Innovative Seed Grant - Enabling Interactive Control of In vivo Robots

Summer 2016 - Fall 2017.  The goal of this research is to enable closed-loop control of mobile in vivo robots in direct support of physicians so that physicians can focus on diagnosis/treatment and not primarily on guiding devices within the body. Realization of this co-robot system will enable control through tortuous paths and around corners (colonoscopy), tracking robot pose within the body (gastrointestinal exploration), and positional control to maintain an object of interest (tumor) within the field of view for detailed viewing, biopsy, or removal. In pursuit of this long-term goal, three research objectives need to be addressed. The first is to create a closed-loop control system for maintaining robot position and velocity. The expectation is that model-based control approaches will prove unfeasible due to highly irregular, and changing environmental factors. Thus, the research will focus on adaptive control approaches, with the second outcome being a set of new design rules formulated for gain scheduling at different in vivo junctures. The second long-term objective is to address the in vivo localization challenge in a deformable environment void of conventional global cues (e.g., GPS). In an in vivo scene (e.g., colon), surface edges are scarce and highly deformable; the environment also exhibits similar color and texture patterns. When combined, these difficulties create high computational complexities not currently addressed in conventional field robotics. The third long-term objective is to create a novel multi-modal user interface to minimize physician burden so that the physician can focus on the imaging. The proposed project’s short-term research aim is to demonstrate the feasibility of effectively modeling the RCE and in vivo environment dynamics. Accomplishing this, would enable development of control techniques to allow the RCE to maneuver with respect to important features in the colon (e.g., navigating tortuous curves, maintaining position against physiological processes such as breathing and peristalsis) so as to allow the user to easily select, interact and respond to important features in the environment (e.g., maintaining a polyp of interest in the camera frame), while preventing patient discomfort and minimizing tissue damage.