Proposing Optimal ARMA Based Model for Measurement Compensation in the State Estimation

The purpose of this paper is to improve state estimation in the event of data loss by augmenting a novel Moving Average Autoregressive-based artificial measurement vector with Kalman filtering. The proposed technique replaces the existing Autoregressive-series based model embedded in the linear prediction techniques through Moving Average Autoregressive-based model. The Autoregressive scheme needs only one type of linear prediction coefficient to be tracked, while the proposed scheme computes two parameters at each recursion. Since Autoregressive Moving Average technique possesses more information, hence it efficiently predicts the future values of a signal. This value is placed as an alternative in the structure (or steps) involved in standard process of state estimations. The ultimate consequences of this extra computations involve more computational efforts. A standard mass-spring damper case study has been provided to show some aspects of the existing and proposed techniques.