Notice: Undefined index: reset in /var/www/html/wp-content/mu-plugins/cets_blog_defaults.php on line 733

Deprecated: define() was called with an argument that is deprecated since version 3.0.0! The constant VHOST is deprecated. Use the boolean constant SUBDOMAIN_INSTALL in wp-config.php to enable a subdomain configuration. Use is_subdomain_install() to check whether a subdomain configuration is enabled. in /var/www/html/wp-includes/functions.php on line 5497
Abstracts - Mathematics for Environment and Sustainability

Home » Abstracts

Abstracts

Here you will find abstracts of some excellent student projects from the Mathematics for Environment and Sustainability module and other modules on the BSci/MSci Mathematical Sciences programmes.

AI and Healthcare: A New Dawn or Apartheid Revisited

Alice Parfett, Flexible Combined Honours

The Bubonic Plague outbreak that wormed its way through San Francisco’s Chinatown in 1900 tells a story of prejudice guiding health policy, resulting in enormous suffering for much of its Chinese population. This article seeks to discuss the potential for hidden “prejudice” should Artificial Intelligence (AI) gain a dominant foothold in healthcare systems. By using a toy model, this piece explores potential future outcomes, should AI continue to develop without bound. Where potential dangers may lurk will be discussed, so that the full benefits AI has to offer can be reaped whilst avoiding the pitfalls. The model is produced using the computer programming language MATLAB and offers visual representations of potential outcomes. Interwoven with these potential outcomes are numerous historical models for problems caused by prejudice and recent issues in AI systems, from police prediction and facial recognition software to recruitment tools. Therefore, this research’s novel angle, of using historical precedents to model and discuss potential futures, offers a unique contribution.

A video of the Gyroboy robot balancing itself and being controlled by an Xbox One controller via Bluetooth. Video by Gareth Willetts.

LQG controller for the LEGO MINDSTORMS EV3 Gyroboy Segway Robot as a Teaching Tool

Gareth Willetts, Jakub Kryczka,  Mathematics

This project details the development of an affordable Segway robot demonstrator for undergraduate and Masters level systems and control courses, such as the Mathematical Sciences degree programme at the University of Exeter (Penryn Campus), based on the LEGO MINDSTORMS EV3 robotics platform. The demonstrator is designed using a standard LEGO MINDSTORMS model from the Education EV3 core set – the Gyroboy. This is interfaced with using the Simulink Support Package for LEGO MINDSTORMS EV3 Hardware, with its model designed using MATLAB and Simulink. Reference inputs can be provided from either a keyboard or an Xbox One gamepad. The purpose of the demonstrator is to provide a physical and interactive device for explaining mathematical concepts that commonly feature on systems and control courses, notably: model-based control, linearization, the Linear-Quadratic Regulator (LQR), pole placement, noise amplification due to differentiation, reference following, Kalman filtering, and the principle of separation of estimation and control. To the best of our knowledge, this is the first example of the successful implementation of an observer-based reference-following feedback controller for a Segway robot built entirely using LEGO MINDSTORMS EV3 components, with previous designs being either not observer-based or based on the now outdated LEGO MINDSTORMS NXT platform. This demonstrator has been used in teaching 3rd and 4th year modules for the Mathematical Sciences degree programme, and a paper on the Segway robot was accepted for the track titled “Benchmarks and Case Studies for Control Education” at the International Federation of Automatic Control World Congress, held online in July 2020.

Traffic Accidents Analytics in UK Urban Areas using k-means Clustering for Geospatial Mapping

Chris Sinclair, Mathematics

The goal of this project is to use the unsupervised machine learning method in road accident analytics, especially using k-means clustering to identify patterns and understand the relationships between variables recorded by the UK police department. These include features like number of casualties, number of vehicles, age of vehicle and age bracket of the driver. We aim to describe clusters of accidents based on similarity measures in the features and identify what separates each one.

Skip to toolbar