TL;DR
This paper characterizes how 28 GHz millimeter-wave signals—the backbone of 5G Fixed Wireless Access (FWA)—penetrate from outdoors to indoors. Using a high-performance real-time channel sounder with phased array antennas, the researchers quantified not just power loss, but also the critical temporal (delay) and spatial (angular) statistics across two different building types.
What Problem Does This Paper Address?
As the demand for high-speed broadband explodes, service providers are looking toward Fixed Wireless Access (FWA) at millimeter-wave (mm-wave) frequencies like 28 GHz as a cost-effective alternative to laying fiber-optic cables. However, mm-wave signals are notoriously fragile; they are easily blocked by walls, glass, and foliage.
While previous research has measured "penetration loss" (how much the signal dims when it goes through a wall), most studies lacked high-resolution data on delay spread (how much the signal echoes over time) and angular spread (the directions from which the signal arrives). These statistics are vital for engineers designing the beamforming algorithms that allow 5G base stations to "track" and maintain a solid connection with indoor receivers.
Methods
The researchers conducted a measurement campaign on the University of Southern California (USC) campus using a sophisticated hardware setup:
- Real-Time Channel Sounder: Operating at 27.85 GHz with a 400 MHz bandwidth. It can measure path loss up to 169 dB, allowing it to "see" very faint signals that have penetrated deep into buildings.
- Phased Array Antennas: Instead of slowly rotating a horn antenna, they used 8x2 phased arrays that steer beams electronically in microseconds. This allows for "directionally resolved" measurements even in dynamic environments.
- Scenarios: They tested two distinct environments to see how construction materials affect the signal:
- Single Family Unit (SFU): A two-story, wood-frame house (typical of California construction).
- Multi-Story Building (MSB): A brick building surrounded by heavy foliage.
Datasets
The paper introduces a new measurement dataset collected from these two locations:
- SFU Dataset: 5 outdoor and 12 indoor receiver (RX) locations.
- MSB Dataset: 3 outdoor and 9 indoor RX locations. The data includes omnidirectional and directional Power Delay Profiles (PDPs), which map out signal strength relative to time and arrival angle.
Key Insights and Ideas
- Construction is Destiny: The material of the building is the primary driver of signal quality. Wood-frame houses are significantly more "transparent" to 28 GHz signals than brick buildings.
- The Window Factor: In the wood-frame house, the loss varied wildly (from 1 dB to 21.5 dB) depending on whether the receiver was directly behind a window or shielded by a wall.
- Temporal Consistency in Homes: Surprisingly, in the single-family home, the "echoiness" (RMS delay spread) was very similar for both indoor and outdoor locations. This suggests that the signal penetrates the structure relatively cleanly without bouncing around excessively.
- The "Hallway Effect" in Brick Buildings: In the brick building, the strongest signals didn't always come through the nearest window. Instead, signals often reflected off other structures and entered through hallways or distant openings, leading to much higher delay spreads than those seen in the wooden house.
Key Results
- Mean Excess Penetration Loss:
- SFU (Wood): 10.6 dB.
- MSB (Brick): 22.7 dB.
- Path Loss: For the SFU, the outdoor path loss averaged 127.8 dB, while the indoor path loss averaged 138.4 dB.
- Delay Spreads: In the SFU, the mean RMS delay spread was ~24 ns (outdoor) vs. ~26 ns (indoor). In the MSB, the indoor delay spreads were significantly larger, indicating a more complex multipath environment.
- Angular Spread: While mean angular spreads were somewhat similar, the range of values varied significantly indoors, suggesting that indoor antennas need high adaptivity to maintain a link.
Limitations and Future Work
The authors acknowledge a few limitations:
- Limited Building Variety: Only two building types were tested. Results might differ significantly for modern "green" buildings with energy-efficient, metal-coated glass, which is known to be highly reflective to mm-waves.
- Receiver Height: All indoor measurements were taken at a fixed height on the first floor.
- Future Directions: The team plans to perform more measurements at different receiver heights and use high-resolution parameter extraction (like the RIMAX algorithm) to more precisely isolate individual multipath components.
Summary updated