Motion Planning And Control Strategy Of A Wheel-legged Hexapod Robot “Long Xiang” Using Centroid Momentum Dynamics

Mobile robots can replace human beings to enter dangerous and complex environments to perform search and rescue,patrol inspection,maintenance and other operation tasks.The wheel-legged mobile robot,which combines the wheel motion mode with high speed,low energy consumption and high efficiency in a structural environment and the leg motion mode with greater mobility and stability in an unstructured environment,has always been a research hotspot in the field of mobile robot.The existing wheel-legged mobile robots face great difficulties in mechanism optimization,system modeling,motion generation and whole-body force control,and cannot fully coordinate the advantages of the combination of wheeled motion and legged motion.This paper aims to propose a wheel-legged hybrid hexapod robot “Long Xiang”,and study its dimension synthesis,centroid dynamics,motion planning and control strategy,so as to greatly improve its moving speed,energy efficiency,and trafficability on rough roads,and enhance the carrying ability and stiffness.The main research contents of this paper are summarized as follows:Firstly,according to the design requirements of wheel-legged robots in complex terrain and limited space,a vertical hybrid leg mechanism configuration is proposed.Taking this configuration as the leg structure,a novel wheel-legged hexapod robot “Long Xiang” with the middle legs that can turn up is designed.The six legs of the robot Long Xiang are symmetrically distributed on both sides of the body,and a driving wheel is installed at the end or in the middle of each leg,which ensures the stability of the robot movement from the structure.Two turnover mechanisms are installed in the middle of the body,so that the two middle legs can be turned over the body,which is convenient to enter the limited space such as pipeline to perform operation tasks.Secondly,the U&2RUS leg mechanism of the robot Long Xiang is modeled.The U&2RUS mechanism is decoupled into two four-bar mechanisms perpendicular to each other,and the accurate and concise forward and inverse kinematics solutions are given.The singularity of the mechanism is analyzed,and the motion/force transmission performance is explored,which provides a basis for the design and optimization of the robot parameters and workspace.In order to obtain the robot with superior performance,the dimension optimization domain of the U&2RUS mechanism is determined using the performancechart based design methodology(PCb DM).Then the optimal dimension combination in the optimization region is found using the trajectory optimization dimension design geometry method.The robot Long Xiang is simplified as a moving 6-SPU parallel mechanism,and the stable workspace of the robot is determined by solving the workspace and the supporting polygon of the parallel mechanism.Thirdly,the velocity and momentum expressions of the robot Long Xiang's centroid are derived using the centroid kinematics and momentum theories,and the centroid mass matrix and the centroid momentum matrix are derived.Then combined with the generalized dynamics model,whole body kinematics and stability theory of the wheel-legged robot,the centroid dynamics model of the robot Long Xiang is established.The centroid dynamics equation in the joint space and the dynamics equation in the operation space based on dynamically consistent generalized inverse are established.An effective recursive algorithm of centroid momentum is developed to calculate the centroid momentum matrix,centroidal composite rigid body inertia,mass displacement matrix,and constrained centroid momentum matrix.Aiming at the actual situation of Long Xiang,a wheel-ground contact geometry model with contact constraints is established to capture the dynamics characteristics of the robot's foot-end in contact with the ground.Combined with the centroid characteristics of the body and each leg,the centroid momentum and dynamics models between the body and each leg are established.By using the contact Jacobian matrix,the dynamic equation of the system is projected into the contact task space,the contact consistent generalized inverse Jacobian matrix is obtained,and control laws in contactconstraint-task consistent null-space is derived.The proposed the kinematics model,wheel center motion model,robot centroid momentum and dynamics model,and layered control model lay the foundation for the whole-body motion control of the robot.Then,using the basic PD controller and zero moment point(ZMP)controller of the wheel-legged robot,combined with the kinematics and centroid momentum and dynamics models,the wheeled motion control model is derived,which is decomposed into three parts related to the body motion,the leg motion and the wheel motion.In order to the coordinated and stable motion control of the robot,the dynamic characteristics of the robot's foot end in contact with the ground are captured based on the contact model,the mathematical model of the terrain estimation is established,and a 3+3 gait motion planning optimizer is designed.The motion planning optimizer relies on the online ZMP optimization,which constantly update the the reference trajectories of the floating base and the wheel-leg.Considering the nonholonomic constraints introduced by wheels,a hierarchical whole-body controller is used to track the optimized motion planning.In order to improve the effectiveness and robustness of the control,the leg movement of the robot Long Xiang follows the biological inspired hierarchical control method,and the optimal feedback control strategy is established.Finally,in order to verify the above theories and methods,an experimental prototype is developed,and the data and results obtained by the robot Long Xiang in the experiment are analyzed.Firstly,the motion stability of the two wheel-leg assembly modes is compared.The robot with wheels in the middle of the leg is selected for testing.Secondly,an indoor slope movement test is carried out to the rationality of slope stability,while the power consumption analysis and endurance test are used to test the running performance and endurance heat dissipation capacity of the robot.Third,the basic motion test verifies the attitude stability and motion performance of longitudinal and lateral motion,and corroborated the effectiveness of walking and wheeled motion control.Fourthly,the comprehensive effectiveness of the proposed control method is verified by an outdoor micro-rugged terrain test.