On adaptive robot systems for manufacturing applications

Abstract

System adaptability is very important to current manufacturing practices due to frequent changes in the customer needs. Two basic concepts that can be employed to achieve system adaptability are flexible systems and modular systems. Flexible systems are fixed integral systems with some adjustable components. Adjustable components have limited ranges of parameter changes that can be made, thus restricting the adaptability of systems. Modular systems are composed of a set of pre-existing modules. Usually, the parameters of modules in modular systems are fixed, and thus increased system adaptability is realized only by increasing the number of modules. Increasing the number of modules could result in higher costs, poor positioning accuracy, and low system stiffness in the context of manufacturing applications. In this thesis, a new idea was formulated: a combination of the flexible system and modular system concepts. Systems developed based on this new idea are called adaptive systems. This thesis is focused on adaptive robot systems. An adaptive robot system is such that adaptive components or adjustable parameters are introduced upon the modular architecture of a robot system. This implies that there are two levels to achieve system adaptability: the level where a set of modules is appropriately assembled and the level where adjustable components or parameters are specified. Four main contributions were developed in this thesis study. First, a General Architecture of Modular Robots (GAMR) was developed. The starting point was to define the architecture of adaptive robot systems to have as many configuration variations as possible. A novel application of the Axiomatic Design Theory (ADT) was applied to GAMR development. It was found that GAMR was the one with the most coverage, and with a judicious definition of adjustable parameters. Second, a system called Automatic Kinematic and Dynamic Analysis (AKDA) was developed. This system was a foundation for synthesis of adaptive robot configurations. In comparison with the existing approach, the proposed approach has achieved systemization, generality, flexibility, and completeness. Third, this thesis research has developed a finding that in modular system design, simultaneous consideration of both kinematic and dynamic behaviors is a necessary step, owing to a strong coupling between design variables and system behaviors. Based on this finding, a method for simultaneous consideration of type synthesis, number synthesis, and dimension synthesis was developed. Fourth, an adaptive modular Parallel Kinematic Machine (PKM) was developed to demonstrate the benefits of adaptive robot systems in parallel kinematic machines, which have found many applications in machine tool industries. In this architecture, actuators and limbs were modularized, while the platforms were adjustable in such a way that both the joint positions and orientations on the platforms can be changed.