Hip Joint Trajectory Generation Based on Human Limb Motion Synergy

Lower limb exoskeleton robot can help hemiplegic patients with rehabilitation training and enhance human movement ability. Lower limb trajectory generation is an indispensable part of lower limb exoskeleton robot control, which can better realize human-machine collaboration and significantly improve the quality of life of patients. One of the main challenges in this field is to predict the wearer's hip joint trajectory in advance. In this study, we analyzed the synergy law between limbs in the process of human walking, and studied how to use the shoulder motion data to predict the hip joint trajectory based on the synergy law. We recruited four subjects, and measured their three-dimensional coordinate motion trajectories data of 8 markers at the left and right shoulders, elbows, hip joints, and knee joints by the optical motion capture system OptiTrack. The non-linear mapping model between the shoulder and the hip joint is established using the least square (LS) method. An error correction model is also proposed based on long short term memory (LSTM) to reduce the hip joint trajectory prediction error. Based on the proposed synergy law of upper and lower limbs, the hip joint trajectory can be predicted in advance only by using the shoulder motion information. The predicted hip joint trajectory error can reach as low as 2.1°.