Search

The purpose of this study was to investigate differences in lower limb muscle activation patterns for females wearing shoes with different heel heights during Sit to Stand Task (STS). Ten female participants with no prior history of neurological disorders participated in this study. Surface electromyography (sEMG) characteristics were recorded for four different heel heights (ranging from 4cm to 10cm) while performing the STS task. Signal processing analysis suggests that muscle activities increases on elevated heel heights, which may induce muscle imbalance for frequent STS tasks. In addition, results of muscle utilisation (percentage) for different heel heights suggest that lower limb muscles tend to compensate in order to maintain postural balance.

In recent years, significant advancements have taken place in human activity recognition using various machine learning approaches. However, feature engineering have dominated conventional methods involving the difficult process of optimal feature selection. This problem has been mitigated by using a novel methodology based on deep learning framework which automatically extracts the useful features and reduces the computational cost. As a proof of concept, we have attempted to design a generalized model for recognition of three fundamental movements of the human forearm performed in daily life where data is collected from four different subjects using a single wrist worn accelerometer sensor. The validation of the proposed model is done with different pre-processing and noisy data condition which is evaluated using three possible methods. The results show that our proposed methodology achieves an average recognition rate of 99.8% as opposed to conventional methods based on K-means clustering, linear discriminant analysis and support vector machine.

This paper introduces a novel shape memory alloy (SMA) material for the controllability in the shape recovery of traditional SMA for orthopedic devices and rehabilitation techniques. The proposed material is formed by doping nano-ferromagnetic particle into porous NiTi alloy. The finite element analysis of shape memory effect property of the different distribution of nano-ferromagnetic particle is done and compared for same load and boundary conditions. The comparative analysis of the percentage change in volume deformation when load is released (for 2 nd step) shows an average of 2.55 % with standard deviation of 1.69 whereas on thermal loading (for 3 rd step) shows an average of 94.94% with standard deviation of 7.75 for all heterogeneous distribution of nano-particles in porous NiTi alloy. Our findings are, all the different conditions of heterogeneous distributions of nano-ferromagnetic particle doped NiTi alloy exhibits its inherent SME property.

Freezing of Gait (FOG) is a highly debilitating and poorly understood symptom of Parkinson's disease (PD), causing severe immobility and decreased quality of life. Turning Freezing (TF) is known as the most common sub-type of FOG, also causing the highest rate of falls in PD patients. During a TF, the feet of PD patients appear to become stuck whilst making a turn. This paper presents an electroencephalography (EEG) based classification method for detecting turning freezing episodes in six PD patients during Timed Up and Go Task experiments. Since EEG signals have a time-variant nature, time-frequency Stockwell Transform (S-Transform) techniques were used for feature extraction. The EEG sources were separated by means of independent component analysis using entropy bound minimization (ICA-EBM). The distinctive frequency-based features of selected independent components of EEG were extracted and classified using Bayesian Neural Networks. The classification demonstrated a high sensitivity of 84.2%, a specificity of 88.0% and an accuracy of 86.2% for detecting TF. These promising results pave the way for the development of a real-time device for detecting different sub-types of FOG during ambulation.

This paper presents a classification of driver fatigue with electroencephalography (EEG) channels selection analysis. The system employs independent component analysis (ICA) with scalp map back projection to select the dominant of EEG channels. After channel selection, the features of the selected EEG channels were extracted based on power spectral density (PSD), and then classified using a Bayesian neural network. The results of the ICA decomposition with the back-projected scalp map and a threshold showed that the EEG channels can be reduced from 32 channels into 16 dominants channels involved in fatigue assessment as chosen channels, which included AF3, F3, FC1, FC5, T7, CP5, P3, O1, P4, P8, CP6, T8, FC2, F8, AF4, FP2. The result of fatigue vs. alert classification of the selected 16 channels yielded a sensitivity of 76.8%, specificity of 74.3% and an accuracy of 75.5%. Also, the classification results of the selected 16 channels are comparable to those using the original 32 channels. So, the selected 16 channels is preferable for ergonomics improvement of EEG-based fatigue classification system.

A great many people suffer from neurological movement disorders that render typical hardware interface devices ineffective. A need exists for a universal interface device that can be trained to accept a wide range of inputs across varying types and severities of movement disorders. In this regard, this paper details the design, testing and optimization of an accelerometer-based gesture identification system. A Bluetooth-enabled IMU mounted on the wrist provides hand motion trajectory information to a local terminal. Several techniques are applied to decrease the intra-class variance and reduce classifier complexity including filtering, segmentation and temporal scaling. Datasets consisted of 520 training samples, 260 validation samples and a further 520 testing samples. A multi-layer feed forward artificial neural network (ML-FFNN) was used to classify the input space into 26 different classes. Initial system accuracy, using arbitrary hyperparameters was 77.69% with final optimized accuracy at 99.42%.

Gait initiation Failure (GIF) is the situation in which patients with Parkinson's disease (PD) feel as if their feet get "stuck" to the floor when initiating their first steps. GIF is a subtype of Freezing of Gait (FOG) and often leads to falls and related injuries. Understanding of neurobiological mechanisms underlying GIF has been limited by difficulties in eliciting and objectively characterizing such gait phenomena in the clinical setting. Studies investigating the effects of GIF on brain activity using EEG offer the potential to study such behavior. In this preliminary study, we present a novel methodology where wavelet transform was used for feature extraction and Support Vector Machine for classifying GIF events in five patients with PD and FOG. To deal with the large amount of EEG data, a Principal Component Analysis (PCA) was applied to reduce the data dimension from 15 EEG channels into 6 principal components (PCs), retaining 93% of the information. Independent Component Analysis using Entropy Bound Minimization (ICA-EBM) was applied to 6 PCs for source separation with the aim of improving detection ability of GIF events as compared to the normal initiation of gait (Good Starts). The results of this analysis demonstrated the correct identification of GIF episodes with an 83.1% sensitivity, 89.5% specificity and 86.3% accuracy. These results suggest that our proposed methodology is a promising non-invasive approach to improve GIF detection in PD and FOG.

This paper presents an electroencephalography (EEG) based-classification of between pre- and post-mental load tasks for mental fatigue detection from 65 healthy participants. During the data collection, eye closed and eye open tasks were collected before and after conducting the mental load tasks. For the computational intelligence, the system uses the combination of principal component analysis (PCA) as the dimension reduction method of the original 26 channels of EEG data, power spectral density (PSD) as feature extractor and Bayesian neural network (BNN) as classifier. After applying the PCA, the dimension of the data is reduced from 26 EEG channels in 6 principal components (PCs) with above 90% of information retained. Based on this reduced dimension of 6 PCs of data, during eyes open, the classification pre-task (alert) vs. post-task (fatigue) using Bayesian neural network resulted in sensitivity of 76.8 %, specificity of 75.1% and accuracy of 76% Also based on data from the 6 PCs, during eye closed, the classification between pre- and post-task resulted in a sensitivity of 76.1%, specificity of 74.5% and accuracy of 75.3%. Further, the classification results of using only 6 PCs data are comparable to the result using the original 26 EEG channels. This finding will help in reducing the computational complexity of data analysis based on 26 channels of EEG for mental fatigue detection.

In this paper we introduce Shape Memory Alloy (SMA) for designing the tibial part of Total Knee Arthroplasty (TKA) by exploiting the shape-memory and pseudo-elasticity property of the SMA (e.g. NiTi). This would eliminate the drawbacks of the state-of-the art PMMA based knee-spacer including fracture, sustainability, dislocation, tilting, translation and subluxation for tackling the Osteoarthritis especially for the aged people of 45-plus or the athletes. In this paper a Computer Aided Design (CAD) model using SolidWorks for the knee-spacer is presented based on the proposed SMA adopting the state-of-the art industry-standard geometry that is used in the PMMA based spacer design. Subsequently Ansys based Finite Element Analysis is carried out to measure and compare the performance between the proposed SMA based model with the state-of-the art PMMA ones. 81% more bending is noticed in the PMMA based spacer compared to the proposed SMA that would eventually cause fracture and tilting or translation of spacer. Permanent shape deformation of approximately 58.75% in PMMA based spacer is observed compared to recoverable 11% deformation in SMA when same load is applied on both separately.

Nowadays portable devices with more number of sensors are used for gait assessment and monitoring for elderly and disabled. However, the problem with using multiple sensors is that if they are placed on the same platform or base, there could be cross talk between them, which could change the signal amplitude or add noise to the signal. Hence, this study uses wavelet PCA as a signal processing technique to separate the original sensor signal from the signal obtained from the sensors through the integrated unit to compare the two types of walking (with and without an exoskeleton). This comparison using wavelet PCA will enable the researchers to obtain accurate sensor data and compare and analyze the data in order to further improve the design of compact portable devices used to monitor and assess the gait in stroke or paralyzed subjects. The advantage of designing such systems is that they can also be used to assess and monitor the gait of the stroke subjects at home, which will save them time and efforts to visit the laboratory or clinic.

In this paper we propose an on-the-fly reconfigurable hardware-software codesign based reconfigurable solution for real-time protein identification. Reconfigurable string matching is performed in the disciplines of protein identification and biomarkers discovery. With the generation of plethora of sequenced data and number of biomarkers for several diseases, it is becoming necessary to have an accelerated processing and on-the-fly reconfigurable system design methodology to bring flexibility to its usage in the medical science community without the need of changing the entire hardware every time with the advent of new biomarker or protein. The proteome database of human at UniProtKB (Proteome ID up000005640) comprising of 42132 canonical and isoform proteins with variable database-size are used for testing the proposed design and the performance of the proposed system has been found to compare favorably with the state-of-the-art approaches with the additional advantage of real-time reconfigurability.

An electroencephalography (EEG)-based counter measure device could be used for fatigue detection during driving. This paper explores the classification of fatigue and alert states using power spectral density (PSD) as a feature extractor and fuzzy swarm based-artificial neural network (ANN) as a classifier. An independent component analysis of entropy rate bound minimization (ICA-ERBM) is investigated as a novel source separation technique for fatigue classification using EEG analysis. A comparison of the classification accuracy of source separator versus no source separator is presented. Classification performance based on 43 participants without the inclusion of the source separator resulted in an overall sensitivity of 71.67%, a specificity of 75.63% and an accuracy of 73.65%. However, these results were improved after the inclusion of a source separator module, resulting in an overall sensitivity of 78.16%, a specificity of 79.60% and an accuracy of 78.88% (p <; 0.05).

This study uses multiscale principal component analysis (MSPCA) signal processing technique in order to distinguish the two different surfaces, tiled (regular) and cobbled (irregular) using accelerometry data (recorded from MTx sensors). Two MTx sensors were placed on the head and trunk of the subject while the subject walked freely over the regular and irregular surfaces during a free walk. 3D acceleration signals, vertical, medio lateral (ML) and anterior-posterior (AP) were recorded for the head and trunk segments and compared for the free walk on a defined route. The magnitude of the ML and AP acceleration obtained from the MTx sensors (for both head & trunk) was higher when walking over the irregular (cobbled) surface as compared to the regular (tiled) surface. The accelerometry data was initially analysed using MSPCA and was later classified using naïve Bayesian classifier with >86% accuracy. This research study demonstrates that MSPCA can be used to distinguish the regular and irregular surfaces. The proposed method could be very useful as an automated method for classification of the two surfaces.

In this paper, we propose a high speed architecture design methodology for the Under-determined Blind Source Separation (UBSS) algorithm using our recently proposed high speed Discrete Hilbert Transform (DHT) targeting real time applications. In UBSS algorithm, unlike the typical BSS, the number of sensors are less than the number of the sources, which is of more interest in the real time applications. The DHT architecture has been implemented based on sub matrix multiplication method to compute M point DHT, which uses N point architecture recursively and where M is an integer multiples of N. The DHT architecture and state of the art architecture are coded in VHDL for 16 bit word length and ASIC implementation is carried out using UMC 90 - nm technology @V DD = 1V and @ 1MHZ clock frequency. The proposed architecture implementation and experimental comparison results show that the DHT design is two times faster than state of the art architecture.

Electroencephalograms (EEGs) are progressively emerging as a significant measure of brain activity and are very effective tool for the diagnosis and treatment of mental and brain diseases and disorders including sleep apnea, Alzheimer's disease and Neurodevelopmental disorders. However, EEG signal is mixed with other biological signals including Ocular and Muscular artefacts making it difficult to extract the diagnostic features. Therefore, the contaminated EEG channels are often discarded by the medical practitioners resulting less accurate diagnosis. In this paper we propose a real-time low-complexity and reliable system design methodology to remove these artefacts and noise in an automated fashion to aid online diagnosis under the pervasive personalized healthcare set-up without the need of any reference electrode. The simulation and hardware performance of the proposed methodology are measured and compared in terms of correlation and regression statistics lying above 80% and 67% which are much improved over the state-of-the art methodologies.

This paper introduces a dual-mode low complex on-chip methodology for processing of ECG (Electrocardiogram) and EEG (Electroencephalography) signals, wherein based on the input switch the architecture can be dynamically configured to operate either as an ECG bio-marker or EEG signal de-noising system. In both the modes the signal processing technique depends on the output of the DWT (Discrete Wavelet Transform), hence a low complex methodology has been developed in which both ECG and EEG processing blocks sharing the same DWT block resulting in low area and low power consumption. The integrated ECG and EEG methodology has been implemented in Matlab, for verifying the ECG processing block the ECG database is taken from MIT-BIH PTBDB and IITH DB, similarly for EEG processing block the EEG signals are taken from PhysioNet database. The outcome of methodology in Matlab is equal to the results obtained from individual ECG and EEG blocks.

This paper reports the classification of finger flexion and extension of surface Electromyography (EMG) and Cyberglove data using the modified Independent Component Analysis (ICA) weight matrix. The finger flexion and extension data are processed through Principal Component Analysis (PCA), and next separated using modified ICA for each individual with customized weight matrix. The extension and flexion features of sEMG and Cyberglove (extracted from modified ICA) were classified using Linear Discriminant Analysis (LDA) with near 90% classification accuracy. The applications of this study include Human Computer Interface (HCI), virtual reality and neural prosthetics.

The proposed signal processing technique uses Fast Independent Component Analysis (ICA) algorithm to improve the quality of the original biosensors recordings, which can be used as valuable pre-processing technique such as cross talk removal, artefact reduction etc. Initially, the ill conditioned original surface Electromyography (sEMG) recordings were separated using ICA methods and later they were reconstructed using modified un-mixing matrix. The simulation results showed huge improvement of the original recorded signal after reconstruction. The proposed method has potential applications in various biomedical signal processing techniques.

Single Channel Blind Source Separation (SCBSS) is an extreme case of underdetermined (more sources and fewer sensors) Blind Source Separation (BSS) problem. In this paper, we propose a novel technique using Local Mean Decomposition (LMD) and Independent Component Analysis (ICA) combined with single channel BSS (LMD&#95;ICA). First, the LMD was used to decompose the single channel source into a series of data sequences, which are called as Product Functions (PF), then, ICA algorithm was used to process PFs to get similar independent components and extract the original signals. A comparison was made between LMD&#95;ICA and previously proposed single channel ICA method (EEMD&#95;ICA). The real time experimental results demonstrated the advantage of the proposed single channel source separation method for artifact removal and in biomedical source separation applications.

The electromyograpy (EMG) signal provides information about the performance of muscles and nerves. The shape of the muscle signal and motor unit action potential (MUAP) varies due to the movement of the position of the electrode or due to changes in contraction level. This research deals with evaluating the non-Gaussianity in Surface Electromyogram signal (sEMG) using higher order statistics (HOS) parameters. To achieve this, experiments were conducted for four different finger and wrist actions at different levels of Maximum Voluntary Contractions (MVCs). Our experimental analysis shows that at constant force and for non-fatiguing contractions, probability density functions (PDF) of sEMG signals were non-Gaussian. For lesser MVCs (below 30% of MVC) PDF measures tends to be Gaussian process. The above measures were verified by computing the Kurtosis values for different MVCs.

This research paper reports an experimental study on identification of the changes in fractal properties of surface Electromyogram (sEMG) with the changes in the force levels during low-level finger flexions. In the previous study, the authors have identified a novel fractal feature, Maximum fractal length (MFL) as a measure of strength of low-level contractions and has used this feature to identify various wrist and finger movements. This study has tested the relationship between the MFL and force of contraction. The results suggest that changes in MFL is correlated with the changes in contraction levels (20%, 50% and 80% maximum voluntary contraction (MVC)) during low-level muscle activation such as finger flexions. From the statistical analysis and by visualisation using box-plot, it is observed that MFL (p ≈ 0.001) is a more correlated to force of contraction compared to RMS (p≈0.05), even when the muscle contraction is less than 50% MVC during low-level finger flexions. This work has established that this fractal feature will be useful in providing information about changes in levels of force during low-level finger movements for prosthetic control or human computer interface.

Classification of surface electromyogram (sEMG) signal is important for various applications such as prosthetic control and human computer interface. Surface EMG provides a better insight into the strength of muscle contraction which can be used as control signal for different applications. Due to the various interference between different muscle activities, it is difficult to identify movements using sEMG during low-level flexions. A new set of fractal features - fractal dimension and Maximum fractal length of sEMG has been previously reported by the authors. These features measure the complexity and strength of the muscle contraction during the low-level finger flexions. In order to classify and identify the low-level finger flexions using these features based on the fractal properties, a recently developed machine learning based classifier, Twin Support vector machines (TSVM) has been proposed. TSVM works on basic learning methodology and solves the classification tasks as two SVMs for each classes. This paper reports the novel method on the machine learning based classification of fractal features of sEMG using the Twin Support vector machines. The training and testing was performed using two different kernel functions - Linear and Radial Basis Function (RBF).

Spectral compression of surface electromyogram (sEMG) is associated with onset of localized muscle fatigue. The spectral compression has been explained based on motor unit synchronization theory. According to this theory, motor units are pseudo randomly excited during muscle contraction, and with the onset of muscle fatigue the recruitment pattern changes such that motor unit firings become more synchronized. While this is widely accepted, there is little experimental proof of this phenomenon. This paper has used source dependence measures developed in research related to independent component analysis (ICA) to test this theory.

This paper has experimentally verified and compared features of sEMG (Surface Electromyogram) such as ICA (Independent Component Analysis) and Fractal Dimension (FD) for identification of low level forearm muscle activities. The fractal dimension was used as a feature as reported in the literature. The normalized feature values were used as training and testing vectors for an Artificial neural network (ANN), in order to reduce inter-experimental variations. The identification accuracy using FD of four channels sEMG was 58%, and increased to 96% when the signals are separated to their independent components using ICA.

Source separation and identification is one of the challenging areas in the bio signal processing. The processing of Electromyographic (EMG) signals can be viewed as the identification and separation of a series of overlapping sources of muscle activity with slowly varying source distribution and/or levels of activity. Blind source separation (BSS) techniques such as independent component analysis (ICA) lend themselves well to the analysis of such problems. The problem, however, still remains largely ill-posed even through the use of powerful assumptions such as those posed in ICA and other such techniques. It is generally the case in EMG signals that a certain level of a priori knowledge is available on the spatio-temporal and/or frequency distribution of the activities of interest, based on neurophysiological expectations. Here we describe limitations and applications of BSS on surface EMG. The problems we consider include the analysis of facial sEMG recordings during vowel utterance and analysis of hand EMG during finger and wrist movements.

To ensure that no biological event that may be important is missed, redundancy of sensors is provided. While this is useful, there are shortcomings when there is need to separate the signals from different sources using blind source separation techniques. An example of such a situation is over-complete surface electromyogram (sEMG) recording. Techniques such as principal component analysis (PCA) and entropy measures are used to identify the suitable channels. The shortcomings in these are the need for prior estimation of the number of channels. This paper has used the determinant of the global matrix of the mixtures to determine the number of independent sources in a mixture. The results indicate that the technique is able to distinguish between dependent and independent channels and this may be applied for determining the number of independent sources. The applications of this include data reduction by identifying redundant data, and for pre-processing of the data prior to use of any data classification techniques.