Blind CP Length Detection

The core mechanism exploits the redundancy introduced by the cyclic prefix. By computing the autocorrelation of the received signal at a lag equal to the useful symbol duration, a correlation peak emerges at the CP boundary. The width of this peak directly corresponds to the CP length.

• Normal CP: Produces a narrow correlation plateau
• Extended CP: Produces a wider correlation plateau
• The lag value at which the peak occurs reveals the FFT size

The blind detector formulates CP length estimation as a hypothesis testing problem. For each candidate CP length, a likelihood metric is computed based on the correlation magnitude within the hypothesized CP region.

• Computes the log-likelihood ratio for normal vs. extended CP
• Integrates correlation energy over the candidate CP window
• Selects the hypothesis that maximizes the accumulated correlation
• Robust to moderate frequency offsets due to differential processing

To improve detection reliability under low SNR, the autocorrelation output is accumulated over multiple OFDM symbols using a sliding window. This temporal averaging reinforces the periodic correlation structure while suppressing noise.

• Window length typically spans 5–10 OFDM symbols
• Coherent accumulation enhances the CP-induced peak
• Non-coherent combining mitigates phase drift from residual CFO
• Trade-off: longer windows improve SNR but increase acquisition time

The detector distinguishes between CP modes by comparing the duration of the correlation plateau. In LTE, the normal CP is approximately 4.7 µs (for symbol 0: 5.2 µs), while the extended CP is 16.7 µs.

• Threshold-based decision: Plateau width > threshold → Extended CP
• Ratio test: Compares correlation energy in early vs. late CP regions
• Handles mixed CP configurations within a subframe
• Critical for proper FFT window alignment in subsequent demodulation

Blind CP detection is often performed jointly with symbol timing and carrier frequency offset estimation. The same autocorrelation computation provides all three parameters simultaneously.

• Symbol timing: Derived from the position of the correlation peak
• Fractional CFO: Estimated from the phase angle at the peak lag
• CP length: Inferred from the peak width or plateau extent
• This joint estimation reduces computational overhead compared to sequential approaches

In frequency-selective channels, the CP correlation peak may be spread or attenuated due to inter-symbol interference. Advanced detectors employ normalized correlation metrics to maintain reliability.

• Delay spread tolerance: Detector remains functional when delay spread < CP duration
• Normalization by signal energy compensates for fading amplitude variations
• Cyclostationary-based methods provide additional robustness in severe multipath
• Performance degrades when channel impulse response exceeds the CP length