【摘要】 针对采用加速度信号进行步态识别错误率偏高的问题，提出一种基于信息熵和改进特征权重融合的AdaboostSSA-BP步态识别方法。基于信息熵理论和改进特征权重算法，提取不同步态下加速度计输出的信号特征，并对其进行特征组合。利用SSA优化BP神经网络，通过Adaboost算法调整网络的样本权重，并进行训练获得步态识别模型。实验结果表明：该方法能够有效捕获步态特征，步态识别的平均准确率可达96.15%，可为后期开展步态康复训练等相关研究提供技术支撑。更多还原

【Abstract】 Aiming at the problem of high error rate of gait recognition using acceleration signals, an Adaboost-SSA-BP gait recognition method based on the fusion of information entropy and improved feature weight is proposed. The features for the outputs from the accelerometer under different gait are extracted based on the information entropy theory and the improved feature weight method algorithm, and then the features are combined for the gait recognition. The BP neural network is optimized by SSA, the weight of the optimized network is adjusted by Adaboost algorithm, and the gait recognition model is obtained by iterative training. Experiment results show that the method can effectively capture gait features, and the average accuracy of gait recognition reaches 96.15%. It can provide the technical support for the related researches such as the gait rehabilitation training.更多还原