A scalable multi-step least squares method for network identification with unknown disturbance topology

Identification methods for dynamic networks typically require prior knowledge of the network and disturbance topology, often rely on solving poorly scalable non-convex optimization problems. While estimating topology are available in literature, less attention has been paid to i.e., (spatial) noise correlation structure rank a filtered white representation signal. In this work we present an identification method networks, which estimation precedes full with known topology. To end extend multi-step Sequential Linear Regression Weighted Null Space Fitting deal reduced noise, use these estimate dynamics measurement situation. As result, provide least squares algorithm parallel computation capabilities that only explicit analytical solutions, thereby avoiding usual optimizations involved. Consequently consistently Box Jenkins model structure, while keeping computational burden low. We consistency proof includes path-based data informativity conditions allocation excitation signals experimental design. Numerical simulations performed clearly illustrate potential method.