Performance of adaptive equalization techniques in wireless underground sensor networks

This paper investigates different adaptive equalization techniques in three different communication scenarios viz; Underground-to-Underground (UG-UG), Underground-to-Aboveground (UG-AG) and Aboveground-to-Underground (AG-UG), using Mineralogy Based Spectroscopic Dielectric Model (MBSDM) which is preferred over the Peplinski Soil Dielectric model because of its ability to generate Complex Dielectric Constant (CDC) spectra of moist soils with considerably better accuracy and lesser prediction error. BPSK QPSK and 32 QAM modulated signals at 1 GHz are sent through the MBSDM channel. The received signals are equalized using RLS, LMS, and NLMS algorithms. Amongst these algorithms, RLS provides better performance in mitigating the Inter Symbol Interference (ISI) induced by the underground channel in a stationary channel environment as it converges much faster than LMS and NLMS. NLMS performs better than LMS as evidenced by plotting BER vs SNR. This study infers that adaptive equalization techniques help not only in recovering the original signal but also provides a deeper insight about underground channel models for efficient WUSNs.