Phase-Rotation-Based CFO Estimation and Compensation Method for Reliable LoRa

In dynamic applications, carrier frequency offset (CFO) may lead to performance degradation in LoRa communication. Especially for the fractional CFO part, energy leakage and peak reduction in the demodulation domain pose additional challenges for accurate symbol detection after noise superimposition. Therefore, we propose a phase rotation based CFO estimation and compensation (PREC) method to achieve accurate fractional CFO estimation by supplementing frequency offset to an integer value to promote energy concentration in demodulation and improve noise resistance. By continuous phase rotations, a frequency offset is first added to the original CFO in the preamble of the received signal to form an integer value. After the integer CFO estimation in peak detection, the introduced frequency offset can be subtracted from the estimated integer value to obtain the original CFO, which helps compensate for the CFO in the data symbol for accurate detection. Monte Carlo simulation results demonstrate that the PREC exhibits robustness to CFO and helps to obtain better bit error rate (BER) performance with fewer upchirps in the preamble.