| 摘要: |
| 为了提高下肢步态识别的准确性和效率，针对K最近邻（KNN）算法参数调节困难的问题，提出了一种基于灰雁优化-K维树-K最近邻（GGO-KD-KNN）算法的下肢步态识别方法.首先，利用表面肌电信号（sEMG）采集下肢肌肉活动信息，并将信号划分为5个步态阶段.然后，进行sEMG去噪，并提取时域和频域特征.接着，用GGO算法基于灰雁群体行为进行启发式优化，优化KNN算法的K值和距离度量，并通过适应度迭代寻找最优解.实验结果表明，通过GGO算法优化的步态识别精度达到了98.23%，标准差为0.264，相较于其他常用算法，基于GGO-KD-KNN算法的步态识别方法展现出更高的分类准确率和稳定性，为下肢智能辅助装置的研究和开发提供了有力的理论支持. |
| 关键词: 下肢步态识别 表面肌电信号（sEMG） 灰雁优化-K维树-K最近邻（GGO-KD-KNN）算法 分类优化 |
| DOI：10.20192/j.cnki.JSHNU（NS）.2025.02.002 |
| 分类号:TN911.7 |
| 基金项目: |
|
| Research on lower limb gait recognition based on GGO-KD-KNN algorithm |
|
LI Chuanjiang, DING Xinhao, TU Jiajun, LI Ang, YIN Shiyi
|
|
College of Information, Mechanical and Electrical Engineering, Shanghai Normal University, Shanghai 201418, China
|
| Abstract: |
| To improve the accuracy and efficiency of lower limb gait recognition， a lower limb gait recognition method based on the greylag goose optimization-K dimensional tree-K nearest neighbor （GGO-KD-KNN） algorithm was proposed to solve difficult parameter adjustment problem. Firstly， the surface electromyographic signal （sEMG） was used to obtain the muscle activity information of the lower limb， and the signals were classified into five gait phases. Secondly， the sEMG signals were denoised， and time-domain and frequency-domain features were extracted. Thirdly， the GGO algorithm was inspired by the behavior of gray geese groups to optimize the K value and distance metric of the KNN algorithm. Finally， the optimal solution was obtained by iterative fitness. The experimental results showed that the gait recognition method based on GGO-KD-KNN algorithm achieved higher classification accuracy （reached 98.23%， with a standard deviation of 0.264） and stability compared with other commonly used algorithms， which provided a theoretical basis for the research and development of human lower limb intelligent assistive devices. |
| Key words: lower limb gait recognition surface electromyographic signal （sEMG） greylag goose optimization-K dimensional tree-K nearest neighbor （GGO-KD-KNN） algorithm classification optimization |
