| The recovery of motor function for patients with lower limb injury,stroke and other motor inconvenience has always been a major problem in the medical field,and the lower limb rehabilitation exoskeleton can effectively assist patients to carry out long-term standardized rehabilitation treatment.Meanwhile,due to the emergence of brain-computer interface technology,information interaction between the brain and external devices has been realized.Combining brain-computer interface with exoskeleton to complete rehabilitation training can enhance the training initiative of patients,improve the plasticity of the brain,and effectively help patients recover motor function.In this paper,a brain-controlled lower limb rehabilitation exoskeleton based on motor imagination is designed.The main research contents are as follows:(1)The lower limb rehabilitation exoskeleton with a flexible waist is designed.The physiological structure and joint movement characteristics of human lower limbs are analyzed to determine the degree of freedom of the joint of the lower limb rehabilitation exoskeleton.The waist mechanism with flexible module and the adjustable leg structure are designed and driven by the linear actuator.The static analysis of the lower limb exoskeleton structure using Ansys to ensure that the strength and stiffness of the exoskeleton meet the requirements.The kinematics and dynamics models of D-H method and Lagrange method are established respectively,and the theoretical joint torque curve is calculated.Simulation of the virtual prototype model by Adams to obtain the simulated joint torque curves.The theoretical data is compared with the simulation data,and the results met the needs of sports training.(2)EEG signals based on motor imagination were studied,and EEG signal features were extracted through preconditioning and multi feature joint algorithms.The generation mechanism,physiological characteristics,and collection methods of EEG signals are analyzed.At the same time,the ERD/ERS phenomenon generated during brain movement imagination is studied,and the sources of EEG signal data are analyzed,including the collection process paradigm and electrode position selection..The feature extraction algorithm for time-frequency spatial motor imaging EEG signals is designed,which uses fast independent component analysis for filtering processing to remove EEG artifacts and high-frequency signals.The signal is decomposed and reconstructed using discrete wavelet transform,and the processed signal is compared with the original signal to verify the effectiveness of signal processing.The multi classification common spatial pattern method is used to extract EEG features to obtain feature values.(3)The classification performance of both support vector machine algorithm and convolutional neural network algorithm with high classification degree is studied.The classification parameters of the four-classification support vector machine model are optimized by using particle swarm optimization method,and the support vector machine optimization model of particle swarm algorithm is established,and iterative experiments are conducted by using Matlab to obtain the optimized fitness curve with optimal penalty parameters and kernel function parameters.The principle and training process of convolutional neural network classification algorithm are analyzed.The hyper-parameters optimized convolutional neural network classification model is established,and the classification accuracy is obtained through experiments with different parameters.(4)The control system based on motor imagination EEG signals is constructed and the classification and control experiments are conducted.The EEG signal acquisition experiment is designed to obtain the original signal,which is processed and converted into a control signal using a written PC control program,and the exoskeleton is controlled for movement through a motor control card.The motor imagination EEG signals are screened,and the classification results are compared and analyzed to determine the optimal EEG signal processing method.The effectiveness of the control system and brain-controlled exoskeleton is verified through brain-controlled exoskeleton experiments. |
