Summary: | Thesis: S.B., Massachusetts Institute of Technology, Department of Mechanical Engineering, 2015. === Cataloged from PDF version of thesis. === Includes bibliographical references (page 37). === High precision sensing is a widely applicable technology that ranges in use case from manufacturing, mobile devices, and robotics among more. Specifically inspired by the need for a high precision orientation sensing for an underwater inspection robot with high maneuverability, we explore the design of an underwater-capable sensor that can provide fast response and easy integration into the overall architecture of the robot. Ultimately, the sensor design opted for cost-minimization (<$50) in tandem with the high precision capabilities to create a design that would be further accessible. The sensor array consists of three Time-of-Flight sensors, each having +/- 1 mm precision and accuracy. The sensors sit in a triangular formation to allow for 3-point range finding and thereby constrain the sensing input to locate the robot in 3D space. In addition, custom fabrication of Printed Circuit Boards and implementation of an Extended Kalman Filter allow for integration into nearly and robot, while this was specifically designed for EVIE (Ellipsoidal Vehicle for Inspection and Exploration), a current project within d'Arbelloff Lab. For the sake of yielding useful data for position and orientation, the sensor array outputs a vector of Distance perpendicular to the surface, Yaw difference, and Pitch Difference. This is then fed back into the trajectory planning algorithms onboard EVIE. With full EKF implementation, and tuned noise parameters, the system exhibited precision and response beyond the typical sensor range.Typical accuracy of the Perpendicular Distance measure output was found to be +/-.52 mm while the Pitch and Yaw respectively held accuracy of +/- 20° and +/- 5°. === by Lawrence "Zack" Bright. === S.B.
|