Heba Lakany

Extracting a diagnostic gait signature

This research addresses the question of the existence of prominent diagnostic signatures for human walking extracted from kinematics gait data. The proposed method is based on transforming the joint motion trajectories using wavelets to extract spatio-temporal features which are then fed as input to a vector quantiser; a self-organising map for classification of walking patterns of individuals with and without pathology. We show that our proposed algorithm is successful in extracting features that successfully discriminate between individuals with and without locomotion impairment.

EEG decoding of semantic category reveals distributed representations for single concepts

Achieving a clearer picture of categorial distinctions in the brain is essential for our understanding of the conceptual lexicon, but much more fine-grained investigations are required in order for this evidence to contribute to lexical research. Here we present a collection of advanced data-mining techniques that allows the category of individual concepts to be decoded from single trials of EEG data. Neural activity was recorded while participants silently named images of mammals and tools, and category could be detected in single trials with an accuracy well above chance, both when considering data from single participants, and when group-training across participants. By aggregating across all trials, single concepts could be correctly assigned to their category with an accuracy of 98%. The pattern of classifications made by the algorithm confirmed that the neural patterns identified are due to conceptual category, and not any of a series of processing-related confounds. The time intervals, frequency bands and scalp locations that proved most informative for prediction permit physiological interpretation: the widespread activation shortly after appearance of the stimulus (from 100 ms) is consistent both with accounts of multi-pass processing, and distributed representations of categories. These methods provide an alternative to fMRI for fine-grained, large-scale investigations of the conceptual lexicon.

Understanding intention of movement from electroencephalograms

In this paper, we propose a new framework for understanding intention of movement that can be used in developing non-invasive brain-computer interfaces. The proposed method is based on extracting salient features from brain signals recorded whilst the subject is actually (or imagining) performing a wrist movement in different directions. Our method focuses on analysing the brain signals at the time preceding wrist movement, i.e. while the subject is preparing (or intending) to perform the movement. Feature selection and classification of the direction is done using a wrapper method based on support vector machines (SVMs). The classification results show that we are able to discriminate the directions using features extracted from brain signals prior to movement. We then extract rules from the SVM classifiers to compare the features extracted for real and imaginary movements in an attempt to understand the mechanisms of intention of movement. Our new approach could be potentially useful in building brain-computer interfaces where a paralysed person could communicate with a wheelchair and steer it to the desired direction using a rule-based knowledge system based on understanding of the subjects intention to move through his/her brain signals.

A fuzzy inference system for fault detection and isolation: Application to a fluid system

This work focuses on the design and implementation of a fuzzy inference system for fault detection and isolation (FDI) which can learn from example fault data, and the determination of a suitable optimisation strategy for the membership functions. A FDI system was developed which is based on adaptive fuzzy rules. A number of optimisation strategies were then applied; it was found that an evolutionary algorithm not only produced the best results but did so with relatively little processing effort and with excellent consistency. The adaptive fuzzy system, thus optimised, was tested against a neural network, which was trained to produce analogue outputs as an indication of fault magnitude. The fuzzy solution produced the best accuracy. We can conclude that an adaptive fuzzy inference system for FDI, using an evolutionary algorithm to learn from examples, can provide an accurate and readily comprehensible solution to diagnosing and evaluating fluid process plant faults.

A generic kinematic pattern for human walking

Abstract The aim of this work is to investigate the existence of a generic feature vector based on kinematic data for normal walking. The paper describes a method to quantify generic features of the sagittal angles of the lower extremities of human subjects. The idea is to extract salient features from hip, knee and ankle sagittal angles to characterise normal and pathological walking. The algorithm is based on transforming the trajectories of the flexion/extension of joints of subjects using the continuous wavelet transform to represent a feature vector which is then fed to a self-organising map for clustering. The algorithm proved to be successful in distinguishing between normal subjects according to their age group, gender and also distinguishing between normal and pathological subjects. Rules are extracted from self-organising map to determine the salient features characterising each cluster as well as differentiating it from others.

Human Behavior Integration Improves Classification Rates in Real-Time BCI

Brain-computer interfaces (BCI) offer potential for individuals with a variety of motor and sensory disabilities to interact with their environment, communicate and control mobility aids. Two key factors which affect the performance of a BCI and its usability are the feedback given to the participant and the subjects motivation. This paper presents the results from a study investigating the effects of feedback and motivation on the performance of the Strathclyde Brain Computer Interface. The paper discusses how the performance of the system can be improved by behavior integration and human-in-the-loop design.

The strathclyde brain computer interface

Brain-computer interfaces (BCI) offer potential for individuals with a variety of motor and sensory disabilities to control their environment, communicate, and control mobility aids. However, the key to BCI usability rests in being able to extract relevant time varying signals that can be classified into usable commands in real time. This paper reports the first success of the Strathclyde BCI controlling a wheelchair on-line in Virtual Reality. Surface EEG recorded during wrist movement in two different directions were classified and used to control a wheelchair within a virtual reality environment. While Principal Component Analysis was used for feature vector quantiser distances were used for classification. Classification success rates between 68% and 77% were obtained using these relatively simple methods.

An Algorithm for Recognising Walkers

In this paper, we present an algorithm to recognise walking people, based upon extracting the spatio-temporal trajectories of the joints of a walking subject.

An Investigation Into the Application of Fuzzy Logic Control to Industrial Gas Turbines

This paper investigates the feasibility benefits of applying fuzzy logic control (FLC) strategies for use with industrial gas turbines. Our main objective is to investigate different designs methods, design implement an FLC strategy, plant simulation in a test environment optimize the FLC, conduct tests to compare the FLC with conventional controls in scenarios relevant to the application. We have designed, implemented, and tested our simulation to the exhaust temperature control problem of a gas turbine problem. The FLC, plant simulation, existing control configuration, and integrated test environment were developed in Java. Heuristic methods were used to optimize the FLC, which proved time consuming. The paper illustrates that while implementation of the FLC is feasible, it requires more effort than the conventional controls examined.

Classification of Wrist Movements using EEG-based Wavelets Features

Our aim is to assess and evaluate signal processing and classification methods for extracting features from EEG signals that are useful in developing brain-computer interfaces. In this paper, we report on results of developing a method to classify wrist movements using EEG signals recorded from a subject whilst controlling a joystick and moving it in different directions. Such method could be potentially useful in building brain-computer interfaces (BCIs) where a paralysed person could communicate with a wheelchair and steer it to the desired direction using only EEG signals. Our method is based on extracting salient spatio-temporal features from the EEG signals using continuous wavelet transform. We perform principal component analysis on these features as means to assess their usefulness for classification and to reduce the dimensionality of the problem. We use the results from the PCA as means to represent the different directions. We use a simple technique based on Euclidean distance to classify the data. The classification results show that we are able to discriminate between different directions using the selected features