Radiation Aware Navigation and Planning for exploration of Hazardous Semi-Known Environments using a Mobile Robot Platform

One of the four main challenges facing the nuclear industry is to remove humans from harm in all aspects
of the nuclear life cycle. One of the main areas being explored to achieve this is through the utilisation of
robotic systems to conduct inspection and maintenance tasks in hazardous areas. Whilst huge advances
have been achieved in deploying robotic systems into active areas, the predictability and reliability of
these systems is an issue. System malfunction due to radiation exposure can cause future mission issues
if the system cannot be recovered, meaning restrictions on operation restrict system potential. One
method that could reduce these restrictions is to implement deterministic radiation-aware path planning
to track and avoid radiation as much as practically possible during operation. Current methods use
a standard occupancy grid consisting of a grid of cells, each a few centimetres on each side. This is
manageable computationally for smaller areas, but the number of cells requiring computation becomes
increasingly expensive and time-consuming if larger areas are considered. This work has developed a
method of abstracting this map into a sparse node network, with a total number of nodes required for
full map coverage several orders of magnitude lower than the cells of an occupancy map. Abstraction
utilising both free space and radiation levels in the map allows for a network of nodes that maintain the
environment's key features, allowing for a map that can be used for path planning in a significantly more
computationally friendly manner. Example of the world can be seen in figure 1a and the performance of
the algorithm with regards to radiation aware abstraction and path planning applied compared to only
a* path planning figure 1b. Additionally, this project aims to build on previous work by A. West et al. to
build a predictive model into the abstraction process to allow these techniques to be applied to unknown
environments. This is achieved through the prediction of radiation using inversely weighted distributions
to allow for the estimation of radiation at a given point based on a series of known measurements taken
during exploration. Reduction in the size of this network from abstraction allows for recompilation and
reactance to a changing environment. Finally, a method for simulating radiation sources in real life
without using actual radioactive sources is being explored. Using ultra-wide bandwidth locator beacons
allows real-world system testing in a lab-safe manner to validate the effectiveness of the above work using
as close to real-world conditions as possible.
(a) Example gazebo simulation (b) Distance travelled against radiation absorbed

In GREEN we envisage there are potentially Impacts in several domains: the nuclear Sector; the wider Clean Growth Agenda; Government Policy & Strategy; and the Wider Public.

The two major outputs from Green will be Human Capital and Knowledge:

Human Capital: The GREEN CDT will deliver a pipeline of approximately 90 highly skilled entrants to the nuclear sector, with a broad understanding of wider sector challenges (formed through the training element of the programme) and deep subject matter expertise (developed through their research project). As evidenced by our letters of support, our CDT graduates are in high demand by the sector. Indeed, our technical and skills development programme has been co-created with key sector employers, to ensure that it delivers graduates who will meet their future requirements, with the creativity, ambition, and relational skills to think critically & independently and grow as subject matter experts. Our graduates are therefore a primary conduit to delivering impact via outcomes of research projects (generally co-created and co-produced with end users); as intelligent and effective agents of change, through employment in the sector; and strong professional networks.

Knowledge: The research outcomes from GREEN will be disseminated by students as open access peer reviewed publications in appropriate quality titles (with a target of 2 per student, 180 in total) and at respected conferences. Data & codes will be managed & archived for open access in accordance with institutional policies, consistent with UKRI guidelines. We will collaborate with our counterpart CDTs in fission and fusion to deliver a national student conference as a focus for dissemination of research, professional networking, and development of wider peer networks.

There are three major areas where GREEN will provide impact: the nuclear sector; clean growth; Policy and Strategy and Outreach.

the nuclear sector: One of our most significant impacts will be to create the next generation of nuclear research leaders. We will achieve this by carefully matching student experience with user needs.

clean growth - The proposed GREEN CDT, as a provider of highly skilled entrants to the profession, is therefore a critical enabler in supporting delivery of both the Clean Growth agenda, Nuclear Industry Strategy, and Nuclear Sector Deal, as evidenced by the employment rate of our graduates (85% into the sector industry) and the attached letters of support.

Policy and Strategy: The GREEN leadership and supervisory team provide input and expert advice across all UK Governments, and also to the key actors in the nuclear industry (see Track Records, Sections 3.3 & 5.1, CfS). Thus, we are well positioned to inculcate an understanding of the rapidly changing nuclear strategy and policy landscape which will shape their future careers.

Outreach to the wider public: Building on our track record of high quality, and acclaimed activities, delivered in NGN, GREEN will deliver an active programme of public engagement which we will coordinate with activities of other nuclear CDTs. Our training programme provides skills based training in public and media communication, enabling our students to act as effective and authoritative communicators and ambassadors. Examples of such activities delivered during NGN include: The Big Bang Fair, Birmingham 2014 - 2017; British Science Week, 2013 - 2017; ScienceX, Manchester; 2016 - 2018; and The Infinity Festival, Cumbria, 2017.