The Research On Algorithm Of Mobile Robot Simultaneous Localization And Mapping

The robot technology had a very fast development along with the rapid developmentof the microelectronic technology, the computer technology, the information fusiontechnology and the network technology, especially that with walking function,environmental awareness and can independently complete a variety of intelligentautonomous mobile robot research services has become a special study. For the mobilerobot, the autonomous navigation is a prerequisite capability to practical applications, andautonomous navigation is based on the autonomous mobile robot to locate and mapping.Simultaneous Localization and Mapping is a process used by robots to build up a mapaccording to the information measured by sensors with in an unknown environment, andmeanwhile keeping track of their current location. In this paper, a variety of algorithms ofthe Simultaneous Localization and Mapping are studied, and made some improvementsbased on the traditional methods, to improving the performance and accuracy of theSLAM. The major research contents are as follows:Firstly, this paper summarizes the background and significance of the research,Introduces the concept and status of mobile robot SLAM, and establishment of a variety ofmobile robot system model, these systems model are the basis of a platform for SLAMproblem.Secondly, aiming at the problem of mobile robot SLAM, this paper has a depthresearch to the present methods. focuses on the Extended Kalman Filter algorithm, anddiscuss the problems of the Extended Kalman Filter algorithm.Thirdly, aiming at the problem of premature of fuzzy adaptive extended kalman filterSLAM algorithm based on the Geese PSO algorithm, this paper make some improved.The use of fractional calculus to improve the speed of evolution of the particle, chaos toinitialize the problem of processing methods and particle particles occurred whenpremature optimization. Then the improved geese particle swarm algorithm used to thefuzzy adaptive extended Kalman filter localization and mapping algorithm training.Contrast with Geese particle swarm algorithm fuzzy adaptive extended Kalman filtersimultaneous localization and mapping algorithm, the new algorithm positioning and composition has greatly improved.Finally, aiming at the problem of nonlinear model linearized of the extended kalmanfilter algorithm made some improved. The extended Kalman filter nonlinear models aregenerally used to obtain approximate linear model solution, this paper introducesTakagi-Sugeno (T-S) fuzzy model, instead of using pseudo-linear fuzzy nonlinear model,and with forgetting factor adaptively adjust T-S fuzzy Kalman filter algorithm to improvethe accuracy of SLAM algorithms.