Browsing by Author "Englehart, Kevin"
Now showing 1 - 20 of 26
Results Per Page
Sort Options
Item A Deep Transfer Learning Approach to Reducing the Effect of Electrode Shift in EMG Pattern Recognition-Based Control(Institute of Electrical and Electronics Engineers, 2020) Ameri, Ali; Akhaee, Mohammad Ali; Scheme, Erik; Englehart, KevinAn important barrier to commercialization of pattern recognition myoelectric control of prostheses is the lack of robustness to confounding factors such as electrode shift, skin impedance variations, and learning effects. To overcome this challenge, a novel supervised adaptation approach based on transfer learning (TL) with convolutional neural networks (CNNs) is proposed which requires only a short training session (a few seconds for each class) to recalibrate the system. TL is proposed as a solution to the problem of insufficient calibration data due to short training times for both classification and regression-based control schemes. This approach was validated for electrode shift of roughly 2.5cm with 13 able-bodied subjects to estimate individual and combined wrist motions. With this method, the original CNN (trained before the shift) was fine-tuned with the calibration data from after shifting. The results show that the proposed technique outperforms training a CNN from scratch (random initialization of weights) or a support vector machine (SVM) using the minimal calibration data. Moreover, it demonstrates superior performance than previous LDA and QDA-based adaptation approaches. As the outcomes confirm, the proposed CNN TL method provides a practical solution for adaptation to external factors, improving the robustness of electromyogram (EMG) pattern recognition systems.Item A multi-variate approach to predicting myoelectric control usability(University of New Brunswick, 2020) Nawfel, Jena; Scheme, Erik; Englehart, KevinPattern recognition techniques leveraging the use of electromyography (EMG) signals have become a popular approach to provide intuitive control of myoelectric devices. Performance of these control interfaces is commonly quantified using offline classification accuracy, despite researchers having shown that this metric is a poor indicator of usability. Several attempts have been made to find alternative training metrics that better correlate with online performance. Moderate correlations have been identified in some cases; however, the relationship between offline training and online usability has yet to be fully defined in the literature. The following work attempts to bridge this information divide by exploring combinations of offline training metrics capable of predicting myoelectric control usability. The results indicate that linear combinations of three offline training metrics provide superior predictive power of future online performance. Additionally, the role of feedback presented to the user during training is explored to determine its effect on performance and predictability. The results of this study suggest that properly designed feedback mechanisms can influence both the quality of the training metrics and the predictive ability of the developed linear models.Item A Multiday Evaluation of Real-Time Intramuscular EMG Usability with ANN(MDPI, 2020) Waris, Asim; ur Rehman, Muhammad Zia; Niazi, Imran Khan; Jochumsen, Mads; Englehart, Kevin; Jensen, Winnie; Haavik, Heidi; Kamavuako, Ernest NlanduRecent developments in implantable technology, such as high-density recordings, wireless transmission of signals to a prosthetic hand, may pave the way for intramuscular electromyography (iEMG)-based myoelectric control in the future. This study aimed to investigate the real-time control performance of iEMG over time. A novel protocol was developed to quantify the robustness of the real-time performance parameters. Intramuscular wires were used to record EMG signals, which were kept inside the muscles for five consecutive days. Tests were performed on multiple days using Fitts’ law. Throughput, completion rate, path efficiency and overshoot were evaluated as performance metrics using three train/test strategies. Each train/test scheme was categorized on the basis of data quantity and the time difference between training and testing data. An artificial neural network (ANN) classifier was trained and tested on (i) data from the same day (WDT), (ii) data collected from the previous day and tested on present-day (BDT) and (iii) trained on all previous days including the present day and tested on present-day (CDT). It was found that the completion rate (91.6 ± 3.6%) of CDT was significantly better (p < 0.01) than BDT (74.02 ± 5.8%) and WDT (88.16 ± 3.6%). For BDT, on average, the first session of each day was significantly better (p < 0.01) than the second and third sessions for completion rate (77.9 ± 14.0%) and path efficiency (88.9 ± 16.9%). Subjects demonstrated the ability to achieve targets successfully with wire electrodes. Results also suggest that time variations in the iEMG signal can be catered by concatenating the data over several days. This scheme can be helpful in attaining stable and robust performance.Item Compression of EMG Signals Using Deep Convolutional Autoencoders(Institute of Electrical and Electronics Engineers, 2022) Dinashi, Kimia; Ameri, Ali; Akhaee, Mohammad Ali; Englehart, Kevin; Scheme, ErikEfficient storage and transmission of electromyogram (EMG) data are important for emerging applications such as telemedicine and big data, as a vital tool for further advancement of the field. However, due to limitations in internet speed and hardware resources, transmission and storage of EMG data are challenging. As a solution, this work proposes a new method for EMG data compression using deep convolutional autoencoders (CAE). Eight-channel EMG data from 10 subjects, and high-density EMG data from 18 subjects, were investigated for compression. The CAE architecture was designed to extract an abstract data representation that is heavily compressed, but from which the salient information for classification can be effectively reconstructed. The proposed method attained efficient compression; for CR = 1600, the average PRDN (percentage RMS difference normalized) was 31.5% and the wrist motions classification accuracy (CA) reduced roughly 5%. The CAE substantially outperformed the state-of-the-art high-efficiency video coding and a well-known wavelet-thresholding compression technique. Moreover, by reducing the bit-resolution of the CAE's compressed data from 24 bits to 6 bits, an additional 4-fold compression was achieved without significant degradation of the reconstruction performance. Furthermore, the CAE's inter-subject performance was promising; e.g., for CR = 1600, the PRDN for the inter-subject case was only 2.6% less than that of the within-subject performance. The powerful EMG compression performance with remarkable reconstruction results reflects the CAEs potential as an automatic end-to-end approach with the ability to learn the complete encoding and decoding process. Furthermore, the excellent inter-subject performance demonstrates the generalizability and usability of the proposed approach.Item Conventional analysis of trial-by-trial adaptation is biased: Empirical and theoretical support using a Bayesian estimator(Public Library of Science, 2018-12) Blustein, Daniel; Shehata, Ahmed; Englehart, Kevin; Sensinger, Jonathon; Maurice A. SmithResearch on human motor adaptation has often focused on how people adapt to self-generated or externally-influenced errors. Trial-by-trial adaptation is a person’s response to self-generated errors. Externally-influenced errors applied as catch-trial perturbations are used to calculate a person’s perturbation adaptation rate. Although these adaptation rates are sometimes compared to one another, we show through simulation and empirical data that the two metrics are distinct. We demonstrate that the trial-by-trial adaptation rate, often calculated as a coefficient in a linear regression, is biased under typical conditions. We tested 12 able-bodied subjects moving a cursor on a screen using a computer mouse. Statistically different adaptation rates arise when sub-sets of trials from different phases of learning are analyzed from within a sequence of movement results. We propose a new approach to identify when a person’s learning has stabilized in order to identify steady-state movement trials from which to calculate a more reliable trial-by-trial adaptation rate. Using a Bayesian model of human movement, we show that this analysis approach is more consistent and provides a more confident estimate than alternative approaches. Constraining analyses to steady-state conditions will allow researchers to better decouple the multiple concurrent learning processes that occur while a person makes goal-directed movements. Streamlining this analysis may help broaden the impact of motor adaptation studies, perhaps even enhancing their clinical usefulness.Item Do Cost Functions for Tracking Error Generalize across Tasks with Different Noise Levels?(Public Library of Science, 2015) Sensinger, Jonathon; Aleman-Zapata, Adrian; Englehart, Kevin; Shu-Dong ZhangControl of human-machine interfaces are well modeled by computational control models, which take into account the behavioral decisions people make in estimating task dynamics and state for a given control law. This control law is optimized according to a cost function, which for the sake of mathematical tractability is typically represented as a series of quadratic terms. Recent studies have found that people actually use cost functions for reaching tasks that are slightly different than a quadratic function, but it is unclear which of several cost functions best explain human behavior and if these cost functions generalize across tasks of similar nature but different scale. In this study, we used an inverse-decision-theory technique to reconstruct the cost function from empirical data collected on 24 able-bodied subjects controlling a myoelectric interface. Compared with previous studies, this experimental paradigm involved a different control source (myoelectric control, which has inherently large multiplicative noise), a different control interface (control signal was mapped to cursor velocity), and a different task (the tracking position dynamically moved on the screen throughout each trial). Several cost functions, including a linear-quadratic; an inverted Gaussian, and a power function, accurately described the behavior of subjects throughout this experiment better than a quadratic cost function or other explored candidate cost functions (p<0.05). Importantly, despite the differences in the experimental paradigm and a substantially larger scale of error, we found only one candidate cost function whose parameter was consistent with the previous studies: a power function (cost ∝ errorα) with a parameter value of α = 1.69 (1.53–1.78 interquartile range). This result suggests that a power-function is a representative function of user’s error cost over a range of noise amplitudes for pointing and tracking tasks.Item Enhancing electromyography based locomotion mode classification when using powered lower limb prostheses(University of New Brunswick, 2012) Razak, Roua; Englehart, Kevin; Stevenson, MaryhelenThe recent introduction of powered leg prostheses enables users to perform many more tasks than passive legs, such as ascending/descending stairs and ramps with greater ease, and standing from a sitting position. With this increased function comes a need for improved control that may be obtained from neural information recorded as the surface electromyographic (EMG) signal. The EMG signal has been shown to be a promising source of autonomous information to characterize the instantaneous mode of locomotion (level walking, ramp ascent/decent, stairs ascent/decent). This is of great value in that the prosthesis can alter its dynamic properties to suit the current mode of locomotion. The greatest challenge in applying EMG to the control of leg prostheses is EMG distortion that is generated due to the fact that leg prostheses must bear the weight of the user. This introduces considerable force/pressure against the socket and motion/compression of the muscle, and may cause an incorrect interpretation of the locomotion mode. This is a significant barrier to the clinical application of EMG pattern classification for neural control of artificial legs, as incorrect classification may cause the user to fall and suffer serious injuries. Therefore, the focus of this research is to investigate and improve the robustness of EMG signal. This includes investigating the nature of distortion and studying its effect on the classification accuracy. Moreover, an EMG distortion detector and filter is proposed that detects distortion by the fact that normally occurring EMG has a normal amplitude distribution and that distortion manifests itself as extremes or outliers of this distribution. The proposed EMG distortion detector and filter was able to remove high amplitudes distortion from EMG signal and it has resulted in 10% locomotion mode classification improvement. Additionally, a new phase detection technique is proposed that improves the reliability of phase detection system being used in phase dependent classifiers.Item Evaluation of the real-time usability of force myography as a human-computer interface(University of New Brunswick, 2018) Belyea, Alexander; Scheme, Erik; Englehart, KevinForce myography (FMG) is an alternative to electromyography (EMG) for the control of powered upper limb prostheses. FMG signals originate from deformations of muscles and surrounding tissue applying pressure to a force sensor. FMG-based pattern recognition classifiers have been shown to yield high classification rates. High classification accuracy, however, does not ensure great device usability. Instead, these control systems for prostheses should be evaluated based on real-time usability metrics. In the first section of this work a proportional control algorithm, critical to the completion of the second phase of work, was derived and compared to a mean signal amplitude-based approach. In the second, the real-time usability of high-density force myography (HDFMG) was compared to that of EMG in a Fitts’ Law virtual target acquisition task. FMG was found to significantly outperform EMG in throughput for both classification (0.901±0.357 bits/s versus 0.751±0.309 bits/s) and regression (0.871±0.325 bits/s versus 0.689±0.269 bits/s) control types. The evaluated regression-based proportional control algorithm also performed significantly better (ρ<0.001) than a standard mean signal amplitude-based approach. Subsequent data collection from an amputee subject achieved comparable classification accuracy to the able-bodied participants, but an R2 correlation coefficient of only 0.375 for regression based proportional control, significantly (ρ<0.001) lower than the able-bodied results. This work provides a comparison between the real-time usability of HD-FMG and EMG-based control in both a traditional classification-based pattern recognition scheme, with an additional proportional controller dictating device velocity and a regression-based control scheme. HD-FMG was shown to outperform EMG in both control schemes in both throughput and efficiency.Item FMG Versus EMG: A Comparison of Usability for Real-Time Pattern Recognition Based Control(Institute of Electrical and Electronics Engineers, 2019) Belyea, Alex; Englehart, Kevin; Scheme, ErikObjective: Force myography (FMG), which measures the surface pressure profile exerted by contracting muscles, has been proposed as an alternative to electromyography (EMG) for human-machine interfaces. Although FMG pattern recognition-based control systems have yielded higher offline classification accuracy, comparatively few works have examined the usability of FMG for real-time control. In this work, we conduct a comprehensive comparison of EMG- and FMG-based schemes using both classification and regression controllers. Methods: A total of 20 participants performed a two-degree-of-freedom Fitts' Law-style virtual target acquisition task using both FMG- and EMG-based classification and regression control schemes. Performance was evaluated based on the standard Fitts' law testing metrics throughput, path efficiency, average speed, number of timeouts, overshoot, stopping distance, and simultaneity. Results: The FMG-based classification system significantly outperformed the EMG-based classification system in both throughput (0.902 ± 0.270) versus (0.751 ± 0.309), (ρ <; 0.001) and path efficiency (87.2 ± 8.7) versus (83.2 ± 7.8), (ρ <; 0.001). Similarly, FMG-based regression significantly outperformed EMG-based regression in throughput (0.871 ± 0.2) versus (0.69 ± 0.3), (ρ <; 0.001) and path efficiency (64.8 ± 5.3) versus (58.8 ± 7.1), (ρ <; 0.001). Conclusions: The FMG-based schemes outperformed the EMG-based schemes regardless of which controller was used. This provides further evidence for FMG as a viable alternative to EMG for human-machine interfaces. Significance: This work describes a comprehensive evaluation of the online usability of FMG- and EMG-based control using both sequential classification and simultaneous regression control.Item High-density force myography: A possible alternative for upper-limb prosthetic control(Rehabilitation Research and Development Service, 2016) Radmand, Ashkan; Scheme, Erik; Englehart, KevinSeveral multiple degree-of-freedom upper-limb prostheses that have the promise of highly dexterous control have recently been developed. Inadequate controllability, however, has limited adoption of these devices. Introducing more robust control methods will likely result in higher acceptance rates. This work investigates the suitability of using high-density force myography (HD-FMG) for prosthetic control. HD-FMG uses a high-density array of pressure sensors to detect changes in the pressure patterns between the residual limb and socket caused by the contraction of the forearm muscles. In this work, HD-FMG outperforms the standard electromyography (EMG)-based system in detecting different wrist and hand gestures. With the arm in a fixed, static position, eight hand and wrist motions were classified with 0.33% error using the HD-FMG technique. Comparatively, classification errors in the range of 2.2%–11.3% have been reported in the literature for multichannel EMG-based approaches. As with EMG, position variation in HD-FMG can introduce classification error, but incorporating position variation into the training protocol reduces this effect. Channel reduction was also applied to the HD-FMG technique to decrease the dimensionality of the problem as well as the size of the sensorized area. We found that with informed, symmetric channel reduction, classification error could be decreased to 0.02%.Item Investigation of muscle synergies as a real-time control strategy for myoelectric control of upper limb protheses(University of New Brunswick, 2016) Atoufi, Bahareh; Englehart, Kevin; Kamavuako, Ernest NianduElectromyogram (EMG) pattern recognition has long been used for the control of powered upper limb prostheses by many researchers. However, several factors such as complexity of motions and variation of applied force challenge the robustness of pattern recognition control in practical use. Such challenging factors must be accommodated to yield truly robust performance of myoelectric control in real task oriented use. Motivated by the growing need to add functionality to current commercially available myoelectric prostheses, the current study is focused on helping with improvement of prosthetic device control toward being intuitive. The novel contribution of this research is the development and test of a platform and control algorithms based on a concept called muscle synergies which was initially introduced to explain the control strategy of the central nervous system (CNS) for coordinating muscles during motions. Based on the physiological attributes muscle synergies, the current research uses this concept toward control of prosthetic devices in a more natural feeling and physiologically expected manner. One important factor in the proportional control of prostheses is to estimate the level of muscle activity produced by the user performing the tasks. Our first study was to investigate the ability of synergies in estimating the produced force with the goal of using this estimation toward proportional control. For this aim, a regression control was performed in which the output was explicitly a single or multi-DOF estimate of force. The extracted muscle synergies demonstrated high repeatability for different repetitions of the same tasks and were quite robust across different force levels. The results indicated that muscle synergies are an effective representation of EMG in force estimation of multi-DoF tasks. Our observations strengthened the idea that predicting the forces produced in unknown levels can be possible by training the model with synergies. Also, it supported the idea that the synergies might be resilient to force changes to some extent. Evaluating their ability of force estimation in an offline test, synergies outperformed MAVs. However, a real-time control test reported no significant difference between the performance of synergies and MAVs. In an attempt to understand the dynamics of synergies, they were used as features of a pattern recognition based task classification. Also the effect of several factors such as force variation, complexity of tasks, features that synergies are extracted from, and the number of EMG channels, on the performance of synergies were examined. In general, relatively low classification errors were yielded by synergies. However, other than the cases with relatively large number of channels and synergies, the study showed that synergies’ performance was generally lagged behind that of TD features. As the performance of the proposed classification model basically depends on the choice of features and synergies, methods of producing more reliable and robust synergies were also investigated. Moreover, alternative heuristic methods were explored in an attempt to improve the synergy results outside of straightforward pattern recognition methods. Accordingly, the final study addresses the training issues and explores the classifier architecture issues. To mitigate the training issues and to improve the consistency of extracted synergies, three strategies were tested: using a validation set to select synergies, increasing the training data size, and constraining the solution space for synergy extraction method. Although, all three methods improved the results achieved by synergies, in all cases TD features still showed better performance than synergies. To explore the classifier architecture issues, strategies such as pooling synergies with TD features and extracting task specific synergies were tested. Both strategies significantly improved the previously achieved results.Item Multi-frame event dependent locomotion mode classification with FIRNNs(University of New Brunswick, 2015) Arsenault, Norman; Stevenson, Maryhelen; Englehart, KevinUsing electromyography (EMG) data obtained from four transfemoral amputees, Finite Impulse Response Neural Networks (FIRNNs) were investigated for the task of locomotion mode identification. Classification accuracy is found to improve as the duration of the observation interval, presented to the FIRNN, increases. Improvement in classifier accuracy is found to depend on the associated gait event; the more locomotion-mode transitions associated with a gait event, the higher the improvement in the classifier accuracy. Overall, the average classification accuracy on transitions improves by 15.9%. FIRNNs prove much more tolerant to increasing input dimensionality when compared with Linear Discriminant Analysis (LDA) classifiers. When Principal Component Analysis (PCA) is used to reduce input dimensionality, LDA performance is nearly equivalent to FIRNN performance without PCA. A confidence based rejection system is implemented from scaled FIRNN outputs and found to increase classification accuracy and average confidence for the nonrejected patterns. Increasing the observation interval also leads to improved confidence and reduced rejection ratios for fixed decision thresholds.Item Multi-objective user priority-based optimal tuning of myoelectric prostheses(University of New Brunswick, 2024-08) Arunachalam, Anjana Gayathri; Sensinger, Jonathon; Englehart, KevinCurrent myoelectric prosthesis control methods lack personalized tuning based on individual user preferences leading to potentially suboptimal user performance. This thesis introduces a model for personalized myoelectric prosthetic control, integrating user preferences into device dynamics through an optimization approach. Drawing on principles of computational motor control, evolutionary multi-objective optimization, and control theory, the model identifies optimal device dynamic parameters based on user preferences for effort, movement time, and reliability. Results from our simulated prosthesis model suggest that customized prosthetic devices could appreciably improve movement outcomes compared to conventional devices. This study provides a foundation for intuitive and effective prosthetic device control. These improvements may have potential applications beyond prosthesis including various human-machine interfaces.Item Multiday Evaluation of Techniques for EMG-Based Classification of Hand Motions(Institute of Electrical and Electronics Engineers, 2019-07) Waris, Asim; Niazi, Imran K.; Jamil, Mohsin; Englehart, Kevin; Jensen, Winnie; Kamavuako, Ernest NlanduCurrently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear discriminant analysis (LDA), artificial neural network (ANN), support vector machine (SVM), K-nearest neighbor (KNN), and decision trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P <; 0.001) between EMG types (surface, intramuscular, and combined), days (1-7), classifiers, and their interactions. Performance on the last day was significantly better (P <; 0.05) than the first day for all classifiers and EMG types. Within-day, classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38%), intramuscular (11.86 ± 7.84%), and combined (6.11 ± 7.46%). The between-day analysis in a leave-one-day-out fashion showed that the ANN was the optimal classifier surface (21.88 ± 4.14%), intramuscular (29.33 ± 2.58%), and combined (14.37 ± 3.10%). Results indicate that within day performances of classifiers may be similar but over time, it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration.Item Multiday Evaluation of Techniques for EMG-Based Classification of Hand Motions(Institute of Electrical and Electronics Engineers, 2019-07) Waris, Asim; Niazi, Imran K.; Jamil, Mohsin; Englehart, Kevin; Jensen, Winnie; Kamavuako, Ernest NlanduCurrently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear discriminant analysis (LDA), artificial neural network (ANN), support vector machine (SVM), K-nearest neighbor (KNN), and decision trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P <; 0.001) between EMG types (surface, intramuscular, and combined), days (1-7), classifiers, and their interactions. Performance on the last day was significantly better (P <; 0.05) than the first day for all classifiers and EMG types. Within-day, classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38%), intramuscular (11.86 ± 7.84%), and combined (6.11 ± 7.46%). The between-day analysis in a leave-one-day-out fashion showed that the ANN was the optimal classifier surface (21.88 ± 4.14%), intramuscular (29.33 ± 2.58%), and combined (14.37 ± 3.10%). Results indicate that within day performances of classifiers may be similar but over time, it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration.Item Multivariate approach for assessing electrode positioning in surface electromyography(University of New Brunswick, 2023-08) Nsofor, Chitom Clare; MacIsaac, Dawn T.; Englehart, KevinPlacing electrodes near or over an innervation zone has been shown to affect the quality and integrity of the recorded signals. The aim of this work was to investigate whether using pattern recognition to analyze electromyography data could lead to an automated approach for estimating the graded effect of an innervation zone on surface electromyography signal. Using a set of features from simulated electromyography signals as input, classification and regression algorithms were explored to predict the graded effect of an innervation zone. The regression technique was observed to be best suited for the application. The effects of physiological parameter variability between the training and test data sets were also investigated. Some physiological parameters, especially the innervation zone distribution and conduction velocity, were found to have the most impact on the performance of the regressor. Regression is a promising approach for subsequent research, especially with recorded data.Item Myoelectric pattern recognition state-based classification method for improved dynamic performance and real-time usability(University of New Brunswick, 2016) Biron, Katerina; Englehart, Kevin; Scheme, ErikPattern-recognition-based myoelectric control systems have been shown to be accurate in controlled laboratory experiments where users are often restricted to perform discrete, segmented and constant force contractions. Such types of contractions are insufficient for real functional use that requires a controller to interpret both constant force contractions and dynamic transitions between motion classes. This explains why pattern recognition based myoelectric control systems have often been unreliable in real-world settings and the dynamic conditions required for functional use. This work develops an original pattern recognition based myoelectric control system that improves system performance during dynamic transitions between motion classes. The proposed architecture uses feature estimate methods inspired by Kalman filters and enables different control strategies depending on whether users perform constant force contractions or dynamic transitions. The system was developed offline and was then evaluated in a real-time (user-in-the-loop) task to verify if it improved usability. For both offline and real-time analyses, the system was compared to state-of-the-art pattern recognition systems based on linear discriminant analysis. The real-time results showed that the proposed system allowed users to perform significantly (p < .05) more tasks in a significantly (p < .05) less amount of time, and the proposed system also obtained a better speed-accuracy tradeoff, suggesting improved usability.Furthermore, the system improved the quality of descriptive features and allowed new users to better learn how to control pattern recognition based myoelectric control system.Item On the feasibility of using pattern recognition based myoelectric control as a human-computer interface for individuals with paralysis(University of New Brunswick, 2016) Chaulk, Mitchell Charles; Scheme, Erik; Englehart, KevinHuman-computer interfaces (HCIs), using electromyogram (EMG) data for control, has been studied for decades as a potential means of restoring functional ability to amputees. Often, these HCIs are used to control a powered prosthesis. However, this technology has potential application outside of the scope of prosthetics. The EMG produced by people with neurological damage could contain enough discriminatory information to distinguish between many classes of motion, including those that they cannot functionally perform. In this study, 10 individuals with spinal cord injuries (SCIs) around the C3-C6 level (ASIA A-C) volunteered to have their EMG studied while performing 10 different classes of motion with their dominant upper limb. Preliminary studies, using high-density EMG, were performed on two volunteers before moving on to using an electrode cuff with 8 bipolar channels. Performing pattern recognition, for the 10 classes, using an LDA classifier referencing 5 features (sample entropy, mean absolute value, zero crossings, slope sign change, and wave length) resulted in a total mean accuracy of 91.5%. This accuracy was increased to 98.0% when evaluating a set of 5 classes. These 5 classes were chosen based on the classes available by the Bioness H200 device, which uses functional stimulation to force user contractions. Such a device could benefit from an accurate EMG controller.Item On the robustness of real-time myoelectric control investigations: a multiday Fitts’ law approach(IOP Publishing, 2019) Waris, Asim; Mendez, Irene; Englehart, Kevin; Jensen, Winnie; Kamavuako, Ernest NlanduObjective. Real-time myoelectric experimental protocol is considered as a means to quantify usability of myoelectric control schemes. While usability should be considered over time to assure clinical robustness, all real-time studies reported thus far are limited to a single session or day and thus the influence of time on real-time performance is still unexplored. In this study, the aim was to develop a novel experimental protocol to quantify the effect of time on real-time performance measures over multiple days using a Fitts' law approach. Approach. Four metrics: throughput, completion rate, path efficiency and overshoot, were assessed using three train-test strategies: (i) an artificial neural network (ANN) classifier was trained on data collected from the previous day and tested on present day (BDT) (ii) trained and tested on the same day (WDT) and (iii) trained on all previous days including present day and tested on present day (CDT) in a week-long experimental protocol. Main results. It was found that on average, the completion rate (98.37% ± 1.47%) of CDT was significantly better (P < 0.01) than that of BDT (86.25% ± 3.46%) and WDT (94.22% ± 2.74%). The throughput (0.40 ± 0.03 bits s−1) of CDT was significantly better (P = 0.001) than that of BDT (0.38 ± 0.03 bits s−1). Offline analysis showed a different trend due to the difference in the training strategies. Significance. Results suggest that increasing the size of the training set over time can be beneficial to assure robust performance of the system over time.Item Real-time simultaneous myoelectric control of multiple degrees of freedom(University of New Brunswick, 2014) Ameri, Ali; Englehart, Kevin; Parker, Philip; Scheme, ErikNatural limb control during activities of daily living involves simultaneous motions of multiple degrees of freedom (DOFs) (such as feeding or grooming, which require simultaneous activation of hand, arm, and wrist in multiple directions). A strategy for simultaneous myoelectric control is to train the system with EMG as the input and the corresponding joint force or position (angle) as the target. For prosthesis control, since it is not possible to measure force/position from an absent limb, a strategy called mirrored bilateral training was proposed (with unilateral amputees) in previous work, in which force/position was recorded from the opposite limb during mirrored contractions. In this work, the effect of alternative feature sets, estimators, feature projection, and coordinate systems on estimation performance was studied with force and position based paradigms. Furthermore, a real-time control test was performed to assess the system usability. It was shown that when the EMG (and effort) levels were similar, no significant difference (p>0.1) was found between the force and position based methods in both offline and real-time control tests. A novel training paradigm for simultaneous control is described, in which users were prompted to synchronize their contractions with a moving target cursor on a computer screen. The cursor displacements were used as targets to train the estimators. The system usability was assessed with a real-time control test, and the performance was found to be equivalent (p>0.1) to that of the mirrored bilateral training. The proposed visual target based training is more practical than mirrored training because it does not require force and position sensing equipment, and can be potentially used by both unilateral and bilateral amputees. Finally, a novel application of a support vector machine (SVM) was evaluated in simultaneous myoelectric real-time control of DOFs. It was shown with able-bodied and amputee subjects that the proposed SVM based method outperformed the widely used multilayer perceptron artificial neural network (ANN) in a Fitts’ law style real-time control test. Moreover, the processing time required for training and estimation with the SVM was significantly lower than that of the ANN. This approach is shown to provide a robust and computationally efficient system for simultaneous and proportional myoelectric control.