Prototyping of Singular Value Reconstruction Precoding for Reliable Non-Orthogonal IoT Signals

Massive connectivity is one of the main research directions for beyond 5G. The cellular based narrowband IoT (NB-IoT), enabled by the orthogonal frequency division multiplexing (OFDM) signal, is an important technique. To evolve into the beyond 5G era, non-orthogonal concepts are preferred to re-shape the NB-IoT to provide higher spectral efficiency, wider coverage and lower power consumed services. This work investigates a non-orthogonal waveform in next generation IoT (NG-IoT) scenarios. Previous work has verified the advantages of zero forcing (ZF) precoding in interference mitigation but with some limitations. This work proposes a singular value reconstruction (SVR) precoding method, which can improve the precoding reliability and greatly reduce noise sensitivity. Simulations show significant spectral efficiency gain when compared with the previous work. An experiment platform is then configured in an over-the-air multiuser multiple input multiple output (MIMO) scenario to verify the practical feasibility of the precoding algorithm