Generation Of Biped Robot's Walking Pattern And Its Following Research Base On Human Walking Trajectory

Driven by the rapid development of key scientific and technological fields such as machinery,control and artificial intelligence,more and more service-oriented robots appear in daily life scenarios.In reality,robots are often required to follow and provide services for humans,which puts forward higher requirements for the ability of robots to adapt to the environment and low energy consumption in walking.The biped walking method adopted by human beings has good stability and coordination after a long period of natural evolution.In this paper,human walking data are collected by motion capture system,human walking trajectory is analyzed,robot walking patterns are generated,human walking parameters are extracted,and biped robot following walking is realized.In this paper,the missing data are supplemented by linear interpolation and geometric solution after the human walking capture experiment.The noise data are removed by the monotonic mutation at the noise point,and the human walking parameters are extracted by the ankle trajectory.From the processed walking data,the trajectory of human body's center of mass and ankle joint under the condition of walking on flat ground,ascending stairs and descending stairs is obtained,and the regular patterns of velocity and acceleration conform to sinusoidal function are analyzed,and the data fitting validation is carried out.Based on the time symmetry parameters of walking gait,the space symmetry parameters are defined by using the walking trajectory of left and right ankle joints,and the walking state of human body is judged.The human body is simplified as a single centroid inverted pendulum model by using the connection between the centroid and the ankle joint.The ZMP trajectory of the human body is always in the supporting area during the course of walking.Similarly,the robot is simplified as a single centroid inverted pendulum,ZMP trajectory is planned,the velocity or acceleration of the centroid and ankle joint is planned by sine function,and the walking pattern of the robot is obtained by inverse kinematics solution.Based on the analysis of the walking trajectory variation regular pattern of human body to generate the biped robot walking pattern,a smooth switching pattern generation method for the robot shifting and gait in the following process is proposed.During the robot variable step and flat to upper stair switching,the joint There are no abrupt change in angular displacement,angular velocity,and angular acceleration.In order to ensure the stability of the robot and reduce the energy consumption,a walking pattern optimization method is proposed.The center of mass drop height and the variable ZMP region coefficient are selected as the optimization parameters during walking.The validity of the parameter optimization is verified by a set of walking patterns.Finally,the virtual prototype of biped robot is built in ADAMS to verify the variable step size of biped robot on the flat ground and the switching gait walking from the flat ground to the stairs.According to the collected human walking data,the walking parameters are extracted,and the walking patterns of biped robot are generated,and the simulation is carried out.The simulation results show that the following walking patterns generated based on human walking regular pattern can realize the following walking of biped robots.