Research On Pose Optimization And Control Of Wheeled-legged Robot In Uneven Terrain

Currently,common quadruped robots mainly adopt legged and wheeled structures.Among them,legged robots have flexible mobility and can adapt to complex terrain environments,which allows them to play an important role in specific environments.However,legged robots face issues of motion instability and slow movement speed.Wheeled robots,on the other hand,have fast movement speed and high efficiency,but pure wheeled structures have certain disadvantages in adapting to rugged terrains.Combining the advantages and disadvantages of legged and wheeled robots,this paper proposes a quadruped wheeled-legged robot that combines wheels and legs.The paper conducts relevant research on the balance control and posture optimization algorithms of the quadruped wheeled-legged robot,and divides the main research content into four aspects: kinematic modeling,balance control motion,posture optimization and control,and simulation analysis,providing a systematic discussion.First,the structural design of classical quadruped robots was analyzed,and a physical model of the quadruped wheeled-legged robot was built based on the principles of bionics design.The D-H modeling of the quadruped wheeled-leg robot legs was performed,and the forward and inverse kinematics were derived.The relationship between the leg joint angles,wheel center position,and wheel contact point position was established,and the variation of the leg workspace was obtained using the Monte Carlo method.The corresponding control strategy was established for stable motion of the quadruped wheeled-legged robot.The balance control uses the center of mass PD control to solve the expected acceleration of the body,and calculates the optimal solution of the contact force between the wheel and the ground through quadratic programming.The contact force of the wheel is converted into joint torque through the calculation of the Jacobian matrix.The rolling control obtains the torque of the wheel movement by using feedback control according to the expected center of mass velocity in the forward direction of the robot,and then controls the movement of the wheel.Then,a control algorithm for posture optimization of the quadruped wheeled-legged robot was established.In the case of known terrain information,the demand for utilizing the traction of wheels to cross the terrain imposes higher requirements on the posture control.In order to satisfy the constraints of stability during robot motion and no collision with the terrain,the optimal problem is transformed into a non-linear programming(NLP)problem.The non-linear programming solver is used to solve for the optimal joint angles and body pitch angle during motion,which are respectively input into the balance controller and joint PD controller to complete the coupling control.Finally,a series of simulation tests were conducted using the Gazebo simulation platform.These tests included simulation of body posture balance control,squatting experiments,straight walking,climbing,and steep slope walking.The simulation results showed that the four-legged wheeled-legged robot had stable motion,indicating that the posture optimization and control algorithms proposed in this paper have some feasibility.