mobile robot,it has played an increasingly important role in household cleaning,medical rehabilitation,and new multimedia entertainments,and other fields.Mobile robot,usually working in relatively complex indoor environment,obtains the surrounding environment by multiple sensors to do map construction,positioning and autonomous navigation task,namely,SLAM(Simultaneous Localization and Mapping).Therefore,multi-sensor information fusion technology has become an important aspect in the SLAM study of mobile robot,directly related to the function of robot and intelligence level.This article,starting with the angle of multi-sensor information fusion,makes further study to map construction,positioning and autonomous navigation of mobile robot and other related theories and algorithms,and then combines with the powerful open-source robot operating system(ROS)to realize map building,positioning and autonomous navigation of mobile robot,or other functions.First,the basic theories of multi-sensor information fusion algorithm are studied.According to JNPF-4WD Pingfang mobile robot platform,working principles and characteristics of several sensors are studied.Basing on extended kalman filter(EKF)algorithm and Rao-Blackwellized particle filter(RBPF)algorithm,the RBPF map building from the observation model matching algorithm is to be the key.Second,localization algorithms are studied.Combining with JNPF-4WD platform,the keys,the inertial navigation prediction positioning algorithm and 2D-laser-radar maps based on MCL(Monte Carlo Localization)algorithm,are studied.Then the localization algorithm of mobile robot based on wireless signal strength is improved to classical Rayleigh damping function model for the improved trilateral positioning and the fingerprint identification algorithm based on fuzzy reasoning through Gaussian filter and Lagrange interpolation.Their positioning accuracy,advantages,disadvantages and their feasibilities are analyzed for applications in the mobile platform to provide theoretical basis.Third,path planning algorithm is studied.Path planning,including global path planning and local path planning,could realize path planning and autonomous navigation of mobile robot through their both combination.Therefore,the traditional field method is improved.To address this problem of local minimum points,the distance factor and rotating the repulsive force field are consided.The simulation results have shown that the improved algorithm can address the problem.Finally,the related experiments of SLAM on the JNPF-4WD robot platform are carried out.The platform,including control system hardware platform,software system and other related configurations about sensors,ROS nodes,gmapping,coordinate,has been studied for map building and autonomous navigation.In this paper,positioning algorithm of mobile robot based on the wireless signal strength and path planning algorithm based on improved artificial potential field method are studied and proposed.And,the keys in the research of mobile robot SLAM,including map construction,positioning and autonomous navigation,have been addressed.The results have shown that it can achieve reliable accuracy requirement,and that path planning algorithm can complete the autonomous obstacle avoidance and path planning tasks.Meanwhile,the intelligence of mobile platform is realized on the JNPF-4WD mobile robot of ROS platform depending on the in-depth study of multi-sensor information fusion.Therefore,this article focuses on the theories and algorithms research of SLAM problem,with much attention to the exploration of practical application,and has certain reference value for scientific research.