This MSc thesis project focuses on the optimisation and control of a robotic injection system for targeted treatment of tumours in soft tissue. A robotic arm equipped with a steerable concentric-needle end-effector is
used to deliver injectable therapeutic substances into a tumour, with the tumour geometry provided as input from medical imaging data, such as CT, with possible extension to MRI. The inner needle is steered through
controlled insertion and rotation inside the outer needle, allowing directional control while minimising damage to surrounding healthy tissue.
The main objective is to optimise injection planning such that tumour coverage is maximised, exposure of healthy tissue is minimised, and the number of injections is reduced. Optimisation approaches, such as
reinforcement learning or other techniques, will be investigated. In addition, trajectory planning and control methods will be developed to enable the robotic arm to accurately follow the planned injection paths and
execute the injections. These methods will be developed and evaluated primarily in simulation and subsequently validated experimentally on soft-tissue phantom models.
For the experimental implementation, the existing roto-translational stage of the injection end-effector, which is currently manual, will be motorised to enable automated execution of the planned injection strategies.