基于触觉P300的脑控下肢康复机器人

基于P300信号的脑机接口技术在康复医疗领域具有广阔的应用前景，但P300信号的诱发方式多为视听刺激诱发，容易导致患者视听疲劳，同时也限制了视听障碍患者的使用。针对这些不足，设计一种基于触觉P300的脑控下肢康复机器人系统，在被试的左右食指处各放置一个振动器，通过调整左右手振动器的刺激间隔、刺激时长及刺激比例让P300信号更容易诱发和区分；利用共空间模式算法和支持向量机对信号进行特征提取和分类。被试通过选择关注左手或右手的振动刺激输出不同指令，从而控制下肢康复机器人进行相应动作。实验证明，被试通过感受振动刺激可以轻松诱发脑电中的P300信号，在不进行P300信号平均叠加的条件下，分类准确率为86.50%，既保证了较高的分类准确率，又缩短了指令输出时间。每位被试均可通过下肢康复系统顺利完成训练任务，证明了该方法的可行性。

Brain-Controlled Lower Limb Rehabilitation Robot Based on Tactile P300

Brain-computer interface technology based on P300 signal has broad application prospects in the field of rehabilitation medicine. However, the induction mode of P300 signal is mostly induced by audio-visual stimulation, which is easy to cause audio-visual fatigue of patients, and also limits the use of patients with audio-visual disorders. In view of these shortcomings, a brain-controlled lower limb rehabilitation robot system based on tactile P300 was designed. A vibrator was placed on the left and right index finger of the subjects, and the stimulation interval, duration and ratio of the left and right vibrators were adjusted to make the P300 signals more easily induced and distinguished. The common spatial pattern algorithm and support vector machine was used to extract and classify the signal. The subjects selected the vibration stimuli of the left or right hand to output different instructions, so as to control the lower limb rehabilitation robot to carry out corresponding actions. Experiments show that subjects can easily induce P300 signals in EEG by feeling vibration stimulation. The classification accuracy is 86.50% without average superposition of P300 signals, which not only ensures high classification accuracy, but also shortens the instruction output time. Each subject can successfully complete the training task through the lower limb rehabilitation system, which proves the feasibility of the method.