Sensing strategies for a flexible manipulator

In this paper, we investigate the combination schemes using multisensors for Kalman filtering for modal estimation of a flexible manipulator. In addition to observability, the performance indices regarding the global estimation error covariance, the sensitivity to the initial estimate errors, and the robustness to the round-off errors of the relative noise intensity between process and measurements, are utilized to evaluate the quality of the Kalman filters. The aim is to identify well-conditioned multisensor combinations for which Kalman filters can achieve a specified level of design performance. Case studies for the assessment of the multisensor combination schemes are carried out for Kalman filters with various relative noise intensities. The useful guidelines for achieving the well-conditioned combinations of multisensors are presented.