基于柔性神经树和表面肌电信号的手势识别模型

表面肌电信号因为具有非入侵式、易于采集特点,被广泛应用于康复医疗和行为识别等领域.传统的基于表面肌电信号sEMG( Surface Electromyography)的手势识别通常采用数字信号处理DSP( Digital Signal Processing)芯片或者集合方法研究实时识别问题.这些方法易导致数学模型参数繁多、硬件连接复杂和实时识别率较低.提出一种基于肌电信号与柔性神经树FNT(Flexible Neural Trees)模型的实时手势识别模型.柔性神经树模型通过简单的预定义建立,能够解决人工神经网络ANN( Artificial Neural Network)的结构高依赖性问题.柔性神经树模型不仅能够避免复杂的计算和电路连接,还具有较高的实时识别率.针对六名参与者的六种手势进行实验,实验结果表明:该模型的均方根误差RMSE(Root Mean Square Error)最低为0.000385,实时识别率最高可达97.53％.

HAND GESTURE RECOGNITION MODEL BASED ON FLEXIBLE NEURAL TREES AND SURFACE ELECTROMYOGRAPHY

Surface Electromyography(sEMG) signals have the features of non-invasive and easy to record,therefore they have been applied widely in many fields of treatment and rehabilitation as well as in behaviour recognition.Traditional hand gesture recognition methods using surface Electromyography(sEMG) signals usually employs digital signal processing chips or ensemble methods as tools to solve the issue of real time recognition.These methods are prone to resulting in complicated mathematical models parameters,complex circuit connection and lower real time recognition rate.A real-time hand gesture recognition model by combining flexible neural trees(FNT) model and sEMG signals is proposed in this paper.The FNT model is generated and evolved based on the pre-defined simple instruction sets,which can solve highly structure dependent problem of the artificial neural network(ANN).FNT method avoids complicated computation and inconvenience of circuit connection and also has a higher real-time recognition rate.Testing has been conducted with six participants to test six hand gestures.The results indicate that the model has lower RMSE of which the lowest is 0.000385,and the highest real time recognition rate is up to 97.53%.