Research On Path Planning Of Home Mobile Service Robot

With the development of robot technology, the robot began to penetrate from thetraditional industrial field to the service field. Major developed countries in the worldregard service robot as a priority in important technologies. As an important branch ofrobot, the home service robot is mainly to help people do some housework, which canmake people’s life more convenient. The premise of capability of service robotautonomous mobile is path planning, because path planning has been one of the keytechnologies of research on mobile robot. With the progress of the study, many researchresults of the global path planning have been obtained, but local path planning whichdependent on sensor model still remains to break through. In this paper, we focus onpath planning technique, and set up a simple robot as the hardware platform. Under thepremise of a known environment, global path planning algorithm is used to obtain theoptimal path, and in allusion to an unknown environment the local path planning is usedto avoid obstacles. In this paper, the development status and trends of service robots athome and abroad are introduced firstly, and the key technologies of robotics researchare summarized, and then three issues (Environment Modeling, Localization, PlanningAlgorithm)involved in path planning are researched respectively, in which the focus ofresearch is Planning Algorithm.With regard to environment modeling the specific procedure of grid method iselaborated and the local environment modeling studies on the hardware platform ismade The measurement error of the ultrasonic sensor is analyzed according to the dataof platform, and obstacle model is set up through simulation, and the simulation resultsshow that this method can basically show the obstacle position.With regard to autonomous localization, dead reckoning is used in this paper. Inallusion to not suitable for long-distance positioning for dead reckoning, monocularvision is used to adjust it. The positioning method is presented through monocularvision identifying the color signpost, and the monocular visual position principle isanalyzed, as well as localization experiment is conducted on the hardware platform. Atlast, the measurement method of camera parameters and specific positioning steps iselaborated.In this paper, we use genetic algorithm for global path planning, introduce the basicprinciples of genetic algorithm, and elaborate specific planning procedures. In allusion to the defect of having poor searching ability and slow rate of convergence for geneticalgorithm, we add a wall following strategy to the process of generating initialpopulation, analyze the effect of mutation rate on genetic algorithm, and set up a kind ofmutation rate which decreases with evolution generation. The simulation results showthat this method can solve the problem of global path planning effectively.The artificial potential field method which has a stronger capacity of real timeplanning is used for local path planning. The defect of the artificial potential fieldmethod is analyzed, and the improved method is put forward against the defects.Computer simulation indicates that this algorithm has better planning ability.