FLEXIBLE MANIPULATORS: MODELING, ANALYSIS AND OPTIMUM DESIGNS

Abstract

Lightweight robotic manipulators play important roles in many applications, such as construction automation, environmental applications, and space engineering. Due to the complexity of the link deformation, the research effort on accurate modeling and high performance control has increased dramatically in recent years. Very few studies with quantitative results have been reported in the literature on the effect of size, shape, mass distribution, tip load, and other factors on the dynamics and operational performance of flexible manipulators. However, such analyses are critical to the effectiveness of any model for optimization and control purposes. The objective of this dissertation is to provide a unified approach that simultaneously considers all factors in mechanical, electrical, sensing, and control components to address modeling, analysis, optimization, and control of flexible manipulators. A systematic study of various models and comparison their pros and cons with respect to specific design and control problems are developed. Those critical factors mentioned above are addressed with systematic but specific numerical investigations based on the current available dynamic models. Two conventional dual optimal design problems have been well studied in our previous studies. To make those designs useful, complicated constraints that are encountered in reality must be included. In most of those cases, analytical procedures developed before would no longer applicable. New methods for optimal design problems with meaningful constraints are looked into. The complexity of a manipulator system is due to the interrelation and interdependency of its subsystems, for example, its kinematic system control system, driver system, and measuring or sensing system. In traditional design, a manipulator's link structure is designed first, followed by its driver system, then a measuring system, and finally its control system. This leads to a sequential design process and a locally optimal solution, and therefore the potential of a flexible manipulator is rarely fully realized. To overcome such problems, a concurrent design procedure that integrates all subsystems must be under taken, that is, a mechatronic approach must be considered in the design of flexible manipulators.