Research on Adaptive Grasping with a Prosthetic Hand Based on Perceptual Information on Hardness and Surface Roughness.

In order to solve the problems of methods that use a single form of sensing, the ease of causing deformation damage to the targets with a low hardness during grasping, and the slow sliding inhibition of a prosthetic hand when the grasping target slides, which are problems that exist in most current intelligent prosthetic hands, this study introduces an adaptive control strategy for prosthetic hands based on multi-sensor sensing. Using a force-sensing resistor (FSR) to collect changes in signals generated after contact with a target, a prosthetic hand can classify the target's hardness level and adaptively provide the desired grasping force so as to reduce the deformation of and damage to the target in the process of grasping. A fiber-optic sensor collects the light reflected by the object to identify its surface roughness, so that the prosthetic hand adaptively adjusts the sliding inhibition method according to the surface roughness information to improve the grasping efficiency. By integrating information on the hardness and surface roughness of the target, an adaptive control strategy for a prosthetic hand is proposed. The experimental results showed that the adaptive control strategy was able to reduce the damage to the target by enabling the prosthetic hand to achieve stable grasping; after grasping the target with an initial force and generating sliding, the efficiency of slippage inhibition was improved, the target could be stably grasped in a shorter time, and the hardness, roughness and weight ranges of targets that could be grasped by the prosthetic hand were enlarged, thus improving the success rate of stable grasping under extreme conditions.

A Hand Gesture Recognition Strategy Based on Virtual-Dimension Increase of EMG.

The electromyography(EMG) signal is the biocurrent associated with muscle contraction and can be used as the input signal to a myoelectric intelligent bionic hand to control different gestures of the hand. Increasing the number of myoelectric-signal channels can yield richer information of motion intention and improve the accuracy of gesture recognition. However, as the number of acquisition channels increases, its effect on the improvement of the accuracy of gesture recognition gradually diminishes, resulting in the improvement of the control effect reaching a plateau. To address these problems, this paper presents a proposed method to improve gesture recognition accuracy by virtually increasing the number of EMG signal channels. This method is able to improve the recognition accuracy of various gestures by virtually increasing the number of EMG signal channels and enriching the motion intention information extracted from data collected from a certain number of physical channels, ultimately providing a solution to the issue of the recognition accuracy plateau caused by saturation of information from physical recordings. Meanwhile, based on the idea of the filtered feature selection method, a quantitative measure of sample sets (separability of feature vectors [SFV]) derived from the divergence and correlation of the extracted features is introduced. The SFV value can predict the classification effect before performing the classification, and the effectiveness of the virtual-dimension increase strategy is verified from the perspective of feature set differentiability change. Compared to the statistical motion intention recognition success rate, SFV is a more representative and faster measure of classification effectiveness and is also suitable for small sample sets.

Design of an Effective Prosthetic Hand System for Adaptive Grasping with the Control of Myoelectric Pattern Recognition Approach.

In this paper, we develop a prosthetic bionic hand system to realize adaptive gripping with two closed-loop control loops by using a linear discriminant analysis algorithm (LDA). The prosthetic hand contains five fingers and each finger is driven by a linear servo motor. When grasping objects, four fingers except the thumb would adjust automatically and bend with an appropriate gesture, while the thumb is stretched and bent by the linear servo motor. Since the change of the surface electromechanical signal (sEMG) occurs before human movement, the recognition of sEMG signal with LDA algorithm can help to obtain people's action intention in advance, and then timely send control instructions to assist people to grasp. For activity intention recognition, we extract three features, Variance (VAR), Root Mean Square (RMS) and Minimum (MIN) for recognition. As the results show, it can achieve an average accuracy of 96.59%. This helps our system perform well for disabilities to grasp objects of different sizes and shapes adaptively. Finally, a test of the people with disabilities grasping 15 objects of different sizes and shapes was carried out and achieved good experimental results.

A Study on the Classification Effect of sEMG Signals in Different Vibration Environments Based on the LDA Algorithm.

Myoelectric prosthesis has become an important aid to disabled people. Although it can help people to recover to a nearly normal life, whether they can adapt to severe working conditions is a subject that is yet to be studied. Generally speaking, the working environment is dominated by vibration. This paper takes the gripping action as its research object, and focuses on the identification of grasping intentions under different vibration frequencies in different working conditions. In this way, the possibility of the disabled people who wear myoelectric prosthesis to work in various vibration environment is studied. In this paper, an experimental test platform capable of simulating 0-50 Hz vibration was established, and the Surface Electromyography (sEMG) signals of the human arm in the open and grasping states were obtained through the MP160 physiological record analysis system. Considering the reliability of human intention recognition and the rapidity of algorithm processing, six different time-domain features and the Linear Discriminant Analysis (LDA) classifier were selected as the sEMG signal feature extraction and recognition algorithms in this paper. When two kinds of features, Zero Crossing (ZC) and Root Mean Square (RMS), were used as input, the accuracy of LDA algorithm can reach 96.9%. When three features, RMS, Minimum Value (MIN), and Variance (VAR), were used as inputs, the accuracy of the LDA algorithm can reach 98.0%. When the six features were used as inputs, the accuracy of the LDA algorithm reached 98.4%. In the analysis of different vibration frequencies, it was found that when the vibration frequency reached 20 Hz, the average accuracy of the LDA algorithm in recognizing actions was low, while at 0 Hz, 40 Hz and 50 Hz, the average accuracy was relatively high. This is of great significance in guiding disabled people to work in a vibration environment in the future.

SEMG Feature Extraction Based on StockwellTransform Improves Hand MovementRecognition Accuracy.

Feature extraction, as an important method for extracting useful information from surfaceelectromyography (SEMG), can significantly improve pattern recognition accuracy. Time andfrequency analysis methods have been widely used for feature extraction, but these methods analyzeSEMG signals only from the time or frequency domain. Recent studies have shown that featureextraction based on time-frequency analysis methods can extract more useful information fromSEMG signals. This paper proposes a novel time-frequency analysis method based on the Stockwelltransform (S-transform) to improve hand movement recognition accuracy from forearm SEMGsignals. First, the time-frequency analysis method, S-transform, is used for extracting a feature vectorfrom forearm SEMG signals. Second, to reduce the amount of calculations and improve the runningspeed of the classifier, principal component analysis (PCA) is used for dimensionality reduction of thefeature vector. Finally, an artificial neural network (ANN)-based multilayer perceptron (MLP) is usedfor recognizing hand movements. Experimental results show that the proposed feature extractionbased on the S-transform analysis method can improve the class separability and hand movementrecognition accuracy compared with wavelet transform and power spectral density methods.