Design and Development of High-resolution Opto-mechatronic Systems.

点击翻译按钮获取中文摘要

The standard definition of mechatronics is: "an approach aiming at the synergistic integration of mechanics, electronics, control theory, and computer science within product design and manufacturing, in order to improve and/or optimize its functionality". The term Opto-mechatronics refers to a mechatronic system whose sensory information is mainly obtained via optical imaging. Incorporating vision sensors into the design of mechatronic systems will bring them to a quantum leap in terms of reliability, resolution, accuracy, and redundancy thanks to the recent developments in digital imaging and the offered computing power of the current processors. No sensory feedback is known to the man kind richer than vision. In fact, machine vision may well surpass most of the biological imaging sensors found in the animal kingdom. However, the complexity of the algorithms used in image processing, feature detection, pattern recognition/classification, and mapping prevents them from being utilized in some real-world applications. Therefore, there is always a trade-off between the resolution and speed. This proposal is a timely response to the world-wide need in formulating intelligent algorithms to replace the time-consuming brute-force techniques used by far. This proposal targets on the development of a cluster of algorithms for metric measurements from sequential images. It also investigates optimal design of opto-mechatronic systems, by which these sequential images are taken, to achieve the utmost effect. The existing systems can provide only a fixed spatial resolution. Development of adaptive information filters and control techniques based on either learning methods or functional analysis techniques will be the main objective of this proposal. The proposed algorithms can keep the utility function in opto-mechatronic systems at its maximum level regardless of the differentials such as lighting condition, and the topological structure of the environment by changing system's configuration and parameters dynamically.