Frequency domain preamble-based channel estimation and equalisation in LoRa

Abstract

This paper deals with multipath channel estimation and equalisation in low range (LoRa). It is suggested to take advantage of the cyclic property of the symbols in the LoRa frame preamble to obtain an interference-free version of the symbols in the frequency domain. Then, estimation methods used in multicarrier systems can be applied, such as the least square (LS), and the minimum mean square error (MMSE) estimators. It is shown that the cyclic property in LoRa is inherently independent of the length of the channel, making these estimation techniques robust to any frequency-selective channel. In addition the frequency domain zero-forcing (ZF) equaliser is used, and an original phase equaliser is introduced, taking advantage of the constant modulus property of LoRa symbols in the frequency domain. The performance of the investigated estimators and equalisers is shown through simulations, and applications to the presented results are further discussed.