PEM-Identification of a block-oriented nonlinear stochastic model with application to room temperature modeling

This contribution proposes a block-oriented nonlinear stochastic model, that can be identified by prediction error minimization (PEM) method. The model consists of a linear time invariant (LTI) stochastic block and a static nonlinear output feedback. Using measured data of a conference and an office room, it is demonstrated that this approach can be used for the identification of models suitable for short term room-temperature prediction.