Autonomous Data-Driven Manipulation of an Unknown Deformable Tissue Within Constrained Environments: A Pilot Study

To ensure safety and precision of autonomous surgical tasks performed on deformable tissues (DTs), in this paper, we propose a model-independent constrained optimization framework that is able to simultaneously learn deformation behavior of an unknown DT while autonomously manipulating it within a constrained and confined environment. To thoroughly evaluate the performance of the proposed framework, we used the da Vinci Research Kit and performed various experiments on an unknown DT phantom. We particularly compared the performance of our algorithm in 10 different configurations with and without the presence of imposed virtual workspace constraints. Results demonstrated the successful performance of the proposed framework in online deformation learning and manipulation of an unknown DT and revealed the effects of imposing constraints on the proposed model- independent framework.