Control of multi-degree-of-freedom catheters in unknown environments - exploring the potential of deep learning and augmented reality
The interaction between a flexible catheter and a deformable lumen is extremely complex as it is affected by slack, friction and a time-varying contact state. ESR11 focuses first on devising controllers for the distal robot part. The problem of deriving a controller that behave robust independent of contact state and actuator limitations (on power and range) is still unsolved. This will require adjusting the kinematic model to governing contact point/forces which can be estimated by fusion proprioception (ESR3) and RT reconstruction (ESR5, ESR6). Second, to establish a desirable global shape a multi-objective control problems is proposed in order to trade off: 1. the amplitude and direction of forces applied on tissue; 2. the need to avoid specific areas (e.g. arterial calcification); 3. the desire to include suggestions by users e.g. provided via teleoperation or shared control.