Kinematic Analysis And Control Study Of A Bionic Deformable Hexapod Robot

In recent years,in the special dangerous areas such as explore star,rescue and relief in the complex terrain,work in the sea floor,field monitoring and military investigation,the perilous task was not suitable for human to accomplish.And people were prompted to research the robot with feet increasingly in order that robots could replace human.This paper studied the bionic hexapod robot whose leg had a deformable joint in the aspects of leg kinematics,dynamics,motion control and the parallel fuselage kinematics.Innovative work of this paper is as follows:(1)By the screw theory and exponential product approach,the kinematic model of this robot's leg was gotten.The explicit solutions of inverse kinematics were deduced.For the intersection point of the last two joints,the solutions of inverse kinematics were infered by Paden-Kahan sub-problems.The Jacobin matrix of the swing leg was gotten by the screw theory.Which was proved correct through simulation.(2)In order to improve the stability and quickness of six-legged robot leg control,the trajectory in joint space was planned by using five times B spline curve.The joint's speed,acceleration and pulse was ensured continuous.By using convex hull of the B spline curve,the curve of the velocity and acceleration was constrained into the scope of B spline curve's control points.(3)When in a three-legged support phase the hexapod robot was a 3-URR parallel mechanism.Its forward kinematics model was established by the screw theory and exponential product method.The inverse kinematics model was gotten by the elimination theory and Paden-Kahan sub-problem method.Which lay the foundation of six-legged robot in parallel dynamics analysis and motion control.(4)The inverse dynamics model of single deformable leg was generated by the Kane approach.which can be applied to design the motion controller of single deformable leg.The inertia moment,the centrifugal moment,coriolis moment,and the gravitational moment among of every joints torque was derived separately,which can help to plan the motion of mechanisms,optimize mechanisms and select appropriate motor.(5)The hybrid controller of computed torque and RBF neural network was designed.The computed torque was a part of the dynamic model.The RBF neural network was used to compensate the model error caused by the unmodeled part,the measure error of the structure parameters and the external disturbance.Simulation and experiment showed that the tracking precision of the controller was higher and its control performance was superior.