Measurement optimization with minimax criteria for parameter estimation in distributed systems

A computational procedure is presented for designing moving sensor trajectories in parameter identification of two-dimensional distributed systems. Two widely used minimax criteria, namely those of MV-optimality and E-optimality, are considered here as the measures of the estimation accuracy. The approach converts the problem to an optimal control one in which both the control forces of the sensors and the initial sensor positions are optimized. A gradient algorithm is developed based on the negative formulation of the Potryagin Maximum Principle which handles various constraints imposed on the sensors´ motions. A summary of numerical experiments with the techniques proposed is given.