Research On Indoor Navigation Technology Of Service Robot Based On Unscented FastSLAM

The service robot industry is helping all walks of life to achieve industrial iterative upgrading.The modern mobile service robot realizes the sensor system to sense and observe the dynamics and uncertainty in the environment through multi-sensor integration and fusion.To complete this complex system process,simultaneous localization and mapping technology is indispensable.The safety,shortest path and smoothness of navigation are the key points of research in related fields.Therefore,it is of great significance to study the autonomous navigation of service robots.The main work of this thesis is as follows:Firstly,by studying the development and current situation of service robots at home and abroad,this thesis lists the achievements and shortcomings in the field of service robots in China.The key technologies of the mobile service robot are analyzed.The environment map construction,location and path planning of the mobile service robot are studied.The location and navigation scheme of the service robot based on slam technology is designed.Secondly,in the process of map building,the particle filter based on Fast SLAM algorithm has the problem of large amount of work to calculate and deduce Jacobian matrix and particle dilution,which leads to further increase of map building and positioning errors in the later stage.In this thesis,we use unscented Fast SLAM algorithm to overcome the shortcomings of the Fast SLAM framework by directly using the non-linear relationship of scale unscented transformation,and improve the resampling process by combining the idea of genetic algorithm to improve the consistency of filtering and the accuracy of state estimation.The experimental results show that the improved algorithm has the advantages of higher accuracy,less computation and less system error when building maps in different environments.Thirdly,aiming at the problems of premature convergence and poor local optimization ability of PSO,a multi-objective grasshopper optimization algorithm based on PSO is proposed,and a mathematical model of multi-objective optimization problem with the goal of path length,smoothness and security is established.In the process of locust population search,for the global search and the local search need different search rates,this thesis introduces the cosine function adaptive strategy to balance the searchrate of the algorithm in different periods.The comparative experimental results show that mogoa has better convergence and local search ability than MOPSO,and the path length is reduced by about 2.01%.Finally,taking the mobile service robot developed by Chongqing University of Posts and Telecommunications as the experimental platform,the navigation scheme of indoor service robot based on Ufastslam algorithm is designed,and the multi-objective locust optimization algorithm is proposed to be used in the path planning of mobile service robot.The feasibility and feasibility of the system are verified by experiments.It has obvious advantages in positioning error,map construction and path planning.