⚡ electrical engineering

Synthesized-Isotropic Narrowband Channel Parameter Extraction from Angle-Resolved Wideband Channel Measurements

This paper addresses the bias in estimating antenna-independent channel parameters from angle-resolved wideband measurements by proposing a unified matrix framework with a beam-accumulation correction factor to accurately synthesize isotropic narrowband power, which is validated through simulations and 154 GHz corridor measurements.

Original authors: Minseok Kim, Masato Yomoda

Published 2026-06-23

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Minseok Kim, Masato Yomoda

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 measure the total amount of sunlight hitting a specific spot on the ground.

The Problem: The "Flashlight" Dilemma
In the world of high-speed wireless internet (like 5G, 6G, and beyond), engineers need to measure how much signal power travels from a transmitter to a receiver. Ideally, they would use a "magic antenna" that listens to signals coming from every direction at once, like a perfect sphere. This is called an isotropic measurement.

However, at the super-high frequencies used for these new networks (millimeter-wave and terahertz), signals are very weak and easily lost. Using a "magic sphere" antenna isn't practical because it's not strong enough to catch the signal. Instead, engineers use high-gain antennas that act like powerful, narrow flashlights. They have to physically rotate these flashlights, pointing them in different directions (scanning) to catch the signal from every angle.

Here is the catch: The flashlights overlap.

If you shine a flashlight at a wall, then move it slightly to the right and shine it again, the two beams overlap. If you simply add up the brightness of the first spot and the second spot, you are counting the overlapping area twice. This leads to an overestimate of the total light. Conversely, if you move the flashlight too far apart, you might miss the light in the gaps between the beams, leading to an underestimate.

This paper tackles the math behind fixing this "double-counting" and "missing-spot" problem to get an accurate total power reading from these directional scans.

The Solution: The "Correction Factor"
The authors propose a clever mathematical trick to fix this. Think of it like a recipe adjustment.

The Map: They create a digital map of exactly how the flashlight beam looks (its shape and how it fades at the edges).
The Grid: They look at the specific steps they took while rotating the flashlight (the "scan grid").
The Calculation: They calculate a "Beam-Accumulation Correction Factor."
- Imagine the flashlight beam is a pile of sand. If you scoop the sand with a bucket that overlaps with the previous scoop, you've scooped some sand twice. The correction factor tells you exactly how much to subtract to get the true amount of sand.
- If you scoop the sand but leave a tiny gap between scoops, the factor tells you how much to add to account for the missing sand.

The "Scalloping" Problem
There is a second, sneaky issue called scalloping. Imagine your flashlight beam is a perfect circle, but the grid you are measuring on is made of squares. If the signal comes from a spot that lands exactly in the middle of a square, you get a perfect reading. But if the signal comes from a spot slightly off-center (between the squares), your flashlight might not catch the full peak of the signal, making it look weaker than it really is.

The paper introduces a new "Offset-Averaged" method. Instead of guessing exactly where the signal is, they mathematically average out all the possible "off-center" positions. It's like saying, "We don't know exactly where the signal is within this square, so let's calculate the average brightness for any position inside that square." This smooths out the errors without needing to know the signal's exact location.

How They Proved It
The authors didn't just do math on paper; they tested it in two ways:

Computer Simulations: They created a fake world with known signal paths and showed that their method could recover the true power almost perfectly, whereas the old "just add them up" method was wrong.
Real-World Tests: They went into a corridor and used real antennas at a frequency of 154 GHz (very high speed). They compared their "flashlight scan" results against a "magic sphere" reference measurement. The results showed that their correction method made the flashlight scan match the reference almost perfectly.

The Bottom Line
This paper provides a simple, reliable "calculator" for engineers. When they use directional antennas to measure wireless channels, they can now easily correct for the fact that their beams overlap and that signals might fall between the grid lines. This allows them to accurately calculate Path Loss (how much signal is lost over distance) without needing expensive, omnidirectional equipment that doesn't work well at these high frequencies.

Technical Summary: Synthesized-Isotropic Narrowband Channel Parameter Extraction from Angle-Resolved Wideband Channel Measurements

Problem Statement
In millimeter-wave and terahertz communications, channel sounding often relies on angle-resolved measurements using high-gain antennas or arrays to overcome limited link budgets. However, extracting fundamental large-scale parameters—specifically path loss (PL) and path gain (PG)—requires converting these directional measurements into isotropic-equivalent values. A significant technical challenge arises because practical scan beams are non-orthogonal and exhibit substantial overlap. Naive power summation across scan angles leads to biased estimates: coarse angular scanning intervals (ASI) cause "scalloping loss" (underestimation) for off-grid angles, while fine scanning intervals cause double-counting of correlated contributions from the same path (overestimation). Existing parametric methods (e.g., SAGE, RIMAX) rely on complex high-resolution parameter extraction which may vary based on modeling assumptions, while non-parametric approaches often remain biased due to beam overlap and sidelobe leakage. There remains a practical gap in obtaining accurate isotropic-equivalent power directly from discretely sampled, non-orthogonal angular beams without resorting to elaborate inversion procedures.

Methodology
The paper proposes a unified framework to synthesize isotropic narrowband power from angle-resolved wideband measurements. The core methodology involves:

Matrix Formulation: The authors formulate the channel power vector in a unified matrix form. By assuming uncorrelated scattering, the total received power is approximated as the incoherent sum of multipath component (MPC) powers. The measured power vector is expressed as the product of a basis matrix (containing squared antenna patterns) and the discrete power distribution.
Beam-Accumulation Correction: Instead of explicitly inverting the large basis matrix to recover path powers, the authors exploit the rotational invariance of the scan. They derive a scalar beam-accumulation correction factor ( $\zeta$ ), defined as the sum of the squared antenna pattern samples over the discrete scan grid. The synthesized isotropic power is estimated by summing the measured power samples and dividing by this factor ( $\zeta$ ).
Scalloping Mitigation: To address the bias introduced when MPCs arrive at angles between grid points (off-grid), the paper introduces an offset-averaged variant of the correction factor. Rather than estimating the specific off-grid angle, this method averages the accumulation factor over a uniformly distributed within-bin offset, providing a robust correction without explicit angle estimation.
Factorization: The total correction factor is factorized into contributions from the delay domain and the angular domains (Transmitter and Receiver, Co-elevation and Azimuth). For delay, the factor depends on the sounding signal's autocorrelation; for angles, it depends on the squared antenna radiation patterns.
Validation: The framework is validated using two approaches:
- Simulation: Using a Saleh-Valenzuela clustered multipath model with varying HPBWs and angular spreads to quantify estimation error against known ground truth.
- Measurement: Conducting double-directional (Horn-to-Horn) angle-scanning measurements in a 154 GHz corridor and comparing the synthesized PL against a reference measurement using an omnidirectional transmitter (Omnidirectional-to-Horn).

Key Contributions

Unified Matrix Formulation: The paper presents a mathematical formulation for isotropic-equivalent power synthesis that explicitly accounts for the non-orthogonality and overlap of practical scan beams.
Beam-Accumulation Correction Factor: It introduces a computationally efficient correction factor derived from the discrete summation of squared antenna patterns. This factor compensates for beam overlap effects directly from the scan grid and antenna pattern data.
Offset-Averaged Variant: A specific contribution is the development of an offset-averaged correction factor that mitigates scalloping loss caused by off-grid angles, enhancing robustness without requiring complex off-grid angle estimation.
Practical Implementation: The method is designed to be implementation-friendly, avoiding the need for high-resolution parametric extraction or elaborate inversion procedures, making it suitable for practical directional-scanning campaigns.

Results

Simulation: Monte-Carlo simulations demonstrated that the proposed correction significantly reduces estimation bias compared to naive summation. The use of the offset-averaged factor further reduced residual errors, particularly in wide-beam scenarios and Co-elevation scanning where scalloping effects are pronounced. The error distribution centered close to 0 dB relative to the true incoherent power sum.
Measurements: In 154 GHz corridor measurements, the synthesized path loss from double-directional horn-to-horn scans showed strong agreement with the reference omnidirectional-to-horn measurements. The cumulative distribution function (CDF) of the path loss difference was centered near 0 dB, confirming that the proposed compensation effectively mitigates the dominant biases caused by beam overlap and scalloping.

Significance and Claims
The paper claims to address a practical gap in channel sounding by providing a straightforward, robust method to recover antenna-independent path loss from directional measurements. The significance lies in its ability to correct for inherent biases in non-orthogonal beam scanning without relying on complex parametric models or additional omnidirectional measurements for every scenario. The authors emphasize that while delay-domain integration is standard, their work provides the necessary correction for the angular domain, enabling accurate large-scale parameter extraction from widely used angle-scanning setups. The method is presented as a practical tool for researchers and engineers conducting millimeter-wave and terahertz channel measurements, offering improved accuracy in path gain estimation while maintaining ease of implementation.

More like this