LoRa Performance Enhancement through List Decoding Technique

Recently, Long Range (LoRa) has developed itself as one of the dominant technologies for Low Power Wide Area Networks (LP-WAN). LoRa is primarily based on its patented chirp spread spectrum (CSS) modulation. The patented LoRa physical layer supports Forward Error Correction (FEC) codes that are basically derived from simple Hamming codes. Such Hamming codes are restricted to either mere error detection or single error correction. However, burst errors actually exceed the correction capability of standard FEC. Accordingly in this paper, we propose a list decoding technique that helps in resolving the burst errors that could be encountered due to noise as well as interference effects. The proposed list decoding technique depends on widening the search in a list of candidate symbols that are considered based on a pre-defined list size. In order to easily evaluate the performance of the proposed list decoding technique, this paper also introduces an analytical model for the resultant block error rate (BLER) under additive white Gaussian noise (AWGN) channel. The proposed technique is evaluated at different coding rates, different Signal to Noise Ratio (SNR) and different Signal to Interference Ratio (SIR) scenarios. Results demonstrate that the proposed list decoding technique significantly enhances the BLER performance compared to the nominal LoRa case. In case of spreading factor sf =7, coding rate 4/7 and SIR γ = 0.25 dB, the SNR gain of the proposed technique reaches 15 dB compared to the nominal LoRa performance at BLER 10−4. In addition, results verify the accuracy of the derived BLER analytical expression compared to the simulations.