Study On Localization Technique Of Omni-directional Mobile Platform Towards Satellite Assembly

In the process of satellite assembly,the traditional manual installation has been unable to meet the requirements of high efficiency.It is a trend to use the mobile robot to complete the assembly of satellite.The mobile robot localization technology is an important area of robotics,.It is the state estimation of complex systems.It is necessary to fuse the data from various types of sensors.It involves sensors,computers,signal processing and other disciplines.This paper focused on the application of mobile robot localization and the state estimation error,and done a lot of work on nonlinear system filtering algorithm and robot localization.The main research results are as follows:First,the status of mobile robot localization at home and abroad are summarized.The mobile robot kinematics model,coordinate transformation,and commonly used sensor principle are analyzed,and the mathematical model and the noise model of observation sensor during mobile robot localization are set up based on the state space model.It is focused on the influence of noise on robot localization.Secondly,focused on the high nonlinear of mobile robot system,the advantages and disadvantages of particle filter and Kalman filter are studied.An improved particle filter algorithm based on genetic algorithm is proposed which can deal with the degradation problem.Simulation results show that the algorithm has high filtering performance in nonlinear systems.Thirdly,a robot positioning system is constructed based on laser range finder and micro inertial navigation system.The laser rangefinder is fastened in the work area,to measure the absolute location of the mobile robot;the micro inertial navigation system is fastened in the robot,to derive the robot pose information based on the mileage information from inertial sensors such as accelerometer and gyroscope.The predict state of the system is calculated by dual models,kinematics and odometer.Its confidence values are calculated by adaptively robust filter.Particle filter is used to estimate the position.Finally,some experiment is done according to the method.The experimental results show that the algorithm has excellent performance and reliable positioning.It can meet the positioning requirements of omni-directional mobile platform in the process of satellite assembly.