Ph.D. Thesis Defense by Anh  Dang
Monday, November 11, 2002

(Dr. Ebert-Uphoff, advisor)

"Theoretical and Experimental Development of an Active Acceleration Compensation Platform Manipulator for Transport of Delicate Objects"

Abstract

In many transport systems, disturbances and/or uneven terrain makes it difficult or impossible to handle and transport delicate and fragile cargo. Therefore, compensation techniques for disturbance rejection are required to circumvent these problems. While passive compensation techniques have been extensively studied for this purpose, research on the application of active compensation is still relatively novel. Active compensation techniques can broaden the scope of compensation applications due to its abilities to handle a larger variety of situations for which passive compensation may not be adequate.

The proposed approach to enhance the capabilities of transport vehicles is to mount a robotic device on top of the vehicle whose motion will seek to compensate for any forces or moments that act on the objects as a result of the vehicle's motion, including disturbances caused by uneven terrain. The objective of this research is to develop a theoretical and experimental framework to explore the feasibility of active acceleration compensation for actual, real-time applications. The theoretical framework seeks to define a performance measure and design a realizable motion planning scheme to achieve acceleration compensation in real-time. The proposed performance measure is to minimize the friction force and reaction moment, while maximizing the normal force acting on an object that is on the parallel platform manipulator (PPM) for any given motion of the mobile robot. Based on the performance measure, the general framework of motion planning is developed to yield different actuation schemes to achieve compensation. The fundamental principle behind the motion planning scheme presented in this study, is based on the emulation of a virtual pendulum to actively compensation for acceleration input disturbances. In addition to the virtual pendulum, a well-developed washout-filter motion planning algorithm is also used. The combination of the pendulum and washout-filter is a novel motion planning approach that provide a robust and general solution to the active compensation problem.

The practical effectiveness of the motion planning algorithm is proven with an experimental test-bed. The test-bed is developed as a proof of concept for active acceleration compensation. The experimental results of the test-bed are compared with the simulation results to seek an agreement between the theory and practice.