A novel state estimation strategy for observation recovery in nonlinear systems based on ExpARMA algorithm

A more effective state estimation scheme of nonlinear systems having loss at output has been discussed in this paper. The existing model of reproducing lost measurement in nonlinear systems is Exponential Autoregressive or ExpAR which incorporates only output measurements for compensation. In this work, however, the compensated measurement is based on Exponential Autoregressive Moving Average (ExpARMA) series. The proposed model incorporates relatively more particulars compared to the existing models based on AR, ARMA or ExpAR etc. The ExpARMA scheme utilizes output, as well as input measurements, which bears more optimal outcomes at the cost of higher computational efforts. Important steps of calculating nonlinear prediction coefficients are carried out to assist the input signals. A trade-off could be considered between efficient results and computational time. To test and compare the performance of proposed algorithm with existing compensation techniques, two-phase Permanent Magnet Synchronous Motor (PMSM) has been simulated as a case study. •Problem of State Estimation with intermittent measurement has been addressed.•A more sophisticated Exponential ARMA (ExpARMA) model is introduced.•The ExpARMA model is integrated with nonlinear estimation tools.•The Nonlinear prediction order is determined using MMSE technique.