Motion Optimization For Hydraulic Quadruped Robot Leg

As Robotics and Artificial Intelligence advancing by leaps and bounds,new ideas come into reality such as biometic robot,drone smart vehicle.Among these robots,the most representative is undoubtedly the hydraulic quadruped robot designed in the United State.It has excellent dynamic performance,high load capacity,good stability and environmental adaptability,attracting both ordinary people and researchers.The research of hydraulic quadruped robot involves the process of perception,planning,decision-making and motion.The motion of robot is an important research direction,including gait planning,inverse kinematics,compliance control,swing leg retraction(SLR),self balancing and other issues.This paper focus on motion optimization of quadruped robot leg.After modeling the virtual robot,inverse kinematics and swing leg retraction is investigated,exploring methods to optimize movement effect and enhance athletic perfoemance.In modeling the hydraulic quadruped robot,the mechanical and kinetic characteristics are analyzed and mathematical model is established.Then the dynamic model of extended spring loaded inverted pendulum of the robot is made according to the actual parameters and its feasibility is studied.On the problem of robot inverse kinematics,the redundant degree of freedom is considered.Different constraints are compared among which the minimal norm of joint angle is chosen.Different algorithms such as gradient projection method,genetic algorithm,particle swarm optimization(PSO)are also compared.Finally PSO is selected and improved in three aspects: initial range,compression factor and interpolation,which makes it more real-time and suitable for this inverse kinematics problem.As to swing leg retraction,the impact on impulse energy loss,friction and total impulse are analyzed.In the simulation software,extended spring loaded inverted pendulum(SLIP)model is used to simulate the relation between swing leg retraction rate and energy loss,friction and total impulse.Empirical data is acquired in the simulation.After that,multi objective is converted into single objective optimization problem and the optimal SLR rate of the hydraulic quadeuped robot is calculated.Then the relation between trajectory planning and SLR rate is analyzed and the normalized SLR rate of specific bezier trajectory is deduced.Finally a trajectory which is based on bezier line and whose SLR rate is optimalin this study is designed.