Adaptive EMS Test Design Method on UAV Data Link Based on Bayesian Optimization

The complex electromagnetic environment makes the electromagnetic susceptibility (EMS) test challenging for the unmanned aerial vehicle (UAV) data link. The EMS test can be regarded as an optimization problem to obtain the maximum signal-to-interference ratio that makes the data link losing lock. In this article, we proposed an adaptive EMS test design method based on Bayesian optimization to expand the ability of the EMS test toward the UAV data link. The sensitive electromagnetic interference parameters can be automatically searched according to the adaptive test design method. The Gaussian process is used as the surrogate model for the EMS threshold. Besides, acquisition functions are discussed, including the probability of improvement, expected improvement, and upper confidence bound to balance the exploration and exploitation. Then, two test examples are applied to verify the adaptive test design method. Compared with the traditional EMS test methods, the proposed method introducing Bayesian optimization into the EMS test design can significantly improve the test efficiency and achieve an excellent human-in-the-loop effect.