| China’s huge population and accelerating aging trend have given rise to a huge rehabilitation population of stroke patients and a growing demand for motor rehabilitation,while the level of domestic rehabilitation development is not high and there is a relative shortage of rehabilitation resources per capita.Development of rehabilitation robots is an effective solution to the current rehabilitation problems,especially lower limb rehabilitation robots that help lower limb rehabilitation patients regain walking ability.The current lower limb rehabilitation products still suffer from the problems of little feedback information,insufficient joint motion coordination,lagging control methods and single training mode,etc.This project develops an intelligent lower limb rehabilitation exoskeleton robot system with a high degree of integration,multi-sensing feedback and various functions.The system adopts multimodal information fusion method to provide real-time,accurate and comprehensive information feedback,building active coordination of hip,knee and ankle and multi-degree of freedom active assistance of trunk and center of gravity.By combining multiple adaptive control algorithms to realize diverse personalized rehabilitation training modes,it can improve the efficiency and effectiveness of lower limb rehabilitation and help lower limb disabled people recover lower limb motor functions faster.The main work and contributions of this thesis are as follows:First,based on the principle of lower limb joint actuation,the method of joint actuation and lower limb gait generation is proposed.Through the method of extra moment motion driving of the knee joint,the flexion and extension motion of the knee joint was successfully achieved,and the effectiveness of external moment driving of the knee joint was demonstrated.The method of lower limb gait generation assisted by multiple joint moments of the lower limb is proposed,and the experimental results verify that the method of gait generation by external moment assisted driving of human lower limb joints is feasible and effective.Second,in response to the current problems of a single mode of lower limb rehabilitation training,insufficient synergy of lower limb joints,and lack of trunk movement,we developed a multiple rehabilitation model of active joint synergy training and trunk movement-assisted lower limb rehabilitation.A comprehensive trunk-lower limb synergistic motion assistance and a hierarchical control software architecture scheme are proposed.The forward and reverse kinematics of the system motion mechanism is analyzed and solved,and the dynamics model of the motion assist mechanism is established to lay the foundation for the system motion control research.The test and simulation results verified that the motion space of the exoskeleton robot overlaps with the human lower limb motion space.The proposed robot system can meet the rehabilitation needs of human lower limb and trunk motion.Third,a multi-source information sensing and monitoring network is proposed to address the problems of single,untimely and unintuitive feedback data for traditional rehabilitation training of lower limbs.Facing the complex usage environment and multiple noise interference problems,the Kalman filter theory is applied and a multimodal fusion algorithm is proposed to improve the reliability of sensor detection and measurement accuracy.The results verify the effectiveness of the designed sensor monitoring network and the filtering and fusion algorithms.Fourth,four motor rehabilitation training modes are proposed to address the problems of uncoordinated trunk movements,inefficient and inaccurate manual demonstrations,poor accuracy of passive training movements,and insufficient active flexibility during lower limb rehabilitation training.First,for the problems of center of gravity undulation motion and trunk motion training,a parametric adaptive control strategy that can adapt to the patient’s weight and an RBF adaptive control strategy are proposed to provide assistance to the patient’s center of gravity and trunk translational and rotational motion.Second,to address the drawbacks of manual demonstration rehabilitation training method,a Kinect camera-based visual demonstration method is proposed to achieve the personalized and repeatable rehabilitation demonstration training of the lower limbs.Next,for the problem of low precision and poor robustness of passive rehabilitation gait,a fuzzy adaptive control strategy is proposed to track the personalized rehabilitation gait.Finally,for the problem of active supple rehabilitation gait training,a human-machine interaction force-based admittance flexible control strategy is proposed.The effectiveness of the proposed control strategy is verified by simulation and experiment,and the system motion assistance function is stable and feasible to meet the motion training demand of lower limb rehabilitation.Starting from the principle of lower limb joints driving,Taking the problems and design requirements indexes of the lower limb rehabilitation robot into account,we developed a complete research platform of the lower limb rehabilitation exoskeleton robot system.Then,we designed a monitoring network for multimodal sensing of the rehabilitation motion process and states.Through a variety of adaptive control algorithms and admittance flexibility control algorithms,rehabilitation training modes including trunk center of gravity,lower limb demonstration,passive training and active flexibility are achieved.Finally,the simulation and experimental results verify that the designed lower limb rehabilitation exoskeleton robot system and control algorithms are effective and feasible. |
