Nonlinear prediction fusion of cost functions in multisensory linear dynamic systems

Abstract

In this paper we propose multisensory fusion prediction algorithms for an arbitrary nonlinear cost function (NCF) consisted of system state variables. Some specific applications demand such measurands that can often be embodied by NCF since NCFs indicate useful information for instrumentation and control systems. Especially, prediction of NCF in multisensory environment is useful in target-tracking area. Hence, we discuss multi-sensor fusion prediction (estimation) of cost function based on optimal mean square solution. To carry out fusion prediction of NCF, we propose centralized and decentralized fusion prediction algorithms. The numerical simulations demonstrate accuracy and effectiveness of proposed work in terms of mean square error (MSE). © 2012 ICIC International.