Internet of underground things (IOUT) is an emerging paradigm which consists of sensors and communication devices, partly or completely buried underground for real-time soil sensing and monitoring. In this paper, the performance of different modulation schemes in IOUT communications is studied through simulations and experiments. The spatial modularity of direct, lateral, and reflected components of the UG channel is exploited by using multiple antennas. First, it has been shown that bit error rates of 10 −3 can be achieved with normalized delay spreads (τ d) lower than 0.05. Evaluations are conducted through the first software-defined radio-based field experiments for UG channel. Moreover, equalization has a significant impact on the performance improvement of an IOUT system. An 8-Tap DFE (decision-feedback equalizer) adaptive equalizer achieves better performance. It is also found that DBPSK, and DPSK are more suitable for digital communications in the UG channel without adaptive equalization. Then, two novel UG receiver designs, namely, 3W-Rake and Lateral-Direct-Reflected (LDR) are developed and analyzed for performance improvement. It has been shown that with a three antenna LDR design, BER of lower than 10 −5 can be achieved. The BER of these two approaches are compared and the LDR has been shown to perform better.