System identification of complex and structured systems

摘要: A key issue in system identification is how to cope with high complexity. In this contribution we stress the importance of taking application into account order issue. We define concept “cost complexity” which a measure minimum required experimental effort (e.g. used input energy) as function complexity, noise properties, and amount, desired quality, information be extracted from data. This gives user handle on trade-offs that must considered when performing fixed “budget”. Our analysis based observation objective guarantee estimated model ends up within pre-specified “level set” objective. geometric notion leads number useful insights: Experiments should reveal properties important for but may also conceal irrelevant properties. The latter, dual, can explored simplify structure selection error assessment issues. discuss practical issues related computation implementation optimal experiment designs. Finally, illustrate some fundamental limitations arise structured systems. topic has bearings networked decentralized