Research On Dynamic Modeling And Low Level Motion Control For Omni-Directional Mobile Robots

The low level motion control of the omni-directional mobile robot is the foundation of trajectory tracking for autonomous mobile robots.It is of great significance for the stability and precise control of omnidirectional mobile platforms.In RoboCup middle size league(MSL),the performance of the omni-directional mobile robot low level motion control,directly affects the implementation of some accurate behavior,such as fixed-point movement and passing over the ball.In this paper,dynamic model based on hierarchical control structure is proposed,in reference to the hierarchical control based on kinematics model.It will not only facilitate modular design software and the improvement of the upper control algorithm,but also simplify the dynamic model.The research is based on the NuBot soccer robot.The NuBot soccer robot has four omni-directional wheels,which are evenly cross distributed on the chassis.A kinematic model without sliding is proposed,based on the geometry analysis for the wheel distribution and assembly way on the platform.Then according to the analysis of the forces on the omni-directional wheel,robot's dynamic model is built up.Finally,the parameters of the model are identified based on the designed experiments and data analysis.The motion control structure of the NuBot soccer robot,at present,is hierarchical control based on kinematics model.And the outputs of the low level motion controller are the speeds of four wheels,where the motor speed control is accomplished by the Elmo digital servo driver.In the actual game,the collision between robots usually lead to motor stall.Then the servo driver will automatically disconnect the motor power supply due to the driver protection scheme.In such situation,the robot has to drop out the game since the robot motion will out of control.In this paper,motor speed controller based on gain scheduling PID algorithm is designed,and the output of the controller is current command.The controller will not only avoid the servo driver disconnecting the motor power supply,but also responds faster to the changes in all speed period and makes the system more stable.For the hierarchical control based on kinematics,without considering the specific dynamics,it tends to be relatively simple to design the motion controller,and has good code portability.However,the controller cannot operate real time since the motor speed controller shall spend some time to reach the target speed command.This paper proposes a new hierarchical control scheme which is based on dynamic model,which will not only make the motion controller operate real-time,but also avoid the complicated tuning of motor speed controller.Finally,in order to solve the wheels sliding problem of the four wheel omni-directional robot,a closed loop speed tracking controller based on control allocation of motors moment is designed.The experimental results validate the effectiveness the validity of the dynamic model and the feasibility of the two different speed tracking methods.Good dynamic performance and small tracking error can be achieved by the controller.