Information Processing And System Design Of Mobile Robot In Conformation Scene

Mobile robots integrate many professional contents such as computer science,automatic control theory,robotics,etc.There are many types of sensors and actuators,and the quality of information processing algorithms will directly affect the overall performance of the system.In multi-robot confrontation scenarios,information extraction of multiple targets,high accuracy and stability pose challenges to information processing algorithms,and are currently hotspots in robot research.This thesis mainly studies the information processing and system control design of mobile robots in confrontation scenarios.The main research work is as follows:Firstly,the system synthesis framework is designed,the Mecanum wheel motion model,IMU measurement model,camera imaging model and lidar model are established,and the robot pose estimation and target trajectory tracking prediction problems are described.Secondly,for IMU data processing problems,error calibration and attitude update algorithms are given;for encoder angular velocity estimation,"T" method speed calculation,nonlinear compensation and filtering algorithms are proposed;a fusion estimation algorithm of IMU and odometer based on Kalman filter is proposed,and the wheel slip detection and maximum torque identification algorithms are given by using the filtered residual information.Thirdly,for multi-camera configuration,an image stitching algorithm is given;for the multi-target tracking problem,a data association algorithm based on the probabilistic multiple hypothesis method is given;considering the problem of target loss or addition,a Bayesian trajectory determination algorithm is given;for the target hit point prediction problem,a trajectory smoothing and prediction algorithm based on Kalman filter is designed;considering the temporal and spatial alignment of multisource heterogeneous sensors,a temporal alignment algorithm based on moving average filter and linear interpolation and a spatial alignment algorithm based on quaternion spherical linear interpolation are proposed.Then,considering the simulation problem of multi-robot confrontation scenarios,a comprehensive simulation platform is designed and developed;a model-free parameter self-tuning algorithm is proposed to realize a platform capable of simulating the actual physical environment with a higher degree of reality.Finally,a Mecanum wheel mobile robot system is built,and the effectiveness of the proposed algorithm is demonstrated through experiments.