Channel Randomness-Based Adaptive Cyclic Prefix Selection for Secure OFDM System

A novel method for cyclic prefix (CP) adaptation based on wireless channel randomness is proposed for orthogonal frequency-division multiplexing (OFDM) systems in urban areas. Especially, a quantization algorithm and an adaptive window-based strong subcarriers selection are designed on the sorted channel frequency response of the legitimate party to select randomly the secure CP (SCP) length for each OFDM symbol without the need to share any information between legitimate nodes. The proposed method creates a problem for illegitimate user's synchronization as it does not have any information about SCP length for every OFDM symbol, even if it uses blind synchronization techniques. The effectiveness of the proposed mechanism is evaluated through representative metrics, such as bit error rate (BER), throughput in both perfect and imperfect channel estimation scenarios, mismatch probability (MP), and cyclic autocorrelation function (CAF). Also, the performance for the presence of a correlated eavesdropper (Eve) in the system is evaluated. Simulation results show a huge secrecy gap between BER and throughput performance of legitimate and illegitimate users in all scenarios.