The design and development of a mobile colonoscopy robot
[Thesis]
Norton, Joseph Christopher
Neville, Anne ; Culmer, Pete ; Boyle, Jordan ; Alazmani, Ali
University of Leeds
2017
Thesis (Ph.D.)
2017
The conventional colonoscopy is a common procedure used to access the colon. Despite it being considered the Gold Standard procedure for colorectal cancer diagnosis and treatment, it has a number of major drawbacks, including high patient discomfort, infrequent but serious complications and high skill required to perform the procedure. There are a number of potential alternatives to the conventional colonoscopy, from augmenting the colonoscope to using Computed Tomography Colonography (CTC) - a completely non-invasive method. However, a truly effective, all-round alternative has yet to be found. This thesis explores the design and development of a novel solution: a fully mobile colonoscopy robot called "RollerBall". Unlike current passive diagnostic capsules, such as PillCam, this device uses wheels at the end of adjustable arms to provide locomotion through the colon, while providing a stable platform for the use of diagnostic and therapeutic tools. The work begins by reviewing relevant literature to better understand the problem and potential solutions. RollerBall is then introduced and its design described in detail. A robust prototype was then successfully fabricated using a 3D printing technique and its performance assessed in a series of benchtop experiments. These showed that the mechanisms functioned as intended and encouraged the further development of the concept. Next, the fundamental requirement of gaining traction on the colon was shown to be possible using hexagonal shaped, macro-scale tread patterns. A friction coefficient ranging between 0.29 and 0.55 was achieved with little trauma to the tissue substrate. The electronics hardware and control were then developed and evaluated in a series of tests in silicone tubes. An open-loop strategy was first used to establish the control algorithm to map the user inputs to motor outputs (wheel speeds). These tests showed the efficacy of the locomotion technique and the control algorithm used, but they highlighted the need for autonomy. To address this, feedback was included to automate the adjusting of the arm angle and amount of force applied by the device; a forward facing camera was also used to automate the orientation control by tracking a user-defined target. Force and orientation control were then combined to show that semi-autonomous control was possible and as a result, it was concluded that clinical use may be feasible in future developments.
Norton, Joseph Christopher
Neville, Anne ; Culmer, Pete ; Boyle, Jordan ; Alazmani, Ali