Online radar screening pulse width allocation strategy based on non-stationary bandit

With the development of jamming technologies, jammers can respond to radar signals within a short period, causing radar frequency-hopping strategies to fail. In the face of jammers with instantaneous frequency measurements (IFM) capabilities, radio frequency (RF) screening techniques are an effective countermeasure. However, in practice, the jamming strategy is usually unknown to the radar. Moreover, when the jammer's strategy changes, failing to promptly adjust the shielding pulse parameters will result in a significant degradation of the RF screening anti-jamming effectiveness. To overcome the limitations caused by the lack of information about the jammer and the non-stationary environment due to changes in the jammer's IFM time, this paper proposes an online radar screening pulse width allocation method based on a non-stationary multi-armed bandit (MAB). This method combines discounted historical rewards and sliding window rewards, allowing it to better adapt to the non-stationary jamming environment. Simulation results show that the proposed method has a faster convergence speed than the discounted method and a stronger exploration capability than the sliding window method. It can effectively enhance radar performance in countering instantaneous frequency measuring jammers in non-stationary jamming environments.