Adaptive Multichannel Scanning Strategies for Electronic Support Using Multi-Armed Bandits

In Electronic Support Measures, the objective is to listen to a wide portion of the radio frequency spectrum to detect potential threat emitters. This can be realized by a relatively cheap system, the superheterodyne receiver. Its narrow instantaneous frequency band can be scanned along the whole spectrum. With this approach, the probability of intercepting emitters is intrinsically low. A first solution to this problem is to use multichannel superheterodyne receivers which can listen to multiple frequency bands at once. In this paper, we propose 4 adaptive stochastic algorithms of increasing complexity that exploit information from every channel to improve interception performance and monitoring capabilities. These algorithms are based on the theory of Multi-armed Bandits and can estimate antenna scan pattern of emitters to improve efficiency.Based on the adopted methodology and simulations carried out, we achieve greatly improved pulse interception rates while keeping intercept times similar to random exploration.