Self Funded PhD – I

Project Title: Anomaly Detection in Sensor Generated Data for Secure Systems

Anomaly detection is important for detecting critical events such as intrusions, faults, and frauds, in many application domains (e.g., aircraft health management, surveillance systems, health information systems). For example, in aircraft health management, during an aircraft’s flight, multiple sensors flight parameters indicate system health. This data is collected as time series. A fault in the aircraft’s operation, such as failure of a component during flight, is manifested as anomalies in one or more of the sensor readings generated by the aircraft. Another application domain is surveillance in cluttered scenarios like museums or transport (airports, train or tube stations) where conventional monitoring systems have to analyse vast amounts of recorded data to extract evidence about criminal activities.

The PhD researcher will investigate 2D/3D data streams using GPGPUs for real-time responses and cloud computing for off-line training and machine learning purposes. In particular, the candidate will be instructed in areas related with anomaly detection techniques for the analysis of deformations and trajectories that define behaviours, as well as high processing performance of GPUs and parallel computing (e,g, CPU-GPU, GPU-GPU) to allow the parallel management of information from different sensors. The proposed system should develop robust machine learning algorithms for the real-time processing and analysis of data streams under constraints (e.g. noise from sensors, network fail). Strong background in mathematics and programming is required, as well as general grasp in the following areas: use of machine learning and parallel computing libraries and architectures (CUDA, OpenCV, PCL, etc.).

This project will enable the student to develop skills in Computational Intelligence, and High-Performance and Parallel Computing architectures with direct impact on Digital Economy and Security. The student will be encouraged to attend relevant conferences, participate in the University’s Doctoral Researcher Development Programme (DRDP) and have the opportunity to obtain a Postgraduate Certificate in teaching and learning.

Funding:

This PhD project is offered on a self-funding basis. It is open to applicants with funding or those applying to funding sources.  Details of tuition fees can be found at https://www.westminster.ac.uk/courses/research-degrees/fees-and-funding

Informal project enquiries and Director of Studies: Dr Anastassia Angelopoulou (agelopa@westminster.ac.uk)

Supervisory team: Dr Epaminondas Kapetanios (kapetae@westminster.ac.uk) and Prof. Gabor Terstyanszky (G.Z.Terstyanszky@westminster.ac.uk)