SLAM Research And Implementation Of Mobile Robot Based On Laser Range Finder

Autonomous robots applications increasingly widespread. Its related problems haveattracted much attention. Robot positioning and map building is important guarantee ofautonomous performing task. Robot localization and map building depend on each otherin unknown environment. Researchers proposed the simultaneous localization and mapbuilding (SLAM) Problem.This paper performs intense research about SLAM problem in the order fromtheoretical analysis to practical application using extended Kalman filter, mainly includesthe following research:Firstly, mobile robot system model is established aiming at SLAM problem in two-dimension indoor environment, including coordinate system definition, environmentmodel, sensor model, robot motion model and environment dynamic model.Secondly, general EKF solution of SLAM problem is illustrated. The uncertaintytransfer process, landmark correlation process, and convergence of SLAM problem isanalyzed theoretically. Simulation environment of point landmark SLAM problem wasestablished using MATLAB. Simulation results for different task verify the results oftheoretical analysis.Thirdly, considering need of practical application, in view of the line segment oftwo-dimensional environment the landmark extraction method based on RANSACalgorithm is studied, and improved according to the specific application requirementsalgorithm. EKF equation for SLAM problem is established in the environment of linesegment landmarks. Data association algorithm is proposed combined the nearestneighbor method with joint compatibility branch and bound.Finally, simulations and experiments about landmark extraction and EKF-SLAMalgorithm are performed using powerboat mobile robot and related softwares. The results verify the effectiveness of algorithm proposed earlier.