Validation and integration of a FEM model for probe-soft tissue interaction in robotic ultrasound applications

Kurzfassung

Robotic systems are increasingly being used to automate contact tasks. Reliable mechanical interaction with the environment is essential here, as the energy exchange that occurs during contact has a significant influence on the robot's behaviour. Contact force is a central parameter and should be taken into account in feedforward and closed loop control algorithms to ensure the successful completion of these tasks. The interaction with deformable objects, such as soft tissue, represents a special challenge: contact forces not only cause motions, but also complex, material-dependent deformations. With the help of a FEM model, the relation between contact force and tissue deformation can be realistically described and used in various control strategies. In this thesis, the model is applied in an impedance-controlled robotic system. In robot-assisted US imaging, impedance control ensures stable and secure contact between the probe and tissue – a central requirement for high and reproducible image quality. Depending on the type of tissue, the probe has to indent several millimetres to centimetres into the tissue to ensure stable contact force and good acoustic coupling. With manual guidance, the indentation depth and image quality depend strongly on the experience of the operator, resulting in varying contact forces, inconsistent image quality, potential tissue damage to the patient and an increased workload for the operator. Robotic systems, on the other hand, enable automated probe guidance. The aim of this thesis is to systematically analyse the interaction between US probe and soft tissue. For this purpose, a simulation model is developed in FEM environment that considers probes with circular and rectangular contact surfaces. The prescribed displacement model used enables controlled displacement of the probe, taking into account the material-specific deformation behaviour of the soft tissue. An experimental test setup is being realised for validation. The measured contact force–indentation depth curves are used for optimisation-based identification of the material parameters. The validated model is then integrated into the robotic system MIRO at DLR. The first use case serves to verify the FEM model, while the second use case utilizes it as a force feedforward control to achieve a desired contact force. In addition, the FEM model is being conceptually integrated into ongoing research work at DLR, where an US servo control system is being developed that uses US image information to adjust the robot position and probe orientation and ensure the desired coupling between probe and tissue. For this approach, the deformation model is integrated into a Kalman filter to enable local estimation of soft tissue properties, such as object stiffness.