Adaptive False-Target Recognition for the Proximity Sensor Based on Joint-Feature Extraction and Chaotic Encryption

Abstract

The false target caused by interference obfuscates the real target detection of the radio frequency proximity sensor (RFPS). In this article, an adaptive false-target recognition method is proposed to classify the target echo and the false target. Two strategies are designed for the recognition of false targets caused by different interference. In the first target recognition strategy, each RFPS is assigned an identity (ID) with chaotic encryption. For a better performance against the false target caused by barrage interference, joint features of the range profiles and Fourier spectrum are extracted in the second strategy. The Naïve Bayesian classifier is applied and it performs well against barrage interference. Since the two strategies perform differently under different interference, the game model between the RFPS and the interference is established to improve the false-target recognition stability of the RFPS. Based on the game model, the payoffs of the two strategies are evaluated, and thus the false-target recognition strategies can be adaptively allocated according to their payoffs. The simulation result verifies the improved stability of the proposed method under different interference strategies and signal-to-interference-plus-noise ratios (SINRs). The false-target recognition performance of the RFPS will be significantly improved with the proposed method.