Study On Approaches Of Optimal Path Planning And Anti-slipping Trajectory Tracking For Wheeled Mobile Robots

Wheeled mobile robots(WMRs)are widely applied to the transport and exploration tasks in the environments of the ground and the extraterrestrial planets.In order to enhance the ability of autonomously accomplishing these tasks,path planning and trajectory tracking technologies of WMRs have long been study hotspots.On the path planning layer,the current studies generally do not consider the effect of the map construction on the performance of path planning,as well as the limitations from the environment for planning and self's locomotion ability to the actual feasibility of the path.On the trajectory tracking layer,most of the current studies are based on the ideal assumption that the odometer information is accurate and the wheel-ground rolling is ideal without slipping.However,when the WMR with limited locomotion and perception abilities is running in a real complex environment,the previous problems that are not considered will land the planned path in poor actual feasibility and also lead to the instability or failure of the trajectory tracking controller.Therefore,it is necessary to design the reconstruction method of the map information and the optimal path planning method based on the objectives of path planning.It is also necessary to study the odometer calculation method and the anti-slipping trajectory tracking method when there existing non-ideal wheel-ground rolling.These studies are of great significance to enhance the ability of a WMR to independently and reliably accomplish complex locomotion tasks.For the position-constrained optimal path planning problems for WMRs,it is discussed from the global and local layers separately about how to construct the cost map matching the path planning objective and how to establish and solve the optimal path planning problems.Considering commanding the planned global path to possess the optimal traversability in a wide area map as the main objective,the improved dual covariant hamiltonian optimization for motion planning(CHOMP),a high-dimensional pointto-region(PTR)path planning method,is designed in the planning environment which is the global traversability cost grid map constructed based on the satellite elevation map.By using the non-monotone gradient projection algorithm(NGPA),an effective set method with high tolerance,the efficient and high-success-rate path planning while there existing non-convex cost functions is realized.Furthermore,the method is probabilistically improved based on the simulated annealing framework and the success rate of planning is greatly increased by overcoming the local minimum deadlock and divergence.The designed method can achieve paths with high traversability while balancing the low geometric span and the decent smoothness.The results of the comparative simulations demonstrate that the designed method can achieve the previous expected goals of planning.To urge the planned local path to possess not only the high quality,but also the optimal local obstacle avoidance performance considering the actual locomotion ability of a WMR,an optimal local PTR path planning method is designed.By treating the local static and dynamic obstacle cost grid maps as the planning environments,the objective cost function is ameliorated to boost the local planning performance of the improved Dual CHOMP method.And the path searching ability of a WMR in local static and dynamic environments is enhanced by the further combination with the Hamiltonian Monte Carlo(HMC)method.The results of the comparative simulations demonstrate that the designed method constantly drives the initial path out of the local minima while maintaining the performance of smooth obstacleavoiding.Thus,the success rate and efficiency of the planning are guaranteed.To ensure the WMR having the ability of steadily tracking the planned path in a complex environment,the anti-slipping trajectory tracking strategy is studied from two layers,which are improving the accuracy of the odometer information and enhancing the robustness of the trajectory tracking system.To eliminate the influence from the complex wheel-ground interaction to the WMR odometer measurement,a non-contact odometer calculation method is designed based on the measurement information of an IMU and a monocular vision blackbox.Based on the features of IMUs and visual units updating information and the basic framework of an extended Kalman filter(EKF),an IMU-aided camera EKF(ICEKF)algorithm is established,which is a semi-tight information coupling algorithm.By coupling the IMU and visual information on the velocity layer and estimating the displacement scale coefficient of the monocular blackbox,the IMU can aid the vision algorithm in outputting more accurate odometer estimation results.By using this method in an IMU aided monocular vision system mounted by a WMR,a visual inertial odometer(VIO)is established.The experiment results demonstrate that the ICEKF can provide more reliable results of the odometer measurement for a WMR.Focusing on the problem of wheel-ground slipping when a WMR performing trajectory tracking task in a complex environment,a feedforward controller is designed based on the coupling kinematics model of the steering and longitudinal slippage,which can realize the active real-time compensation of the time-varying slip rates.Considering the interferences occuring to the trajectory tracking system owing to the strong wheel-ground coupling,which are hard to be analytically expressed,a sliding mode control(SMC)method is designed based on the fast double power approaching law and the back-stepping method.Furthermore,a new adaptive SMC method is designed by combining the former SMC method with the integral adaptive function,which guarantees the tracking performance of the target trajectory.The results of comparative simulations and experiments,performed when the odometer information of the WMR is known,demonstrate that the output chattering of the tracking control system is significantly suppressed by the designed method,the response ability is improved and the tracking errors are suppressed.Thus,the antislipping trajectory tracking control with strong robustnes is realized.Through designing and performing the joint experiments of local path planning and trajectory tracking with a WMR under the non-ideal rolling condition,the practical feasibility of the designed methods in real environments is validated.