A least squares interpretation of sub-space methods for system identification

So called subspace methods for direct identification of linear models in state space form have drawn considerable interest. The algorithms consist of series of quite complex projections, and it is not so easy to intuitively understand how they work. They have also defied, so far, complete asymptotic analysis of their stochastic properties. This contribution describes an interpretation of how they work. It specifically deals with how consistent estimates of the dynamics can be achieved, even though correct predictors are not used. We stress how the basic idea is to focus on the estimation of the state-variable candidates-the k-step ahead output predictors