Phase-Coherent Wireless Channel Prediction

Extend the Physics-Unrolled Hybrid Neural Operator beyond received signal strength prediction to phase-coherent prediction of complex-valued wireless channels, including channel impulse responses and MIMO channel matrices, for applications such as beamforming, localization, and communication-system design.

Background

The paper’s Physics-Unrolled Hybrid Neural Operator predicts scalar received signal strength fields for indoor, single-transmitter environments. Its current output does not represent phase, channel impulse responses, angular spectra, or multiple-input multiple-output channel matrices, which limits its direct applicability to phase-sensitive wireless tasks.

A phase-coherent extension would require the operator and its supervision to represent complex-valued channel behavior rather than only scalar power. The paper identifies this extension as future work because such richer channel representations could support beamforming, localization, and communication-system design.

References

Phase-coherent extension is left to future work.

Physics-Unrolled Neural Operator for Wireless Field Modeling  (2608.18495 - Murshed et al., 19 Aug 2026) in Appendix, Section “Broader Impact and Limitations,” subsection “Scope and limitations”