基于单通道脑机接口的智能轮椅控制及人机交互研究

发布时间：2018-04-24 13:03

本文选题：单通道脑机接口 + 眨眼脑电信号　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着伤老病残人群对生活品质需求不断提高,对基于脑电控制的人机交互式智能轮椅研究有很大必要性,然而基于运动想象脑电信号的智能轮椅控制系统庞大,信号采集设备成本昂贵,轮椅自由运动受限,致使目前仅限于实验室研究而无法大规模应用。为此本课题开发了低成本、较好性能的人机交互式智能轮椅系统,该系统采用便携式的单通道脑机接口设备采集脑电信号,利用脑电信号眨眼特征和集中度参数结合人机交互界面控制轮椅运动,最终能够实现基本运动,实时性和可操控性可较好满足要求。首先,研究了脑电信号的产生机理和特点,分析了脑机接口系统的工作原理及其国内外研究现状,研究了快速傅里叶变换、小波变换理论和功率谱估计等脑电信号粉刺处理算法并分别matlab仿真实现,改进基于小波的软硬阈值收缩算法并评估改进效果。其次,选定了脑电信号的采集与传输方式,接收到原始脑电信号后采用改进的小波阈值收缩算法对其进行了去伪迹处理,通过特征识别算法识别出眨眼特征和精神集中特征并转换为控制信号,在MFC平台设计了基于眨眼脑电信号的人机交互系统。选取一定的样本量设计了眨眼脑电信号的特征参数预标定实验并验证,根据预标定结果设定眨眼信号的阈值和集中度参数完成了人机交互实验研究。再次,完成了基于单通道脑机接口的智能轮椅控制与人机交互系统的软硬件设计。设计了STM32处理器核心控制电路和脑电信号采集传输、直流电机控制、交互显示界面、辅助避障等系统硬件电路,在STM32上移植了μC/OS-III嵌入式实时操作系统和STemWin图形界面,在系统中进行了脑电信号接收解码与眨眼特征识别,编写和调试智能轮椅运动状态控制和交互界面信息显示等程序的代码,实现了智能轮椅与人机交互控制的软件系统。最后,联合调试基于眨眼脑电信号的智能轮椅控制与人机交互的软硬件系统,在不同类型的路径下分别利用眨眼脑电信号特征和其集中度参数控制轮椅运动,对系统性能进行实验测试,给出了相关的测试数据并分析其结果。测试结果表明,基于单通道脑机接口的智能轮椅控制与人机交互系统具有较好的实时性和可操控性,为便携式脑机接口技术在实际应用中的探索研究奠定了基础。
[Abstract]:With the increasing demand for the quality of life among the elderly, the disabled and the injured, it is necessary to study the man-machine interactive intelligent wheelchair based on EEG control. However, the intelligent wheelchair control system based on EEG is huge. Signal acquisition equipment is expensive and wheelchair free movement is limited, so it is limited to laboratory research and can not be applied on a large scale. Therefore, a low cost and better performance man-machine interactive intelligent wheelchair system is developed. The system uses portable single-channel brain-computer interface equipment to collect EEG signals. The basic motion of wheelchair can be realized by using the blinking characteristic and concentration parameter of EEG combined with the man-machine interaction interface. The real-time and maneuverability can meet the requirements well. Firstly, the mechanism and characteristics of EEG signal generation are studied, the working principle of BCI system and its research status at home and abroad are analyzed, and the fast Fourier transform (FFT) is studied. Wavelet transform theory, power spectrum estimation and other EEG signal processing algorithms are implemented by matlab simulation, which improves the soft and hard threshold shrinkage algorithm based on wavelet transform and evaluates the effect of the improved algorithm. Secondly, the acquisition and transmission mode of EEG signal is selected. After receiving the original EEG signal, the improved wavelet threshold shrinkage algorithm is used to deartifact the EEG signal. The blinking feature and the concentration feature are recognized by the feature recognition algorithm and converted into control signals. A human-computer interaction system based on the blinking EEG is designed on the MFC platform. A certain sample size is selected to design the pre-calibration experiment of the characteristic parameters of the blinking EEG signal and verify it. According to the pre-calibration results, the threshold and concentration parameters of the blinking signal are set up to complete the human-computer interaction experiment. Thirdly, the software and hardware design of intelligent wheelchair control and man-machine interaction system based on single channel brain-computer interface is completed. The core control circuit of STM32 processor and the hardware circuits of EEG acquisition and transmission, DC motor control, interactive display interface and auxiliary obstacle avoidance are designed. 渭 C/OS-III embedded real-time operating system and STemWin graphic interface are transplanted on STM32. The software system of intelligent wheelchair and man-machine interactive control is realized by the code of EEG receiving and decoding and blink feature recognition in the system, and the code of intelligent wheelchair motion state control and interactive interface information display are compiled and debugged. Finally, the software and hardware system of intelligent wheelchair control and man-machine interaction based on blinking EEG signal is debugged, and the wheelchair motion is controlled by the characteristics of blinking EEG signal and its concentration parameter under different paths, respectively. The performance of the system is tested, the relevant test data are given and the results are analyzed. The test results show that the intelligent wheelchair control and man-machine interaction system based on single channel Brain-Computer Interface have better real-time and maneuverability, which lays a foundation for the exploration and research of portable Brain-Computer Interface technology in practical application.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318;TN911.7

【参考文献】