ABSTRACT
From an engineering point of view it is clear that sensor placement is an integral part of control design, particularly in control of DPSs, e.g., flexible structures, air-pollution processes, oil and gas production from deposits, etc. Its choice is fundamental in the sense that it determines the accuracy of the system characteristics which are identified from an identification experiment. On the other hand, an engineering judgement and trial-and-error analysis are quite often used to determine spatial arrangements of measurement transducers, in spite of the fact that the problem has been attacked from various angles by many authors and a number of relevant results have already been reported in the literature. What is more, although it is commonly known that this area of research is difficult, since the nonlinearity inherent in the sensor-location problem precludes simple solution techniques, some systematic attempts at obtaining optimal sensor positions are still made and the progression is towards more general models, more realistic criteria and better understanding of the nature of the optimal locations. Logically, the number of applications should proliferate, yet this is not the case. It seems that two main reasons explain why strong formal methods are not accepted in engineering practice. First, with the use of the existing approaches, only relatively simple engineering problems can be solved without resorting to numerical methods. Second, the complexity of most sensor location algorithms does not encourage engineers to apply them in practice.
