机电工程技术2024，Vol.53Issue(4)：36-40,5.DOI:10.3969/j.issn.1009-9492.2024.04.008
基于AdaBoost-WOA-HKELM的下肢关节角度预测
Lower Limb Joint Angle Prediction Based on AdaBoost-WOA-HKELM
摘要
Abstract
To address the issue of low accuracy in continuous lower limb motion prediction,a novel hip and knee joint angle prediction method based on AdaBoost-WOA-HKELM is proposed.Surface electromyographic signals and joint angle data during normal human walking are collected,the feature extraction from the preprocessed signals are carried out and the joint angle information is combined to establish the characteristic data set.The hybrid kernel extreme learning machine(HKELM)model is selected as the weak learner,with the whale optimization algorithm(WOA)employed to optimize the parameters of the HKELM model.The AdaBoost ensemble learning algorithm is then used to train the weak learner into a strong one.The AdaBoost-WOA-HKELM model is trained and tested by using the feature data set,and compared with the HKELM and WOA-HKELM models in the simulation experiment of hip and knee joint angle prediction.The results show that the AdaBoost-WOA-HKELM model demonstrates excellent performance in predicting hip and knee joint angles,with mean squared errors of 2.086 9 and 2.284 9,and determination coefficients of 0.988 2 and 0.988 7 respectively.The above indexes are significantly better than the other models,highlighting the model's exceptional accuracy in predicting lower limb joint angles.The determination coefficient close to one indicates that the model fits the actual data extremely highly,which further verifies the validity and accuracy of the AdaBoost-WOA-HKELM model.关键词
肌电信号/混合核极限学习机/AdaBoost/WOA/下肢关节角度预测Key words
surface electromyography/hybrid kernel extreme learning machine/AdaBoost/WOA/lower limb joint angle prediction分类
信息技术与安全科学引用本文复制引用
李花宁,吴生彪,冯丽,刘瑾,熊书慧..基于AdaBoost-WOA-HKELM的下肢关节角度预测[J].机电工程技术,2024,53(4):36-40,5.基金项目
江西省教育厅基金项目(GJJ202216) （GJJ202216）
