Design of Radar Waveform Using Adaptive-Network-Based Fuzzy Inference Systems

Abstract

Radar’s transmission waveform will affect the radar’s performance and the extraction of information from the echo of target. Therefore, radar waveform design is of great significance. There are two commonly used radar waveform design methods: the waveform design based on SNR criterion and MI criterion, but both two methods have their own limitations. In this paper, a new waveform design method that takes into account the advantages of both criteria is proposed. Adaptive-Network-Based Fuzzy Inference System (ANFIS) not only has the self-learning ability and self-adaptation ability of neural network, but also has the ability to understand rules in the fuzzy field. The data sets produced by two criteria are training data sets for the neural network. The trained ANFIS is to design waveform. Finally, it is found that the mutual information of the ANFIS waveform is better than the MI criterion, and the signal-to-noise ratio of the ANFIS waveform is better than the SNR criterion, that's to say that ANFIS-based waveform design takes into account the advantages of both based on SNR and MI criterion.