欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
为了提高下肢表面肌电信号步态识别的准确性，提出了一种基于遗传算法（GA）优化的BP神经网络分类器设计方法。首先，对采集的下肢表面肌电信号进行小波滤波及特征提取，其次，构造基于GA优化的BP神经网络分类器，然后，以提取的表面肌电信号特征作为输入对分类器进行训练，最后对训练好的分类器进行测试。实验结果表明，基于GA优化的BP神经网络分类器能成功识别下肢正常行走的5个步态，平均识别率达到98％以上，可见基于GA—BP神经网络分类器的识别效果明显优于BP神经网络分类器。
英文摘要：
In order to improve the accuracy of the lower limb gait recognition using limb surface electromyography, a classification method of BP neural network optimized by genetic algorithms （GA） is put forward. Firstly filter and extract features of limb SEMG, Secondly construct the BP neural network classifier based on GA optimization, then train the classifier with extracted features of SEMG, and finally test the classifier which has been trained. The experi-mental results show that the BP neural network classifier based on GA optimization can successfully identify five normal walking gaits of lower limb and the average recognition rate is above 98%. The recognition effect of GA-BP method is obviously better than that of BP neural network.
同期刊论文项目
同项目期刊论文
期刊信息
位置：