The work of Ahsan et al. [16] represents the detection of four hand motions (left, right, up and down) using an artificial neural especially network (ANN). According to Hudgins et al. [17], the success of a pattern next classification system depends completely on the choice of features used to represent the raw EMG signals. Ahsan et al. [16] used seven statistical time and time-frequency based features: mean absolute value, root mean square, variance, Inhibitors,Modulators,Libraries standard Inhibitors,Modulators,Libraries deviation, zero crossing, slope sign change and Wilson amplitude.Rajesh et al. [18], used wavelets and classification using Euclidean distance. The general features of the hand gestures from EMG signal patterns were: hand extension, hand grasp, wrist extension, wrist flexion, and pinch and thumb flexion.
They used the entropy, rms and standard deviation in the analysis of features.
Shenoy et al. [19], used a simple Inhibitors,Modulators,Libraries feature (rms value over windows) and continuously Inhibitors,Modulators,Libraries classified windows of data collected while the subject maintained a static hand gesture (gestures correspond to pairs of actions: grasp-release, left-right, up-down and rotate). The classification was with Linear Support Vector Machines. Boschmann et al. [20], introduce an approach for classifying EMG signals taken from forearm muscles using support vector machines. In a single experiment run, the test subject had to perform a sequence of eleven different movements: extension, flexion, ulnar deviation, radial deviation, pronation, and supination, open, close, key grip, pincer grip and extract the index finger.
Matrone et al.
[21] presents the work of a robotic hand employing a two differential channels (four electrodes) EMG acquisition system and a principal components analysis (PCA) based controller. Participants volunteered in experimental tasks consisting in grasping, Inhibitors,Modulators,Libraries transporting and releasing Inhibitors,Modulators,Libraries different kinds of objects, by employing a five-fingered (and six motors) robotic hand. Inhibitors,Modulators,Libraries The control system decoded and converted the Inhibitors,Modulators,Libraries subjects’ 2-DoF wrist contractions (flexion/extension and adduction/abduction) into hand posture control commands, implementing the algorithm based on PCA [22].Within the last dozen of years, different structures of neuro-fuzzy networks have been presented, often referred to in the world literature as neuro-fuzzy systems [23].
They combine the advantages of neural networks and fuzzy systems. Kurzynski et al.
[24], used the following methods of sequential classification (five types): Bayes approach with Markov model, multilayer perceptron, multiclassifier with competence function, classifier based on fuzzy logic and classifier Batimastat based on Dempster-Shafer theory of evidence. In this paper it is proposed a method to determine AV-951 the input features based selleck kinase inhibitor ref 3 on autoregressive (AR) model. Six different types of grapes depending on the grasping object (a pen, a credit card, a computer mouse, a cell phone, a kettle and a tube) were chosen for recognition.