This project will be based at: The University of Liverpool
When an autonomous robot needs to move within a partially known environment with high radiological or similar hazards, the control unit needs to perform autonomous navigation, graceful degradation and graceful end of life. This scenario is common in nuclear facilities for which there is a “floor plan” available but exact location of sources in unknown, Solving this challenge involves five main tasks: i) planning the route to reach a desired target position, ii) navigate along the desired path, iii) continuously monitor the environment to check if there are any obstructions/dangers to be avoided, iv) monitoring the effects of the radiation of the internal state of the robot at all times to avoid unplanned failures, v) arrival on a dedicated area (i.e. robot graveyard) before the end of life to avoid being an obstacle itself. Several approaches have been considered in the literature to solve these tasks, but their integration and effectiveness in partially known environment still poses challenges. Moreover, the internal state of the robot itself (e.g. battery level, potential damage due to sources, etc.) rarely features in existing algorithms, causing a major bottleneck in achieving graceful degradation and end of life.
In this project, the student will develop and test algorithms to allow mobile ground robots to navigate partially known environments where sources are placed at unknown locations. It is assumed that such sources need to either be monitored by the robot (e.g. to accurately measure intensity) or avoided to prevent damage to materials and/or on-board electronics. To enable this, the control unit has to not only monitor the surrounding environment, but also estimate and predict its internal state based on distance travelled and amount or radioactivity it has been/will be exposed to along the path to achieve graceful degradation and perform graceful end of life. The path will need to be optimally updated to ensure efficient navigation within the environment while ensuring safety of the autonomous robot. The algorithm will be developed in a simulated environment (Gazebo/CoppeliaSim) first, then tested on ground rovers in a lab environment and, ideally, assessed on the facility of the industrial partner toward the end of the project.
The successful candidate should have, or expect to have, at least a 2:1 degree or equivalent in, one of the core subjects, i.e. Mechanical Engineering, Electronics, Computer Science, or closely related subject. The candidate should be highly motivated, curious, have competent English communication skills, computer skills and be able to work both as part of a team and independently. Prior experience in mechatronics or AI would be an advantage but is not a prerequisite. Applicants, whose native language is not English, or those that do not have a degree from an English speaking University, will need to meet the Postgraduate English Language entry requirements.
The student will be part of the EPSRC-CDT in Nuclear Energy – GREEN and will be based in the Department of Mechanical, Materials and Aerospace, Department of Chemistry and the @LERT Robotics lab at the University of Liverpool. The student will also interact with the RACE robotics team at the centre of the fusion in Culham which is a unique opportunity to learn broader aspects of robotics, fusion and remote maintenance.
The GREEN Centre for Doctoral Training (GREEN CDT) is a a consortium of five universities: The University of Manchester, Lancaster University, The University of Leeds, The University of Liverpool and The University of Sheffield, which aims to train the next generation of expert nuclear scientists and engineers.
Students within the GREEN CDT are invited to undertake a four-year PhD programme. Students will attend taught courses (Year 1) in various subject of nuclear technology followed by subject specific training (Year 1) leading to research activities (Year 2-Year 4).
Based on your current searches we recommend the following search filters.
Based on your current search criteria we thought you might be interested in these.