Markerless Human Gait Analysis using Ultra-Wideband Radar approaches augmented by Artificial Intelligence at London South Bank University

This fully funded PhD project builds on fundamental work by the South Bank Applied BioEngineering Research Centre (SABER) on self-regulating and non-contact impulse radio ultra-wideband (IR-UWB) based human gait analysis augmented by artificial intelligence. The work provisions a non-tactile assistive biomedical application, which can be placed at home or at multiple field sites, for remote human gait monitoring, and action classification to identify traits and abnormalities negating wearable sensor needs. This research brings together AI, UWB radar technologies, and biomechanics, developing a low-cost, robust, and reliable UWB radar-based gait analysis system. This work is beyond state-of-the-art and has formed LSBU’s unique selling point in two EU-funded research projects hitherto.

The research comprises the subjective human motion data collection from able-bodied participants and those suffering from gait pathologies, using existing IR-UWB radar in both anechoic chamber and multi-path environments. The physical gait parameters must be determined via signal processing, using theoretical, and subsequently computational study. Other state-of-art marker-based systems will be available in-house to validate the gait parameters obtained from LSBU’s IR-UWB system. The extracted parameters will be further employed to recognise the gait patterns through machine learning and deep learning methods. This PhD studentship involves a combination of theoretical, computational, and experimental work, providing the student with complete and holistic training and a diverse research experience. 

Keywords: – Gait analysis, IR-UWB, Signal Processing, Mathematical Modelling, Machine Learning, and Deep Learning

This project will concentrate on:

1. Theoretical Modelling to Determine Physical Gait Features: Investigate the current UWB Radar system architecture and associated algorithms and compare them with the literature to evaluate, improve or propose new approaches for 3D gait reconstruction.

The main objectives of this project will be:

1. Bio-mechanical Gait Parameter Extraction: To develop mathematical models to extract the lower body motions, and extract gait parameters such as knee and hip angles, step length, cadence, and gait cycle.
2. Machine Learning Application for Gait Classification: To will utilise the extracted gait features to train and validate machine learning and deep learning models for classifying pathological gaits.

Supervisory Team: The successful applicant will be working at LSBU with Prof Sandra Dudley (peoplefinder.lsbu.ac.uk/researcher/804zz/professor-sandra-dudley-mcevoy).The successful applicant will join The SABER  Research Centre, working alongside a range of new and experienced PhD students in a vibrant, collaborative, multidisciplinary and international environment.

Requirements: Applicants must be of outstanding academic merit and should have (or be expected to gain) either a first class or an upper second-class Honours degree (or the international equivalent), or an MSc/MRes with distinction. Enthusiastic and self-motivated candidates from all countries with a background in Biomedical Engineering, Electrical Engineering, Biology, Physics, or a related discipline are encouraged to apply. A good knowledge of Matlab, python and relevant languages and an interest/experience in RF/microwave hardware use and development is advantageous. Enquiries and applications should be directed to Prof. Sandra Dudley (dudleyms@lsbu.ac.uk). Due to funding constraints, this PhD must start by January 2024, no extensions permitted.