PSO Adaptive Fading Memory Kalman Filter Based State Estimation of Indoor Thermal Model with Unknown Inputs

Abstract

An adaptive filtering approach is proposed in this paper to address the thermal state estimation methodology along with the model parameters jointly for an indoor thermodynamic resistance capacitance model with uncertain stochastic heating inputs. The adaptive dynamics of the state of the model is combined with a particle swarm optimization (PSO) based metaheuristic approach to feed the knowledge of measurement noise statistics and the initial estimation error covariance along with forgetting factor for implementation of fading memory Kalman filter (FMKF). This study has been carried out with the variation of uncertain influential input information to enhance the estimation efficiency with the proposed PSO adaptive FMKF (PSO-AdFMKF) strategy for a real life the test thermodynamic environment scenario inside the building space. Potential observations demonstrate that the proposed estimation algorithm performs encouragingly, with a satisfactory improvement of estimation performance in terms of evaluating error metrics.