Wire-Actuated Parallel Robots for Cochlear Implantation with In-vivo Sensory Feedback
Pile, Jason Anthony
Robot-assisted cochlear implant (CI) surgery is a new research area that emerged in the last decade. The goal of robotic assistance is to improve patient hearing outcomes through improved surgical access and the minimization of intracochlear trauma during implantation. This thesis presents several research efforts converging on a system for robotic atraumatic CI insertion. The work begins with the characterization of fundamental aspects of CI implantation. This leads to the synthesis of a robot design for the implantation task. Lastly, intelligent control through in-vivo sensory feedback is investigated for improvement in CI insertion and final placement. The motivation behind this research stems from fundamental knowledge gaps in both characterization of the CI surgical domain and in robot design and control. Current solutions for robot-assisted CI surgery do not exhibit adaptability to changes from nominal CI insertion plans. This lead to exploration of a new domain of in-vivo sensory guided robotic insertion of CI electrode arrays. The contributions of this work include a system architecture derived from the clinical specifications of CI surgery while simultaneously exploring theoretical gaps in the areas of mechanism design and static balancing of serial and parallel mechanisms. Methodical derivation of specifications for surgical access during CI implantation are presented and include available workspace, kinematic behavior of under-actuated implants, and baseline expectations of insertion forces. From this, the synthesis of a robotic wire-driven insertion platform for CI is presented. Lastly, both force and intra-cochlear impedance data collected by the proposed robot are used to add intelligent correction to the implantation procedure. These corrections include physical misalignment of the robotic system to the patient anatomy and incorrect models of the non-visible intra-cochlear geometry. Novel algorithms utilizing in-vivo sensory feedback for robot-assisted CI insertion guidance and fault detection are proposed and experimentally demonstrated using several robotic platforms. The implications of this research extend to providing new methods of CI insertion and also design of compact parallel robots with remote actuation.