Millimeter Wave Propagation Modeling in Urban Environments

mmWave (30–300 GHz) propagation in urban envs faces severe path loss, atm abs (O2/H2O res at 60/183 GHz), blockage from buildings, vehs, humans. Fund: Friis eq modified for high-freq path loss; PL(d) = PL0 + 10n log10(d/d0) + Xσ, where n=env-dep path loss exp, Xσ=SHADOW (log-normal). LOS prob models: alpha-complex, ray-launching (RL), stochastic geom (e.g., Poisson line proc). Key: blockage stat via geom blockage models—buildings modeled as random rect obstacles; human blockers as dynamic cyls. Urban microcells (UMi) & street canyons induce multipath: strong spec refl (concrete εr≈4–9), diff, surface wave creep, non-spec BR via vol scat (fog, rain). MIMO & beamforming gain essential to combat path loss; narrow beams (≈5°) increase dir gain but sens to align & blockage. Ch models: 3GPP TR 38.901 std defines UMi, UMa, RMa scenarios w/ clustered delay line (CDL) & tapped delay line (TDL) models. CDL-A/B/C for low/mid/high delay spreads; TDL for fixed links. Param: delay spread (DS ≈10–100 ns urban), ang spread (AS), Doppler (veh mobility ≈30–100 km/h → fD ≈0.1–1 kHz at 28 GHz). Time-variance: blockage dynamics (pedestrian ≈1.2 m/s, veh ≈15 m/s) → link outage; handover freq ↑ due to small cell densification. Sim tools: Remcom Wireless InSite, Altair Feko, NYUSIM (open-source, based on 3D mmWave meas). NYUSIM outputs: time-clustered rays w/ Poisson arrival, log-normal pow within cluster. Real-world cal: channel sounder meas @28/39/73 GHz in NYC, Seoul show: avg path loss exp n≈2.0–4.0 (LOS/NLOS), SHADOW σ≈3–8 dB, RMS DS≈30–150 ns. Material dep: glass τ≈0.9 @28 GHz, brick τ≈0.1; foliage loss ≈0.2–0.5 dB/m. Rain fade: specific att ≈0.1–20 dB/km (f↑, rate↑). Mitigation: multi-connectivity (MC), reconfig metasurf (RIS), predictive beam track (ML-based), V2X coord. ML apps: RF lensing via CNN to pred path from GIS; LSTM for blockage state trans; GNNs for topo-aware propagation pred. Pitfalls: over-reliance on ray-optics (neglect surf diffr, leaky modes); poor human blockage modeling (static vs. flow); ignoring polarization dep (HV/VH cross-pol rise in scat envs); misestimating coherence time (Tc≈1/fD) → beam misalign. SOTA: hybrid deterministic-stochastic models (e.g., D-TDD), RIS-aided propagation control, sub-THz (100–300 GHz) urban ch char (＞100 GHz show higher directivity but extreme atm loss). Challenges: real-time ch est in mobile mmWave; energy eff beamforming; scalable site-specific modeling. Future: AI-driven digital twins, joint com-sens framework, THz urban ch models (＞300 GHz) with molecular abs lines.