在摆动相中用于下肢外骨骼跟踪人体踝关节轨迹的方法

摘要

为实现外骨骼摆动腿踝关节对人体踝关节运动轨迹的跟踪，建立了外骨骼摆动腿逆运动学模型与卡尔曼滤波预测模型，并在外骨骼摆动腿踝关节处设置人机位姿误差传感器，同时进行了人体步态实验。仿真中以实验测取的人体踝关节相对于其髋关节的运动轨迹作为外骨骼踝关节的跟踪对象，并依此计算外骨骼踝关节处人机位姿误差。采用卡尔曼滤波器对外骨骼摆动腿的髋、膝关节位置进行预测，再利用外骨骼逆运动学模型对预测结果进行校正。结果表明：该方法能良好地实现外骨骼摆动腿踝关节对人体踝关节的运动轨迹跟踪，并且卡尔曼滤波预测能明显改善跟踪的滞后性，同时仅需检测外骨骼踝关节处的人机位姿误差也有助于简化人机之间的传感器布置。%To achieve the tracking of the exoskeleton's ankle joint motion trajectory to the human's in the swing phase of a gait,an inverse kinematics model of the exoskeleton's swing leg and a Kal-man filtering prediction model are established.Human-machine posture error sensors are set on the exoskeleton's ankle joints.Meanwhile,human gait experiments are carried out.In the simulation, the trajectory data of the human's ankle joint relative to the hip joint obtained in the experiments are taken as the reference trajectory of the exoskeleton's ankle joint and the calculation basis of the hu-man-machine posture error at the exoskeleton's ankle joint.The Kalman filter is employed to predict the angular positions of the exoskeleton's hip and knee joints.Then,the results of inverse kinematics of the exoskeleton are utilized to correct the prediction.The simulation results indicate that this meth-od can well achieve the tracking of the exoskeleton's ankle joint motion trajectory to the human's in the swing phase of a gait.The Kalman filtering prediction model can significantly decrease the track-ing lag.And the demand of only detecting the human-machine posture error at the exoskeleton's an-kle joint can simplify sensor arrangement between the human and the exoskeleton.