Efficient System Identification of Heterogeneous Distributed Systems via a Structure Exploiting Extended Kalman Filter

We consider the problem of computationally efficient system identification of spatially heterogeneous distributed systems. We show how to parametrize distributed interconnected system models such that an extended Kalman Filter can be used for parameter estimation (system identification) in linear computational complexity with respect to the number of subsystems. Furthermore, the identified model will preserve the special sequentially-semi separable (SSS) structure, allowing for the use of efficient structured controller synthesis routines.