本项目面向增能型下肢外骨骼机器人的系统集成，研究下肢外骨骼机器人的本体设计、系统仿真、人体运动意图感知与智能控制，以及建立人机协同下肢外骨骼机器人的评价体系。为实现良好的人机协同性，首先，我们在仿真设计中充分考虑了人体生物学特性，在仿真环境下结合人体肌肉骨骼模型进行人机协同仿真，选取最优设计；其次，在人体下肢运动意图感知上，采用含有人体运动意图的表面肌电信号结合强大的深度学习模型对人体运动意图进行预测，且利用表面肌电和其他传感器信号相结合的多传感器融合技术，对人体自然步态进行分析和预测；最后在控制策略上，将强化学习引入传统的阻抗控制中，以实现在变化的人机环境下的自适应控制。人机协同的增能下肢外骨骼机器人的系统集成研究，在工业制造，军事增能领域具有广泛的应用价值，同时对于增强我国在外骨骼机器人领域的研究积累具有重要的意义。

The main contents of the project include: Design of the lower limb exoskeleton robot, simulation system, perception of human motion intention, adaptive control, evaluation system of the lower limb exoskeleton robot. For a good human-machine integration, firstly, we fully consider lower limb biological characteristics and conduct human-machine cooperative experiments on the exoskeleton with human’s musculoskeletal model in the simulation environment. Secondly, we process surface electromyogram (sEMG) data which contain human motion information with deep learning methods to predict human motion intention and combine sEMG with other data using Multi-sensor Fusion to model and analysis human gait; Finally, when dealing with the control strategy, we introduce reinforce learning in traditional impedance control to make the exoskeleton adapt itself to the changing human-machine environment. The research of the human-machine cooperative lower human-powered augmentation exoskeleton has a good prospect in both the industrial filed and military field because it can expand the limit of human body, and to provide energy assistance for the human.