Centralized fusion estimation for uncertain multisensor system based on LMI method

This paper gives the models of uncertain multisensor system based on norm-bounded parameter uncertain model method and convex bounded uncertain model method, respectively. The corresponding centralized robust fusion estimations are developed, too. Simulation compares different fusion estimation methods detailedly in frequency and time fields and concludes that each of two filters has its strong point at restraining disturbance attenuation within some frequency range. The filter based on former model can obtain less estimation covariance, but may encounter numerical value problem and lead to a ´bad´ solution. The filter based on latter model will become complicated when the number of sensors increasing. The results of this paper develop the theory basis of state fusion estimation for the uncertain multisensor system in practical.