Research On Wheeled Mobile Robot System And Obstacle Avoidance Technology

The key technology of wheeled mobile robot includs controller designs,sensor and data fusion technology,motion control,obstacle avoidance path planning,etc.,which has been closely watched by domestic and foreign researchers.Therefore,this thesis is oriented to warehousing applications,designing and implementing a wheeled AGV system,and studying its obstacle avoidance path planning,which has practical application value.The thesis is oriented to the application of warehousing,clarifies the main design goals of the wheeled AGV,combines the advantages of various sensors,designs the system framework and hardware framework of the AGV based on multi-sensors such as lidar and ultrasound,and designs the AGV hardware controller schematic and PCB,and according to their higher requirements for the stability of the controller in the warehouse environment,adopt a hierarchical modular design,optimize the layout and wiring of the interface,components that are easily interfered,and thus improve the controller stability.Ontology design and implementation laid the foundation for subsequent research.The multi-sensor information fusion method is studied.In view of the shortcomings of using gyro or accelerometer alone to produce larger errors,the Kalman filter is used to effectively fuse the sensor data of gyro and accelerometer,which improves the accuracy of AGV angle estimation.The use of sliding filtering algorithm solves the problem of large encoder noise and improves the accuracy of speed estimation.The extended Kalman filter is used to fuse accelerometer,gyroscope and encoder data,which solves the problem of cumulative error generated by the encoder and improves the accuracy of the estimated pose.The problem of obstacle avoidance path planning is studied.The improved dynamic window algorithm is used as the local obstacle avoidance algorithm of AGV.By introducing the distance information between the current speed of the wheeled mobile robot and the obstacle in the calculation of the speed weight of the dynamic window evaluation function,Improves the obstacle avoidance safety of AGV when the distance between obstacles is small and the efficiency of obstacle avoidance when the distance between obstacles is large.In order to solve the problem that the staff may suddenly showing cause the algorithm to be too late to plan the obstacle avoidance path and cause collision accidents in the actual warehouse use,an improved scheme combining the top dynamic window algorithm and the bottom control is proposed to solve the emergency obstacle avoidance.In view of the blind spot of the lidar sensor due to the installation height limitation,it is impossible to find a low object and cause a collision.A barrier avoidance method that combines an ultrasonic sensor and a laser sensor is proposed to coordinate control between the top and bottom programs that are used in the local obstacle avoidance of AGV that improves the safety of the AGV system during operation.An experimental environment was constructed in the experiment,and a series of experimental tests and analyses were conducted on the AGV independently developed,including small obstacle recognition experiment,emergency obstacle avoidance experiment,obstacle avoidance experiment in single obstacle and multiple obstacle scenarios.The results show that the designed wheeled AGV runs stably in a dynamic environment and it can plan obstacle avoidance paths in real time and reach the target point safely.Experiments also verify the improved local obstacle avoidance scheme based on dynamic window algorithm proposed in the thesis and that are feasibility and safety.