Computational intelligence based EMG for automated classification of foot drop rehabilitation

Publication Type:
Issue Date:
Full metadata record
Foot drop is a complication that arises from the weakness that occurs in specific muscles in the ankle and foot; such as the Anterior Tibialis muscle (AT) during the foot flexion and extension. A lesion in the Lower Motor Neuron (LMN) will cause Foot Drop. Foot drop has been found to arise in 52% to 67% of patients with spinal Upper Motor Neuron (UMN) pathology. Foot Drop (FD) is a common disorder and is not specific to age; it affects around 1% of women and 2.8 % of men. The affected muscles impact on the motion of the ankle and foot both downward and upward. To overcome this problem and to improve rehabilitation devices, this thesis introduces several methods to improve the performance of the Myoelectric Pattern Recognition (M-PR) in both the offline and real data experiments. The thesis proposes a new M-PR system that will work satisfactorily on both the healthy and non-healthy leg by classifying movements in the offline experiments. The thesis describes the state-of-the-art procedures for M-PR and studies all the conceivable structures for the best M-PR features and classifiers. This thesis presents new classifiers to develop the performance of the M-PR. The Self-Advised-Support Vector Machine (SA-SVM) modified from a single class to multiclass. Developments in methodology lead to more significant applications for overall use. The thesis adapted and altered label classification methods resulting in a new classification named Label Self- Advised-Support Vector Machine LSA-SVM. Further, a development to LSA-SVM is to upgrade from a single class into Multi-LSA-SVM and then to evolve the methodology to match Extreme Learning Machine (ELM) with LSASVM to acquire a new rapid method, named ELM-LSA-SVM. For the real data experimental option, a collected data from using the sEMG device from Foot Drop Patients in Metro Rehabilitation Hospital in Sydney, Australia using Ethical Approval (UTS HREC NO.ETH15-0152). The collected used to apply the latest and fastest method to FD patients to use the myoelectric pattern recognition (M-PR) for leg movement detection. The experimental results for the EMG dataset and benchmark datasets exhibit its benefits. Furthermore, the experimental results on UCI datasets indicate that ELM-LSA-SVM achieves the best performance when working together with LSA-SVM and SVM. The whole sets of experimental results are encouraging as recorded and reported in the thesis.
Please use this identifier to cite or link to this item: