An Automatic Modulation Classification Method for Multi-signal Scenario Based on ULA

Traditional digital modulation recognition methods exist following defects including performance deterioration at low SNR and lack of recognition capability when facing multi-signal scenario. Array antenna can separate signals individually by spatial filtering, improve SNR of received signal, and will contribute to multi-signal digital modulation recognition using traditional methods designed for single source. Taking advantage of uniform linear array, a multi-signal modulation recognition method was proposed which uses high-order cumulant feature, instantaneous statistical amplitude feature and high-order cumulants of signal after frequency conversion, and finally classical decision-tree classifier is used for digital modulation classification. Simulation results showed that the method proposed in this paper can recognize multi-signal digital modulation efficiently with satisfactory recognition rate at low SNR, and meanwhile influence factors including DOA estimation error and multi-signal array manifold were studied and analyzed.