Skip to Main Content
The paper considers the stochastic modelling of radar returns. In particular, returns from a typical airport surveillance radar (ASR) system have been modelled as autoregressive-moving average (ARMA) processes. Both maximum-likelihood (ML)- and autocorrelation-based techniques have been used. Order selection algorithms were studied and modified to optimise their performance for short-data records necessitated by the nonstationary radar environment. Distinctively different models have been found for typical combinations of ground, weather and aircraft returns.