Research On Localization And Path Planning For Mobile Robots Based On GPS/INS

In recent decades, with its rapid development and wide application in industry,service industry, military, outer space exploration and so on, mobile robotictechnology promotes social advance of science and technology enormously.Localization and path planning, as key technology of mobile robots, is one of theimportant research focuses, and it has received tremendous attention. Localization ofmobile robots needs internal and external sensors to obtain accurate position ofworking environment. Path planning, which means robots search the optimal orsuboptimal route according a certain performance index, is also an important link ofnavigation research. Therefore, based on issues of localization and path planning formobile robots, this thesis mainly investigates information fusion for GPS/INSintegrated positioning system and path planning under complex environment model.Moreover, in order to meet practical needs, the GPS/INS integratd positioningsystem is designed.In order to meet the requirements of precision and stability of integratedpositioning system, a information fusion algorithm, which is based on an improvedSage-Husa adaptive filter, is presented. The time-varying fading factor is introduced,when calculating forecasting error covariance matrix, to improve filteringadaptability. Meanwhile, the innovation threshold is used in calculation ofmeasurement noise covariance matrix, which is weighted then influences filteringplus matrix dynamically to judge innovation. Therefore the filtering process couldreduce estimatation bias adaptively, strengthen the effect of the new observations,improve the filtering accuracy, and restrain the divergence effectively. Thesimulation results show that the improved Sage-Husa adaptive filtering algorithm,which is compared with the performance of simplified Sage-Husa adaptive filter andH-infinity filter, makes the process better both in localization accuracy androbustness.According to the visualization requirement for mobile robots localization,integrated positioning system is designed and developed with GPS/INS integratednavigation system. Function modules of serial communication, data analysis, loadingand operation of electronic map, real-time positioning and route display developed by MapX control are included in the software system. Using this system, one canobtain the position and trace of mobile robot. The established scheme isexperimented on the mobile P3-DX and excellent performances are obtained.A global path planning approach of mobile robot based on improved ant colonyalgorithm is proposed to solve the problem of trapping into local minimum, in which,a grid map is used to model the environment. The pheromone threshold is limited toavoid the local optimum, and the parameters are adjusted dynamically in thisimproved ant colony algorithm. Based on the improved algorithm, the pathgenerated by the elitist ant is checked to find the reason of local minima. And then,by embedding a sub-ant colony algorithm into the route choice model, some localroutes can be optimized and updated. Simulation results indicate that this approachcan solve the U-shaped groove problem, and reduce the probability of trapping intothe local minimum. Compared with several path planning approaches based on antcolony algorithm, the proposed one could generate a shorter path which is globaloptimal and have the capability of searching path in complicated environments.