We currently have 2 PhD studentships available, with the descriptions as below. This is a fantastic opportunity to develop skills in field robotics. The two students will work closely together and with our industrial partner, FCDO Services, who are responsible for securing sensitive UK government facilities. Please see the links to the official adverts and contact me if you are interested.
Autonomous robot teams or ‘swarms’ have great potential to enhance monitoring of different environments, by providing a mobile sensor network that can cover a complex, dynamic area. These PhD projects will help to develop important principles for the deployment of robot swarms in the field, with a focus on security applications. These principles will be validated through meaningful experimental data collected in robot field trials.
Robot Swarms for Environmental Monitoring: Investigating Behavioural Plasticity
Deploying robots into real-world environments, even enclosed environments such as office buildings, is a difficult challenge. This is because the world is inherently unpredictable and uncertain, and hence robot designs that work well in carefully controlled laboratory environments are liable to encounter difficulties in real-world operation. A key enabler could be to have a suitable level of plasticity (flexibility) in behaviour, such that robots can adapt online to their environment.
This studentship will investigate the value of variability in fundamental robot behaviour, across what have been called ‘personality’ axes, such as slow—fast, shy—bold, or exploratory—neophobic. Such variation could result from an implementation of reinforcement learning, for example, in relation to task performance (e.g. anomaly detection success). This could be introduced alongside other aspects of plasticity, for instance in robot movement patterns.
The investigations will be undertaken in realistic simulations of ROS-based rover robots in the Gazebo simulator, and at a University of Bristol field site. Experimental work (simulated and real) will be complemented by development of a conceptual framework to measure and predict the positive (or negative) impact of behavioural plasticity.
Robot Swarms for Environmental Monitoring: Investigating Functional Heterogeneity
The classic starting point for robot swarm research is to assume agents are identical in their characteristics, for example in the ‘behavioural rules’ they follow in reacting to the environment and each other. However, in biology it is increasing recognised that diversity (heterogeneity) among group members can be functional (useful) for the whole group. Such heterogeneity includes variation in behaviour, robot form (e.g. rover, quadruped) and sensory capabilities. Such variation can confer important benefits, such as enhanced awareness of the environment, better response to risk, or more accurate collective decision-making.
This studentship will investigate heterogeneity in robot characteristics – starting with behaviour – to assess its effect on performance in relation to detecting anomalies that may be a security concern. The student will develop realistic simulations of ROS-based rover robots in the Gazebo simulator and carry out real robot trials at a University of Bristol field site. Experimental work (simulated and real) will be complemented by development of a conceptual framework to measure and predict the positive (or negative) impact of swarm heterogeneity.
I am happy to support your application for a postdoctoral fellowship. I have previously had success myself with the EPSRC Doctoral Prize and the UK Intelligence Community Fellowship. The Doctoral Prize is awarded through the university, while EPSRC also has a Postdoctoral Fellowship that is externally assessed. Please see these and other schemes below and get in touch if you would like to discuss the possibility of doing postdoctoral work in my group.