Q-Learning Based Equivalent False Target Generation for Multi-Point Source Jamming System
Abstract
In the multi-point source jamming (MPSJ) system, the antenna phase error, as an intrinsic issue, seriously affects the distribution of equivalent false targets (EFTs). To ensure that the position of EFTs remains within the constraint range in the presence of phase error, a multi-point source jamming system equivalent false target generation method is proposed in this paper. First, by establishing the spatial geometric relationship between the triplet antenna and the target, the distribution law of EFT in the MPSJ system is derived. Second, based on the feasibility of the MPSJ system EFT generation method, the MPSJ system EFT generation model is constructed by introducing the constraint function. Third, a dual dynamic inductive Q-learning (DDIQL) algorithm with a head-tail dynamic inductive strategy is proposed to solve this model, wherein initial optimization and constraint adjustment are emphasized. Experiments results demonstrate that when the phase error is 9°, the EFTs corresponding to the generated amplitude and phase parameters are all controllable in position. When the radius of the constraint range is reduced to the baseline length of the MPSJ system, the positions of EFTs corresponding to more than 10% of the amplitude and phase parameters are outside the constraint range.