⚡ electrical engineering

On the Spatial Consistency of Sub-Terahertz Channel Characteristics for Beyond-6G Systems

This paper experimentally demonstrates that sub-terahertz channel characteristics in indoor environments remain spatially consistent over distances of tens of centimeters, suggesting that ray-tracing models for beyond-6G systems can utilize significantly coarser spatial grids (10–50 wavelengths) along stable line-of-sight paths to drastically reduce computational complexity.

Original authors: Hossein Amininasab, Huda Farooqui, Dmitri Moltchanov, Sergey Andreev, Michele Polese, Mikko Valkama, Josep M. Jornet

Published 2026-04-21

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Hossein Amininasab, Huda Farooqui, Dmitri Moltchanov, Sergey Andreev, Michele Polese, Mikko Valkama, Josep M. Jornet

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to paint a hyper-realistic, 3D map of a room using invisible light beams (radio waves) to send data. This is what engineers do when designing the "Beyond-6G" internet, which will use super-high-frequency waves (sub-terahertz) to make internet speeds incredibly fast.

The big problem? These waves are so sensitive that, theoretically, if you move your phone even a tiny bit—like the width of a human hair—the "map" of the signal changes completely. If you had to redraw this map for every single hair-width movement, your computer would need to work harder than a supercomputer for a thousand years just to plan a single video call.

The Big Question:
Do these radio waves actually change that fast? Or is the "map" stable enough that we can skip drawing every single hair-width and just draw a grid instead?

The Experiment:
The researchers in this paper acted like detectives with a high-tech flashlight. They set up a transmitter and a receiver in a university classroom (which had tables and chairs, making it a bit messy). They moved the receiver along a cross-shaped path, stopping at hundreds of points.

They didn't just move it by inches; they moved it by millimeters (thinner than a pencil lead) all the way up to a meter. They measured the signal at every single stop to see how much the "picture" of the room changed.

The Discovery (The "Aha!" Moment):
Here is the surprising part: The signal didn't change nearly as fast as the theory predicted.

Think of the radio signal like a beam of sunlight shining through a window into a dusty room.

The Direct Beam (LoS): If you are standing in the direct sunlight, moving your head a few inches (or even a few feet) doesn't change the fact that you are still in the sun. The light is bright and steady.
The Dust Bunnies (NLoS): However, if you look at the dust motes dancing in the air (the scattered, indirect signals), moving your head just a tiny bit makes the dust pattern look completely different.

The researchers found that:

The "Sunlight" is Stable: The main, direct path of the signal stays consistent over distances of 10 to 50 times the length of the wave (which is about 2 to 10 centimeters). This is huge news! It means engineers don't need to recalculate the map for every millimeter. They can use a "coarse grid" (like a low-resolution photo) for the direct path and save massive amounts of computing power.
The "Dust" is Chaotic: The scattered signals (bouncing off walls, tables, and people) do change rapidly. If you move even a tiny bit, the reflections look different.

What This Means for the Future:
This paper gives engineers a "rule of thumb" for building the next generation of wireless networks.

The Analogy: Imagine you are navigating a city.
- Old Way: You had to stop and take a photo of every single brick on every building to know where you were. (Too slow, too expensive).
- New Way (Based on this paper): You realize that as long as you are on the main highway (the direct signal), you can drive 50 meters without needing a new photo. You only need to stop and take a new photo when you turn into a narrow, bumpy alleyway full of obstacles (the scattered signals).

The Takeaway:
We don't need to be as paranoid about tiny movements as we thought. By understanding that the "direct path" is stable, we can design smarter, faster, and less expensive systems for Beyond-6G. We can use a "coarse grid" for the easy parts of the room and a "fine grid" only for the tricky, cluttered corners. This saves computing power and makes the future internet faster and more reliable.

1. Problem Statement

As Beyond-6G systems move toward sub-terahertz (sub-THz, 100–300 GHz) bands to achieve extreme data rates, accurate channel modeling becomes critical.

The Challenge: Theoretically, wireless channels vary over distances as small as one wavelength ( $\lambda$ ). At sub-THz frequencies (e.g., 145 GHz, $\lambda \approx 2$ mm), this implies that ray-tracing simulations would require computational grids with millimeter-scale resolution to capture channel variations accurately.
The Bottleneck: Such fine-grained modeling results in prohibitive computational costs for large-scale network planning.
The Gap: While it is hypothesized that channel characteristics might remain stable over larger distances in practice, there is a lack of experimental data regarding spatial consistency at sub-wavelength scales (specifically < 1 m) in indoor environments. Existing models often rely on assumptions derived from lower frequencies (mmWave) or lack the granularity needed for sub-THz ray tracing.

2. Methodology

The authors conducted a large-scale, high-precision measurement campaign to empirically characterize spatial consistency.

Environment: An indoor-hall (InH) environment at Northeastern University (UNLab), resembling a classroom (20m $\times$ 12m $\times$ 3m) with typical furniture and equipment.
Frequency Band: 140–150 GHz.
Equipment:
- Channel Sounder: A custom spread-spectrum sliding-correlator system developed at Northeastern University.
- Antennas: WR-6.5 VDI horn antennas (21 dBi gain, 13° beamwidth) at both Transmitter (Tx) and Receiver (Rx).
- Bandwidth: 10 GHz (providing a delay resolution of 200 ps).
Experimental Setup:
- Tx: Fixed at 1.5m height, oriented toward the center of the measurement area.
- Rx: Mounted on a motorized rotary table at 1.5m height.
- Trajectory: A cross-shaped path (two orthogonal 1-meter lines) with the Rx moving from the center to the edges.
- Sampling Resolutions: Three distinct scales were tested:
  - Coarse: 10 cm
  - Fine: 1 cm
  - Ultra-fine: 2.5 mm (approx. 1.2 $\lambda$ to 50 $\lambda$ range).
- Angular Sweeps: At each point, the Rx antenna swept azimuth ( $-100^\circ$ to $100^\circ$ ) and elevation ( $-20^\circ$ to $20^\circ$ ).
Data Processing:
- Extraction of Power Delay Profiles (PDP) and Channel Impulse Responses (CIR).
- Metrics Calculated: Root-Mean-Square (RMS) Delay Spread, RMS Angular Spread (Azimuth/Elevation), K-factor (LoS dominance), and Coherence Bandwidth.
- Analysis: Spatial autocorrelation was computed for pairs of measurement points to determine how quickly channel metrics decorrelate as distance increases.

3. Key Contributions

First Experimental Study at Sub-Wavelength Scales: This is the first known open-literature study to experimentally characterize sub-THz spatial consistency at distances smaller than 1 meter (down to 2.5 mm).
Empirical Validation of Stability: The study provides empirical evidence that key channel metrics (delay spread, angular spread, K-factor) remain highly stable over small-scale displacements in Line-of-Sight (LoS) conditions.
Differentiation of LoS vs. NLoS Behavior: The research explicitly separates the spatial consistency of the dominant LoS path from Non-Line-of-Sight (NLoS) multipath components, revealing distinct behaviors for each.
Guidelines for Ray Tracing: The findings offer concrete parameters for optimizing ray-tracing mesh grids, balancing computational efficiency with physical accuracy.

4. Key Results

Channel Stability: In the considered LoS-dominated indoor environment, channel characteristics (Delay Spread, Angular Spread, K-factor) change only slightly over distances of multiple tens of centimeters.
Spatial Correlation Distance:
- LoS Components: Exhibit strong spatial autocorrelation ( $\rho > 0.8$ ) over distances up to 10–50 wavelengths (approx. 2–10 cm at 145 GHz).
- NLoS Components: Show rapid decorrelation ( $\rho \approx 0.45$ ) due to sensitivity to sub-wavelength displacements and environmental scattering.
Metric Values:
- Average RMS Delay Spread: ~22.3 ns (intrinsic channel definition).
- Average RMS Angular Spread: ~7.4° (Azimuth) and ~7.6° (Elevation).
- K-factor: ~18.5 dB, indicating a strong LoS dominance.
Autocorrelation Findings:
- Omni-directional PDPs showed near-perfect correlation ( $\rho \approx 1$ ), masking the underlying angular variations.
- Directional analysis revealed that while the LoS path is stable, the NLoS scattered components vary rapidly.
- The "All Angles" correlation is lower than "LoS Only" correlation because it includes the rapidly varying NLoS components.

5. Significance and Implications

The results have profound implications for the design of Beyond-6G systems and channel modeling:

Optimized Ray Tracing: In indoor LoS scenarios, the mesh grid for ray-tracing simulations does not need to be as fine as one wavelength. A grid spacing of 10–50 wavelengths (approx. 2–10 cm) is sufficient along stable LoS directions, drastically reducing computational load.
Adaptive Meshing: A hybrid approach is recommended: use coarse grids for LoS regions and fine grids (or advanced interpolation) for NLoS/scattered regions where channel characteristics change rapidly.
Beam Tracking & Beamforming: The stability of the LoS path over tens of centimeters suggests that beam tracking algorithms do not need to update at every millimeter, allowing for more robust and less frequent beam alignment updates in indoor environments.
Model Parameterization: The study provides quantitative data to parameterize statistical channel models (like NYUSIM extensions) and hybrid models, ensuring they accurately reflect the spatial consistency of sub-THz channels.

In conclusion, the paper demonstrates that while sub-THz channels are theoretically sensitive to minute movements, practical indoor LoS channels exhibit significant spatial consistency, enabling more efficient and scalable modeling for future 6G networks.

1. Problem Statement

2. Methodology

3. Key Contributions

4. Key Results

5. Significance and Implications

More like this