Publications

Many industries contain confined, clustered environments that require routine inspection. Nuclear facilities, in particular, have extensive pipe networks that pose additional radiological challenges, making human inspection unsuitable. This work formulates a whole-body controller for a non-holonomic mobile manipulator as a quadratic programme, utilising Vector Field Inequalities to constrain the system. It introduces a stochastic point-to-plane constraint to handle uncertainty in wall localisation and an online path planner combining coverage path planning with 3D reconstruction. The approach is demonstrated experimentally using a physical mobile manipulator in a confined environment.

Bettles, J. R., Wu, J., Adorno, B. V., Carrasco, J., & Yamashita, A. (2026). Planning and Control of Mobile Manipulators for Inspection of Clustered Environments. 7th UK Robot Manipulation Workshop.

This work introduces stochastic Vector Field Inequalities (SVFIs), extending the VFI framework to account for uncertainty in workspace geometric primitives such as obstacles. Chance constraints are incorporated into a constrained optimal task-space controller to provide probabilistic guarantees for the desired distance dynamics and collision avoidance. A deterministic surrogate is developed for implementation in constrained optimal control, with formulations for point-to-point and point-to-plane collision avoidance.

Bettles, J. R., Song, Y., Adorno, B. V., & Carrasco, J. (2026). Robust Local Collision Avoidance via Stochastic Vector Field Inequalities. Geometric Representations in Robotics Workshop, IEEE/RSJ IROS 2026, Paper 05.

Radioactive contamination monitoring is important for radiological protection, but automating surface monitoring raises challenges around probe coverage and collision avoidance. This work presents an accessibility framework that models radiation probes as convex hulls and assesses collisions against 3D point clouds. The approach extracts surface properties, approximates the probe and effective scan volume with geometric primitives, and categorises points as accessible or inaccessible. The resulting assessment can support collision-free path planning over objects and forms part of the foundation for a robotic survey assistant.

Bettles, J., West, A., Andrew, J., Darby, I., & Lennox, B. (2024). Accessibility Framework for Determining Collisions and Coverage for Radiation Scanning. Lecture Notes in Computer Science. Springer Nature Switzerland.

Robotic inspection of radioactive areas enables operators to be removed from hazardous environments; however, planning and operating in confined, cluttered environments remain challenging. These systems must autonomously reconstruct the unknown environment and cover its surfaces, whilst estimating and avoiding collisions with objects in the environment. This paper proposes a next-best-view planning approach that enables simultaneous exploration and exploitation by reformulating coverage path planning in terms of information gain. To handle obstacle avoidance under uncertainty, the vector-field-inequalities framework is extended to account for stochastic measurements of geometric primitives through chance constraints in a constrained optimal control law. Experiments on a mobile manipulator demonstrate autonomous inspection and coverage of a simplified pipe network, online estimation of geometric primitives, and collision avoidance while moving between viewpoints.

Bettles, J. R., Wu, J., Adorno, B. V., Carrasco, J., & Yamashita, A. (2026). Coverage First Next Best View for Inspection of Cluttered Pipe Networks Using Mobile Manipulators. arXiv. https://doi.org/10.48550/arXiv.2603.16471