Research On Auxiliary Appliance Safety Action For Huaman Lower Limb Rehabilitation

With the aging of our country’s population, gait instability caused by the phenomenon of falls in older people gradually increased. Human lower limb rehabilitation robot plays an important role on helping injured people or people with disabilities to restore basic walking ability. With some external disturbance, different strategies balance adjustment will be selected based on the different size and direction of the disturbance. Gait data in this process showing a high-dimensional, high variability, non-linear characteristics, this gait analysis include multi-variables and interaction data will be a difficulty part.Considering human gait analysis, complex and high-dimensional characteristics, human lower limb biped robot stability and usability, ultimately establish a simulation design for walking rehabilitation actions:(1) overall robot design:overall structure and the overall control program of the robot program (2) Kinematic Modeling and Simulation Analysis:From the establishment of the kinematic equations, carry forward kinematics and inverse kinematics analysis; obtain the equations of motion of human lower limb robots, using the MATLAB Simulink and SimMechanics toolbox to build kinematic model and simulation analysis.(3) Using neural network structure establishes adaptive control of gait simulation.According to the above characteristics, we can design a controller. Adaptive neural network control is compare required performance with system actual performance information to obtain the controller or control parameters and amendment them. The control method of the system can maintain optimal or sub-optimal operating condition. Specifically, the controller should according to control objects and external interference with the dynamic characteristics promptly correct and change its own characteristics, so that the entire control system to maintain satisfactory performance. Due to the RBF neural network can establish the appropriate network topology based on the problem which we need to be solved. RBF neural network has self-learning, self-organizing, adaptive function, nonlinear continuous function approximation, learning fast, can be a wide range of data integration and data the high-speed parallel processing. So we select the RBF neural network as adaptive controller to control gait trajectory of robots.In this thesis, the results of the adaptive control strategy, which makes the walker design more humane, while providing some reference for the rehabilitation of disabled people of lower extremity function.