Accepting PhD Students

PhD projects

EY1: Biologically-Inspired Multi-Objective Design Optimisation for Space Mechatronic Systems

In future space missions, autonomous, intelligent and massively distributed mechatronic systems will play important roles. The space application domain presents unique challenges to the design of mechatronic systems. For example, due to the extremely limited and expensive resources available onboard, design optimisations are particularly needed for both power savings and performance improvement in satellite-based sensing and imaging. Thus, the development of efficient multi-objective design optimisation algorithms capable of optimising the space-based mechatronic systems is essential under the stringent requirements for power consumption, cost, mass, reliability, and performance improvement, etc.
The aim of this research is to develop novel and efficient bio-inspired optimisation approaches for multi-objective design-space exploration of space mechatronic systems. Its main research objectives are summarised as below:
1. Develop new bio-inspired algorithms for design exploration of space mechatronic systems under multiple design and environmental constraints.
2. Investigate algorithm performance, convergence, and design efficiency trade-off.
3. Investigate the computing requirement of bio-inspired algorithms for real-time response to application requirements under different environmental constraints.
4. Investigate the designs produced by the developed bio-inspired algorithms in terms of multiple objectives, such as cost, mass, reliability, and performance, etc.

EY 2: Brain-inspired Intelligent Control of Multiple Autonomous Systems for Space Applications.
In space application, multiple autonomous systems (MASs) can be more effective than a single autonomous system, for example, in information gathering and exploration tasks with multiple planetary robots. The potential for MASs cooperating together to accomplish challenging tasks has drawn together researchers from several fields, including robotics, control systems, and computer science. Biologically-inspired and intelligent control systems for MASs, including for Mars’ rovers and DARPA Challenges, have received a lot of research attention. For example, a fault-tolerant, Biologically Inspired System for Map-based Autonomous Rover Control has been developed in NASA’s Jet Propulsion Laboratory for long duration missions with multiple autonomous vehicles.
The general scientific objectives of this research are to address two fundamental research challenges related to the development of novel intelligent coordinated control of multi-agent systems, particularly in the context of MASs for space applications. The first research challenge is related to real-time information processing and utilisation, i.e., how to quickly and efficiently extract and analyse information acquired by the MASs. The second research challenge is concerned with designing adaptive autonomous controllers by exploiting the extracted and analysed information to cooperatively control the MASs, and also improve the MAS’s capabilities, such as surveillance, target acquisition and tracking, etc.
To realise these general scientific objectives, novel brain-inspired approaches are particularly appealing for extracting information, processing the extracted information into the design, online tuning, and adaptive switching of autonomous multi-agent controllers under complex and dynamic environments.

EY 3: Towards an Integrated Cognitive Control System for Autonomous Vehicles
The field of autonomous vehicle is a rapidly growing one which promises improved performance, fuel economy, emission levels, comfort and safety. The UK government is determined to address the challenges of tackling climate change, maintaining energy security, and solving transportation in a way that minimises costs and maximises benefits to the economy. Among all sources of CO2 emissions in the UK, the energy supply accounts for about 40%, followed by the transport for over 25%, and emissions from cars and vans account for 70% in domestic transport sector. Intelligent and efficient control is one of key issues in developing fully autonomous vehicles.
Given the similarity between the problem' domains of autonomous vehicle’s control and action selection in animals, this research aims to leverage new results from psychology and neurobiology and apply them to the control of autonomous vehicles. Towards this end, an integrated cognitive control system for autonomous vehicles is targeted in this research. We aim to harness general strategies for control based on high level analysis of human behavioural control.
The primary objective of this research is to uplift research collaborations from the current separate, point-to-point collaborations to a broader and deeper context with a more systematic, more coherent, and more coordinated synergised approach for developing an integrated cognitive control system for autonomous vehicles towards more reliable, more flexible and efficient, and more environmental friendly control solutions.

EY 4：Novel Intelligent and Optimal Decision-Making Paradigm for Autonomous Manufacturing
Today’s manufacturing has become more competitive as manufacturers need innovative and extremely agile processes along with increasing manufacturing automation and informatics complexity. The main aim of this research is to develop novel intelligent and optimal decision-making paradigm for autonomous manufacturing in an Industry 4 environment.
The focus of this research will be on the development of innovative smart decision-making strategies to coordinate multiple agents and make collaborative and optimal decision and take group actions by concurrently dealing with multiple objectives under extreme environmental constrains arising from both internal (such as manufacturing process disturbances, e.g., one industrial robot breaks down) and external (e.g., partial and inconsistent information). In the smart factory, all the manufacturing resources are modelled as intelligent agents/entities. Each agent has the ability to percept, reason, make decision, and take actions without (or with limited) human interferences.
The key research challenge lies in the intelligent and optimal decision-making mechanism for the agents in the smart manufacturing system to independently make decisions and plan their own tasks based on their own reasons about their environment, state/situation and the likely actions taken by other agents.

EY 5: Advanced Artificial Intelligence in Automatic Human-Machine Knowledge Transfer
This research is concerned with the way that how to automatically transfer human expert/operator’s knowledge obtained in existing processes and experiences into intelligent agents (e.g., industrial robots) to advance the capabilities of intelligent agents in an autonomous manufacturing system. Toward this end, an advanced intelligent system consisting of intelligent knowledge-based expert system and artificial neural network (ANN) models will be explored in this research. The use of ANN models will make the system intelligent by learning patterns from existing manufacturing data and process knowledge and use them for predicting the behaviour of manufacturing, which would result in reducing the lead time and cost considerably.