Classification of Smart Jamming Using CNN for Anti-Jamming in the OFDM System

Abstract

Orthogonal frequency division multiplexing (OFDM) systems are vulnerable to smart jamming attacks that utilize a structure of a transmission signal. The smart jammer includes pilot tone jamming (PTJ) and cyclic prefix jamming (CPJ) for destroying channel and synchronization estimation, respectively. Thus, a performance comparison for each smart jamming and an anti-jamming strategies depending on the jamming types are required for the reliable system. In this paper, we analyze the effect of smart jamming on CP-based synchronization which is related system performance. Directly attacking CP used for synchronization causes the greatest degradation in symbol timing and carrier frequency synchronization.

Jamming classification is required to properly select anti-jamming strategies against each smart jamming for the reliable system. This paper suggests CNN with concatenation of time-wise and frequency-wise pooling considering the characteristics of jamming signals on time-frequency plane. We show an accuracy, a false alarm rate, and a miss rate of proposed model for each jamming signals. Finally, we also propose an adaptive strategy that differentiates power in the limited transmission power allocation according to each jamming situation. the power allocation with the proposed classifier improves the BER performance of the OFDM system.