Marco P. Schoen

Feature extraction of forearm EMG signals for prosthetics

This paper presents a new technique for feature extraction of forearm electromyographic (EMG) signals using a proposed mother wavelet matrix (MWM). A MWM including 45 potential mother wavelets is suggested to help the classification of surface and intramuscular EMG signals recorded from multiple locations on the upper forearm for ten hand motions. Also, a surface electrode matrix (SEM) and a needle electrode matrix (NEM) are suggested to select the proper sensors for each pair of motions. For this purpose, EMG signals were recorded from sixteen locations on the forearms of six subjects in ten hand motion classes. The main goal in classification is to define a proper feature vector able to generate acceptable differences among the classes. The MWM includes the mother wavelets which make the highest difference between two particular classes. Six statistical feature vectors were compared using the continuous form of wavelet packet transform. The mother wavelet functions are selected with the aim of optimum classification between two classes using one of the feature vectors. The locations where the satisfactory signals are captured are selected from several mounted electrodes. Finally, three ten-by-ten symmetric MWM, SEM, and NEM represent the proper mother wavelet function and the surface and intramuscular selection for recording the ten hand motions.

Wavelet basis functions in biomedical signal processing

Research highlights? Daubechies 44 wavelet basis is the most similar function across various biosignals. ? High order Daubechies functions are useful to extract features from 1-D biosignals. ? For wavelet signal processing, selection of mother function similar to signal is not always a proper strategy. During the last two decades, wavelet transform has become a common signal processing technique in various areas. Selection of the most similar mother wavelet function has been a challenge for the application of wavelet transform in signal processing. This paper introduces Daubechies 44 (db44) as the most similar mother wavelet function across a variety of biological signals. Three-hundred and twenty four potential mother wavelet functions were selected and investigated in the search for the most similar function. The algorithms were validated by three categories of biological signals: forearm electromyographic (EMG), electroencephalographic (EEG), and vaginal pulse amplitude (VPA). Surface and intramuscular EMG signals were collected from multiple locations on the upper forearm of subjects during ten hand motions. EEG was recorded from three monopolar Ag-AgCl electrodes (Pz, POz, and Oz) during visual stimulus presentation. VPA, a useful source for female sexuality research, were recorded during a study of alcohol and stimuli on sexual behaviors. In this research, after extensive studies on mother wavelet functions, results show that db44 has the most similarity across these classes of biosignals.

Adaptive intelligent control of ionic polymer–metal composites

Electroactive polymers undergo physical deformation in response to external voltage stimuli. These electrically activated polymers possess extraordinary features making them capable of use as lightweight sensors and actuators in manifold applications. The characteristics of applied voltage and environmental conditions, especially the moisture content surrounding the polymer, have a combined influence on the dynamical behavior of these polymers. In order to characterize these polymers under varying environmental conditions, this paper discusses the experimental procedure and modeling techniques used to derive a representative model. Validation of the model derived is provided by comparison tests of the simulated model results and those for experimental specimens. Ionic polymer–metal composites are used for this humidity and electrodynamical study. Insight into the numerous applications of electroactive polymers as actuators is given. The extended model allows for controller design for typical tracking problems. The control architecture presented includes a model reference adaptive scheme along with pole-placement control strategies for achieving the goal of tracking. A genetic algorithm approach is employed to carry out the optimization for the control action. The resulting tracking control of ionic polymer–metal composites, acting as actuators, is simulated. Simulations show that tracking results can be achieved with a correlation of 99% and a root mean square error of less than 30%.

Prosthetic devices: Challenges and implications of robotic implants and biological interfaces:

Abstract Although among designs of prosthetics there have been some successes in the design of functional robotic implants, there remain many issues and challenges concerned with the failure to meet the ‘ideal’ requirements of a satisfactory prosthetic. These ‘ideals’ require the device to be easy to control, comfortable to wear, and cosmetically pleasing. Because the literature on prosthetics and robotic implants are voluminous, this review focuses on four topics to determine key challenges and opportunities underlying these interdisciplinary research areas: firstly, an artificial hand as a biomimetic; secondly, prosthetic implants (electromyography signals and control); thirdly, prosthetic implants and tissue reactions to the material(s) of implants; fourthly, how inflammatory responses of cells and tissues surrounding implanted sensors interfere with the signal transmission of such sensors. This review also notes the importance of the biological interfaces that robotic implants and other prosthetic devices are in contact with and how an improved knowledge of pathophysiological changes at such biological interfaces will lead to improved and more biocompatible designs of prosthetics. This review concludes with the vision that, to develop a design that satisfies the above ‘ideals’, an interdisciplinary team of biomedical and tissue engineers, and biomaterial and biomedical scientists is needed to work together holistically and synergistically.

Wave Prediction and Robust Control of Heaving Wave Energy Devices for Irregular Waves

This paper presents a comparison of different existing and proposed wave prediction models applicable to control wave energy converters (WECs) in irregular waves. The objective of the control is to increase the energy conversion. The power absorbed by a WEC is depending on the implemented control strategy. Uncertainties in the physical description of the system as well as in the input from irregular waves provide challenges to the control algorithm. We present a hybrid robust fuzzy logic (FL) controller that addresses the uncertainty in the model and the short-term tuning of the converter. The proposed robust controller uses the energy from the power take off as input, while the FL controller portion is used for short-term tuning. The FL controller action is based on a prediction of the incoming wave. Simulation results indicate that the proposed hybrid control yields improved energy production, while being robust to modeling errors.

Control strategies for smart prosthetic hand technology: An overview

A chronological overview of the applications of control theory to prosthetic hand is presented. The overview focuses on hard computing or control techniques such as multivariable feedback, optimal, nonlinear, adaptive and robust and soft computing or control techniques such as artificial intelligence, neural networks, fuzzy logic, genetic algorithms and on the fusion of hard and soft control techniques. This overview is not intended to be an exhaustive survey on this topic and any omissions of other works is purely unintentional.

System Identification and Robust Controller Design Using Genetic Algorithms for Flexible Space Structures

This paper is concerned with the problem of identifying and controlling flexible structures. The structures used exhibit some of the characteristics found in large flexible space structures (LFSSs). Identifying LFSS are problematic in the sense that the modes are of low frequency, lightly damped, and often closely spaced. The proposed identification algorithm utilizes modal contribution coefficients to monitor the data collection. The algorithm is composed of a two-step process, where the input signal for the second step is recomputed based on knowledge gained about the system to be identified. In addition, two different intelligent robust controllers are proposed. In the first controller, optimization is concerned with performance criteria such as rise time, overshoot, control energy, and a robustness measure among others. Optimization is achieved by using an elitism based genetic algorithm (GA). The second controller uses a nested GA resulting in an intelligent linear quadratic regulator/linear quadratic Gaussian (LQR/LQG) controller design. The GAs in this controller are used to find the minimum distance to uncontrollability of a given system and to maximize that minimum distance by finding the optimal coefficients in the weighting matrices of the LQR/LQG controller. The proposed algorithms and controllers are tested numerically and experimentally on a model structure. The results show the effectiveness of the proposed two-step identification algorithm as well as the utilization of GAs applied to the problem of designing optimal robust controllers.

A Hybrid Optimal Control Strategy for a Smart Prosthetic Hand

This paper presents a hybrid of a soft computing or control technique of adaptive neuro-fuzzy inference system (AN-FIS) and a hard computing or control technique of the hybrid finite-time linear quadratic optimal control for a two-fingered (thumb and index) prosthetic hand. In particular, the ANFIS is used for inverse kinematics, and the optimal control is used to minimize tracking error utilizing feedback linearized dynamics. The simulations of this hybrid controller, when compared with the proportional-integral-derivative (PID) controller showed enhanced performance. Work is underway to extend this methodology to a five-fingered, three-dimensional prosthetic hand.Copyright

Frequency domain surface EMG sensor fusion for estimating finger forces

Extracting or estimating skeletal hand/finger forces using surface electro myographic (sEMG) signals poses many challenges due to cross-talk, noise, and a temporal and spatially modulated signal characteristics. Normal sEMG measurements are based on single sensor data. In this paper, array sensors are used along with a proposed sensor fusion scheme that result in a simple Multi-Input-Single-Output (MISO) transfer function. Experimental data is used along with system identification to find this MISO system. A Genetic Algorithm (GA) approach is employed to optimize the characteristics of the MISO system. The proposed fusion-based approach is tested experimentally and indicates improvement in finger/hand force estimation.

Intelligent Techniques for Star-Pattern Recognition

This work presents the study of two different approaches for the attitude determination of space vehicles. The Neural Network approach is based on a simple Kohonen network, where the characteristics of a star distribution within the Field Of View (FOV) are matched against an on-board stored star map. The second approach utilizes a Genetic Algorithm (GA). The GA minimizes the discrepancy between the characteristics of the stars inside the FOV and a candidate FOV selected from the star map. Preliminary simulations indicate that both approaches work well and deliver good accuracy in determining the bore sight direction of the space vehicle with respect to the star map.Copyright