Research on RCS Sequence Prediction Based on ARIMA-TPA-LSTM Algorithm

The existing sequence prediction algorithm is difficult to meet the requirements of prediction accuracy when dealing with complex RCS. In this paper, a coupling method is proposed to effectively combine time series algorithm and neural network to accurately predict the relationship between single station RCS and incident Angle. Based on auto regressive integrated moving average(ARIMA) algorithm, bayesian information criterion (BIC) was proposed to optimize the order of the model.Then, temporal pattern attention(TPA) mechanism that pays more attention to RCS historical information is proposed and the residual value was combined with the long short term memory algorithm(LSTM) with TPA, and the ARIMA-TPA-LSTM prediction algorithm based on BIC was proposed to improve the prediction accuracy. At the same time, simulation test is carried out. The simulation results show that the accuracy of ARIMA-TPA-LSTM prediction algorithm proposed in this paper is higher than the existing prediction algorithm on average, and it can accurately deal with complex RCS prediction problems.