Research On Motion Modelling And Tracking Control Of Six-wheeled Planetary Rover On Soft Terrain

Recently,many countries have carried on series of research on the planetary exploration.Since the planetary rover can work as an effective carrier platform for the scientific instruments,it plays an important role in the planetary exploration.However,since the planetary is covered with a layer of soft terrain,the ideal pure rolling for the planetary rover's wheel is not held any more,which brings new chanllenges for its motion tracking control.Currently,aiming at the motion tracking control for wheeled mobile robots(planetary rover),most research is done under the ideal assumption of wheel's pure rolling,and always ignores the wheel's longitudinal/lateral slippage on the soft terrain.Therefore,aiming at such a kind of soft terrain like planetary soil,under the case of wheel's longitudinal/lateral slippage,this paper researches on the planetary rover's motion trakcing control:In order to keep the balance of the rover's base while the rover is running on a rugged terrain,this paper mainly focuses on a kind of six-wheeled rover with rocker-bogie.In the rover's running process,in order to control the multi-DOF of each motion joint,based on the motion relationship of all related joints,the motion command of the base can be devided into each actuator in order to implement the desired motion,and further the Jacobi model of the rover from the base's velocity to the wheel's driving motors and the steering motors is deduced;on the basis of the motion phenomenon of the wheel's longitudinal/lateral slippage on the soft terrains,the planetary rover 's kinematic model influenced by the wheel's slippage is analyzed and modeled in detail;combining with the traditional terramechanical model and vehicle's kinematical/dynamical theory,the planetary rover 's dynamic model is deduced,which also takes the wheel's longitudinal/lateral slippage into account.On a soft rugged terrain,with the influence of the wheel's longitudinal/lateral slippage,the performance of wheel's trajectory tracking becomes worse,and further the rover's path will deviate from the required one under the traditional controllers.Since the geometry parameters and the mechanical parameters of the soft terrain are always position-varying,the wheel's longitudinal/lateral slippage is also different with each other.Based on the six-wheeled planetary rover 's forward kinematic model,the Jacobi matrix between the referrering point of the base and each actuator(driving motors and steering motors)can be decoupled,then the trace tracking control input on the base can be mapped to the six driving motors and four steering motors;then,this paper proposes a method of training the feedforward neural network's weight based on the unscented Kalman filter,which can predict the planetary rover's slipping velocity and compensate for its motion control;based on the optimal control theory and the planetary rover's model considering wheel's lateral sliding,this paper deduces and obtains the optimal trajectory tracking control algorithm of the planetary rover,which can guarantee the rover can run and succeed following its tasks.Although the given tasks can be done with the automatic trajectory tracking controllers,the risky factors exist everywhere on the planetary surface which may induce the rover 's unsafety,and the command's deviation always brings irretrievable results.Therefore,for the rover 's tele-control,the control mode of teleoperation is absolutely required.In order to tele-drive the planetary rover on the earth,this paper proposes a rover bilateral teleoperation system based on a force feedback of velocit y tracking error,and its absolute stability conditions are obtained by the Llewellyn criteria;in order to compensate for the command tracking error induced by the wheel's slippage phenomenon,a local self compensating algorithm is also proposed for the velocity commands at the slave site.Based on the slippage-sinkage mechanism on the soft terrain for the planetary rover,this paper proposes an optimization algorithm with the PE-TE as the objective to coordinate each wheel's slippage;this paper constructs a coordinated slippage tracking system,while usin g the neural network to converge the system's nonlinear function with unknown parameters;in addition,a compensation controller designed by the sliding mode control is used to improver the system's robustness,and the sustem's stability is guaranteed by the Lyapunov approach.In order to validate the above control algorithms,through developing the electical control system of the planetary rover's sample machine,combining the Novint Falcon haptic device and the motion capture system,this paper develops an in-situ experimental platform,which can simulate the exploration process of the planetary rover on the soft planetary surface;then the proposed theory and controllers in this paper can be validated by this platform.The experimental results show that the proposed motion decomposing method can effectively map the base's velocity commands to the driving motors and steering motors;by the proposed trace tracking controllers with feedforward neural network and the optimal control,the influence of the wheel's longitudinal/lateral can be eliminated,and the planetary can guarantee a good tracking performance;with the proposed semi-automatic teleoperation controllers,the rover can track the commands velocity well,with good performance.In general,the research results of this paper can support many ideas and basis for the trajectory tracking control of the planetary rover;and it can also be seen as a reference for the future planetary exploration projects of China.