SGER: Robust Multiobjective Active Vision Systems

点击翻译按钮获取中文摘要

AbstractIIS-9911161Camps, OctaviaPennsylvania State University75,000 - 12 mos.SGER: Robust Multi-objective Active Vision SystemsThis is a SGER standard award. Recent hardware advances have rendered active vision a viable option for a very diverse spectrum of applications ranging from MEMS manufacture to vision assisted surgery. However, in order for active vision to gain widespread application, the problem of fragility (i.e. lack of robustness) of systems designed using classical approachesmust be addressed. This project aims to explore the feasibility of solving this difficulty through a multidisciplinary approach, which brings to bear on the problem new multi-objective robust control and robust vision tools that are just starting to emerge in the respective communities. Specifically, The PIs will use robust feature modeling and identification techniques to recast the problem into a multi-objective control form that can be exactly solved using convex optimization techniques. If successful, this effort will result in an expanded robust active vision framework, capable of addressing multiple performance specifications and uncertainty. The results obtained from this research are expected to show that a tight integration of the control and vision aspects of the problem in a unified framework leads to substantial improvements both in performance and flexibility.