Indoor Wireless Channel Properties at Millimeter Wave and Sub-Terahertz Frequencies

I. INTRODUCTION

• Spectrum from 100 GHz to 3 THz is presented as promising for next-generation wireless systems because of its wide swaths of unused and unexplored spectrum.The passage names applications including broadband IoT, sensing, communications, and wireless fiber links.Source: §I · p. 1
• At sub-THz frequencies, wavelengths comparable to or smaller than many objects’ surface roughness suggest scattering may not be negligible relative to reflection and diffraction.Source: §I · p. 1
• Prior indoor measurements at 100–400 GHz found 1 Gbps achievable on LOS and specular-reflection NLOS links, while frequencies beyond 100 GHz bring propagation challenges.The cited challenges include high phase noise and Doppler, limited output power, and more directional beams.Source: §I · p. 1

II. SCATTERING AND REFLECTION MEASUREMENTS AT 28, 73, AND 142 GHZ

• Measured 142 GHz path loss agrees well with the Friis free-space path-loss equation, supporting the reported calibration accuracy.The verification used transmitter-to-receiver separations of 1, 2, 3, 4, and 5 m.Source: §II · p. 2
• The close-in path-loss model with a 1 m reference distance fits the 142 GHz measured data, indicating its viability above 100 GHz.Source: §II · p. 2
• The measurements use a standard calibration and verification approach intended to support repeatable measurements across research teams and frequencies.Source: §II · p. 2

A. Measurements Setup

• An earlier comparison of 1.7 GHz and 60 GHz reported frequency- and temperature-related variation in building-material electrical parameters.Source: A. Measurements Setup · p. 2
• TX and RX antennas were placed 1.2 m high on a 1.5 m-radius arc, with the radius chosen to exceed the Fraunhofer distance.Narrow-beam horns with 10°, 7°, and 8° half-power beamwidths were used at 28, 73, and 142 GHz, respectively.Source: A. Measurements Setup · p. 2
• Drywall was tested at incidence angles of 10°, 30°, 60°, and 80°, while received power was sampled from 10° to 170°.Power at 0° and 180° could not be measured because of antenna size.Source: A. Measurements Setup · p. 2

B. Reflection at mmWave and THz

• At 142 GHz, reflection loss ranges from 0.36 dB at 80° incidence to 9.81 dB at 10°, and decreases linearly as incidence angle increases.Source: B. Reflection at mmWave and THz · p. 2
• The magnitude of the drywall reflection coefficient increases linearly with incident angle, and this linear fit performs better than the Fresnel equation at all three measured frequencies.Source: B. Reflection at mmWave and THz · p. 2
• Reflections are stronger at higher frequencies; the estimated drywall permittivity values are 4.7, 5.2, and 6.4 at 28, 73, and 142 GHz, respectively.Source: B. Reflection at mmWave and THz · p. 2

• At 142 GHz, measured peak power was 9.4 dB higher at 80° than at 10°, with most energy due to reflection rather than scattering.The comparison is for drywall.Source: C. Scattering at mmWave and THz · pp. 2–3, Figure 2 · p. 3
• At 142 GHz, backscattered power at 10° and 30° incidence was more than 20 dB below peak received power, leaving specular reflection the main propagation mechanism.The authors therefore consider drywall smooth even at 142 GHz.Source: C. Scattering at mmWave and THz · p. 3
• The dual-lobe scattering model plus reflected power matched measured peak power at the specular angle within 3 dB, while scattering at other angles fell off rapidly.The model used estimated permittivity values for each frequency.Source: C. Scattering at mmWave and THz · p. 3

• Partition loss is the difference between signal power immediately before and after a partition.Source: §III · p. 3
• The measure includes both reflection/scattering loss and material absorption loss.Source: §III · p. 3
• Accurate partition-loss models for common building objects are needed for wideband mmWave and Terahertz networks and precise ray-tracing algorithms.Source: §III · p. 3

• XPD values calculated across five distances were within 1 dB at each frequency, validating the measurements.Source: Figure 3 · p. 4
• Measurements used far-field LOS distances of 3–5 m with boresight-aligned antennas and no nearby reflectors or obstructions.Cross-polarization was produced by rotating an antenna 90° with a waveguide twist, whose insertion loss was calibrated out.Source: A. Antenna XPD Measurements · p. 4
• XPD was calculated as the path-loss difference between co-polarized and cross-polarized antenna configurations at a given distance.Source: A. Antenna XPD Measurements · pp. 3–4
• At a fixed separation, H-H and H-V received powers were within 1 dB of V-V and V-H, respectively, showing reciprocity in the cross-polarization measurements.Source: A. Antenna XPD Measurements · p. 4

B. Partition Loss Measurements

• Measurements used 14.5-cm drywall and 0.6-cm clear glass, four antenna polarization pairs, and 3–5 m transmitter–receiver separations.Five measurements at slightly shifted positions were averaged using the first-arriving multipath component; the setup placed the material in the transmitter's far field.Source: B. Partition Loss Measurements · p. 4
• Clear-glass co-polarized partition loss rose from 1.50 dB at 28 GHz to 10.33 dB at 142 GHz.Table III reports standard deviations of 0.50 dB and 0.24 dB, respectively.Source: B. Partition Loss Measurements · p. 4, Table III · p. 4, Figure 4 · p. 5
• Drywall co-polarized partition loss was 3.73 dB at 28 GHz, 2.87 dB at 73 GHz, and 8.89 dB at 142 GHz.Source: Table IV · p. 4, B. Partition Loss Measurements · p. 5, Figure 4 · p. 5
• After XPD subtraction, drywall partition loss was 6.40 dB at 28 GHz and negative at 73 and 142 GHz; clear-glass V-H loss exceeded H-V by 9.55 dB at 142 GHz.The drywall results indicate depolarization at 73 and 142 GHz; clear-glass cross-polarization loss differences were negligible at 28 and 73 GHz.Source: B. Partition Loss Measurements · pp. 4–5
• At 142 GHz, drywall reflected about 18.8% and transmitted 14.3% of incident power, with about 66.9% absorbed.The estimated absorption corresponds to about 4.8 dB real absorption loss.Source: B. Partition Loss Measurements · p. 5

IV. INDOOR PROPAGATION MEASUREMENTS AND PATH LOSS MODEL AT 142 GHZ

• Measurements used the same indoor locations as prior 28 and 73 GHz work, with TX antennas at 2.5 m and RX antennas at 1.5 m.The multipath-rich site included hallways, meeting rooms, cubicle offices, a laboratory, and an open area.Source: §IV · p. 5
• At 142 GHz, indoor path-loss models are similar to those below 100 GHz; LOS has slightly more loss and NLOS slightly less than at lower frequencies.The authors attribute the LOS difference likely to atmospheric attenuation and the NLOS difference likely to greater reflected power.Source: §IV · p. 5, Figure 5 · p. 5
• LOS path-loss exponents are 1.7 at 28 GHz, 1.6 at 73 GHz, and 2.0 at 142 GHz.Source: §IV · p. 5

V. CONCLUSION

• At 140 GHz, drywall reflection loss ranges from 0.4 dB at 80° incidence to 9.8 dB at 10°; backscatter is more than 20 dB below peak received power.For smooth surfaces such as drywall, a reflective-surface approximation is reasonable, especially near grazing incidence.Source: §V · pp. 5–6
• Clear glass and drywall induce depolarization, making partition loss polarization-dependent, with the effect becoming more prominent as frequency increases.Cross-polarized partition loss is below the value expected from antenna XPD and co-polarized partition measurements.Source: §V · p. 6
• Indoor path loss is broadly similar across 28, 73, and 140 GHz, with slightly greater large-scale LOS loss and slightly less NLOS loss at 140 GHz.The authors suggest absorption may explain the LOS difference and stronger reflections the NLOS difference.Source: §V · p. 6