Optimal sensor placement in structural control systems

Due to the accuracy requirements of system performance and suppressing the cost of monitoring associated with the physical constraints, many researchers has focused on the design of optimal sensor systems along with the requirement to perform a specific observation extensively. It is obvious that the optimal sensor allocation is a required tool in enhancing performance qualities from structural control problems. Measures of optimality often include the quality of controllability, the observability and information matrix. The problem of optimal sensor placement for decentralized discrete-time systems was investigated in the literature, and that work is based on optimizing the observation matrices with constraints. The observation matrices with constraints can be considered as a natural optimality measure, since the coefficients of these matrices are a valid evidence to measure state variables by the controller. Therefore, a smaller number of output variables can be achieved by optimally allocating a linear combination of these measurements. In this paper, we extend this idea to the decentralized unified regulator problem using output feedback. For application, structural systems having natural frequency can be considered, where sensors can be located in a finite number of positions. To verify the validity and effectiveness, numerical results are presented.