Optimized Gaussian mixture models for upper limb motion classification.

This paper introduces the use of Gaussian mixture models (GMM) for discriminating multiple classes of limb motions using continuous myoelectric signals (MES). The purpose of this work is to investigate an optimum configuration of a GMM-based limb motion classification scheme. For this effort, a complete experimental evaluation of the Gaussian mixture motion model is conducted on a 12-subject database. The experiments examine algorithmic issues of the GMM including the model order selection and variance limiting. The final classification performance of this GMM system has been compared with that of three other classifiers (a linear discriminant analysis (LDA), a linear perceptron neural network (LP) and a multilayer perceptron (MLP) neural network) . The Gaussian mixture motion model attains 96.3% classification accuracy using four channel MES for distinguishing six limb motions and is shown to outperform the other motion modeling techniques on an identical six limb motion task.

Single motor unit myoelectric signal analysis with nonstationary data.

The information content of the myoelectric signal (MES) is commonly revealed by statistical measures in the time or frequency domain. Empirical analyses of the MES from a single motor unit have generally assumed that features are invariant with time. Theoretical and experimental work has been done to demonstrate how nonstationary behavior in the discharge statistics of a motor neuron may affect estimates of features extracted from the motor unit's contribution to the MES. Specifically, it has been shown that nonstationary behavior can markedly influence estimates of features describing motor neuron firing behavior and consequently, the low-frequency portion of the MES power spectral density. These results may help to explain the discrepancies in the literature which report empirical models of motor neuron firing statistics.