[Paper Review] Path loss, beamforming gain and time dynamics measurements at 28 GHz for 90% indoor coverage.

Dmitry Chizhik, Jinfeng Du+4 more|arXiv (Cornell University)|Dec 18, 2017

Millimeter-Wave Propagation and Modeling21 references12 citations

TL;DR

This study presents extensive 28 GHz indoor measurements using a narrowband channel sounder to characterize path loss, beamforming gain, and time dynamics for 90% coverage. It demonstrates that 30–32 dB excess loss occurs when turning corners or entering rooms, and system simulations show every hallway requires an access point to achieve 1 Gbps in 90% of locations with 1 GHz bandwidth.

ABSTRACT

Achieving adequate coverage with high gain antennas is key to realizing the full promise of the wide bandwidth available at mm/cm bands. We report extensive indoor measurements at 28 GHz, with over 1000 links with and without Line-of-Sight (LOS) using a specialized narrowband channel sounder, capable of reliable measurements up to 171 dB path loss to characterize 90% coverage. Azimuthal power spectra were captured with 1-deg granularity using a 10-deg receive horn spun at speeds up to 300 rpm. Measured path gain-distance dependence in LOS and NLOS are well represented by power-law models, with 3.1 dB standard deviation in NLOS, and by a mode-diffusion model with 3.5 dB RMS error. Excess loss at 28 GHz suffered in turning a corner or into a room was found to be 30 and 32 dB respectively, in contrast to 20 dB found previously at 2 GHz. Degradation of azimuthal gain by scattering was within 4 dB in the hallway and 7 dB inside a room with 90% probability. System simulations in a canonical building indicate that every hallway needs an access point to provide 1 Gbps rate in 90% of locations using 1 GHz of bandwidth. For stationary terminals, with temporal fading caused by pedestrians, long term average-power-based aiming was within 3.7 dB of rapid beam switching in 90% of high traffic locations.

Motivation & Objective

• To quantify path loss and beamforming performance at 28 GHz in indoor environments with and without line-of-sight (LOS).
• To determine the excess path loss incurred when signals propagate around corners or into rooms at 28 GHz.
• To evaluate the impact of scattering on beamforming gain in hallways and indoor rooms.
• To assess the feasibility of achieving 1 Gbps data rates in 90% of indoor locations using 1 GHz bandwidth at 28 GHz.
• To compare long-term average power beamforming with rapid beam switching under pedestrian-induced temporal fading.

Proposed method

• Conducted over 1,000 indoor links at 28 GHz using a specialized narrowband channel sounder capable of measuring up to 171 dB path loss.
• Captured azimuthal power spectra with 1-degree resolution using a 10-degree receive horn spun at speeds up to 300 rpm.
• Modeled path loss-distance dependence using power-law models for LOS and non-LOS (NLOS) conditions.
• Applied a mode-diffusion model to represent NLOS propagation with a 3.5 dB RMS error.
• Simulated system performance in a canonical building to determine access point density requirements for 1 Gbps coverage.
• Evaluated beamforming strategies by comparing long-term average power-based aiming with rapid beam switching under temporal fading from pedestrians.

Experimental results

Research questions

• RQ1 What is the path loss-distance dependence at 28 GHz in LOS and NLOS indoor environments, and how well can it be modeled using power-law functions?
• RQ2 What is the excess path loss when signals propagate around corners or into rooms at 28 GHz, and how does it compare to lower frequencies?
• RQ3 How much degradation in beamforming gain occurs due to scattering in hallways and indoor rooms, with 90% probability?
• RQ4 What access point density is required to achieve 1 Gbps data rates in 90% of indoor locations using 1 GHz of bandwidth at 28 GHz?
• RQ5 How does long-term average power-based beamforming compare to rapid beam switching in terms of performance under temporal fading from pedestrians?

Key findings

• Non-line-of-sight (NLOS) path loss at 28 GHz exhibited a 3.1 dB standard deviation when modeled with a power-law function.
• Excess loss was measured at 30 dB when turning a corner and 32 dB when entering a room, significantly higher than the 20 dB observed at 2 GHz.
• Beamforming gain degradation due to scattering was within 4 dB in hallways and 7 dB inside rooms with 90% probability.
• System simulations indicated that every hallway requires an access point to achieve 1 Gbps data rates in 90% of locations using 1 GHz of bandwidth.
• For stationary terminals, long-term average power-based beamforming achieved performance within 3.7 dB of rapid beam switching in 90% of high-traffic locations.