The Research On Perception And Interaction Methods Of Lower Limbs Rehabilitation Exoskeletons Robot

The lower limb rehabilitation exoskeleton robot is an intelligent robot device,which is a combination of the exoskeleton bionic technology and information control technology,involves the human kinematics,robotics,information science,intelligent control,rehabilitation medicine,etc.As the aging of the population,the number of the mobility of the elderly is expanding due to the attenuation of the function of the body.In addition,the number of the lower limb disabled and motor dysfunction patients is also larger,which caused by natural disasters,traffic accidents and diseases.Hence,the research and development of lower limb rehabilitation exoskeleton robot has become a hot direction.The main focus of this paper is to design the robot's perception system,study the interactive control method,analysis the movement mechanism of human lower limbs,build motion state recognition model based on sensor data,take model online validation on the engineering prototype platform.This paper firstly introduces the research background and the domestic and foreign research status of the lower limb exoskeleton robot,then analyzes the movement mechanism of human lower limbs combined with the theory of human anatomy,determine the type,number and layout of the sensor.Then it designed the sensor information acquisition module circuit,which connects the embedded PC master controller through CAN bus,composed of the distributed sensing system of the robot.Collect and save the wearer's motion status information through the perceptual system,then divide the phase based on human gait motion analysis.Combined with the characteristics of human lower limb movement,it can implement multi-sensor signal data fusion analysis,and built the transition model of sitting state,standing state and walking state based on finite state machine.The gait motion phase recognition model is constructed in combination with the partition of gait cycle based on fuzzy inference method.Then we use the learning vector quantization neural network model to realize automatic stable switching between supporting legs and swinging legs during walking.In addition,we develop an application to realize the remote monitoring and parameter modification of the control system through wireless network.Finally,the experimenter wore the engineering prototype robot to collect data.The parameters of the identification model are determined according to the sample data.Then the model is embedded into the control system for online test verification.The results show that the constructed perceptual interaction model has good practicability and can meet the basic needs of exoskeleton robot's perception and interaction.