A waveform designed to carry on-off keying, or OOK, signals alongside conventional OFDM data was associated with better OOK detection in simulations and lower cyclic-prefix (CP) mismatch, which can leak interference into legacy data subcarriers. The pattern was not uniform: mismatch improved for most tested sequences, while the channel and the sharpness of the OOK signal shaped the reported interference.
How the signal is fitted to an OFDM frame
The design generates OOK over the full duration of a CP-OFDM symbol. It reserves W subcarriers for the OOK waveform and leaves D for data, following the relation . Here, N is the total number of OFDM subcarriers, W is the number used for OOK and D is the number left for data.
The model requires the first and last L samples of each OOK waveform block to be identical at the CP boundary. A failure to meet that condition creates CP mismatch. The analysis uses the mismatch energy to bound leakage from OOK subcarriers into the data subband, so reducing the mismatch can lower average leakage.
The method chooses the block-wide time shift that minimizes CP mismatch. The OOK cyclic prefix is generated independently before the OOK signal is superimposed on the CP-OFDM data signal.
That added control requires additional signal processing. The proposed generator is described as using two W-DFT processes and two N-IDFT operations, compared with one W-DFT process and one N-IDFT operation in the conventional system.
The reported numerical setup used 32 chips per OFDM symbol, eight OFDM symbols in an OOK block, a DFT size of 512 and 64 OOK subcarriers. It tested 256 arbitrary OOK sequences.
The gains were strongest in the mismatch and OOK tests
Across the 256 arbitrary OOK sequences, optimal shifting was reported to improve mismatch for most sequences. Some needed no shift, while others showed little reduction. Because the analysis uses mismatch energy to bound leakage, a lower mismatch can provide a tighter interference bound and potentially lower average leakage into legacy data subcarriers.
In the OOK receiver tests, the proposed scheme was reported to produce the same bit-error rate, or BER, across the tested OOK sequences. The conventional orthogonal method without receiver-side cyclic-prefix removal showed a clear error floor in both single-tap and multi-tap channel cases. Receiver-side CP removal was reported to improve BER and remove the error floor in the single-tap case, although performance depended on the device's sampling-frequency offset.
Existing data still feels the channel trade-off
The legacy-data simulations reported greater SINR degradation as the channel included more taps. Leakage was described as tolerable for a four-tap channel, but the worst reported case, Seq1, showed about a 10 dB SINR drop at an SNR of 20 dB in a 16-tap channel. SINR is the study's measure of the desired signal against interference and noise, so the drop marked more degraded conditions for the existing data stream.
In a separate moderate-SNR case, the reported worst-case leakage did not exceed 1 dB at an SNR of 4 dB. Among the OFDM data users, the worst-case BER performance loss was around 0.5% at an SNR of 20 dB and was less evident at lower SNR values.
The waveform shape also mattered. The ideal rectangular OOK signal produced more pronounced interference than the smoother OOK signal described as more realistic. The comparison shows that OOK generation details can affect leakage into legacy data subcarriers.
What the simulations leave open
Subcarrier orthogonality, the separation that helps keep OFDM data streams apart, is intentionally relaxed in the proposed design, so leakage into legacy data subcarriers remains possible. The design also adds processing by independently generating the OOK cyclic prefix. These trade-offs matter because the reduction in mismatch was not uniform across all tested sequences.
The reported results varied with channel taps, SNR and OOK sequence. The analytical bound also depends on channel delays and coefficients, while the comparison between rectangular and smoother OOK signals shows that waveform shape matters. The numerical findings therefore describe the tested combinations rather than guaranteeing the same behavior for every channel or waveform.
The study was supported in part by TUBITAK under Grant Number 124N803. The document front matter displays an arXiv version dated 28 August 2026.
Paper data and sources
Original title: CP-Aware OFDM-Based OOK Signaling
Authors: Badr Eddine Ouakouak, Salah Eddine Zegrar, Hüseyin Arslan
Journal/Repository: Ouakouak, B.E., Zegrar, S.E. and Arslan, H., 2025. CP-Aware OFDM-Based OOK Signaling. IEEE Wireless Communications Letters, 15, pp.935-939
Status: Peer-reviewed
First online: 2026-08-28
DOI: 10.1109/lwc.2025.3638749
Original paper · Full text