⚡ electrical engineering

Joint Segment Activation and Antenna Placement for Uplink SWAN Systems

This paper analyzes the achievable sum-rate of multiuser uplink segmented waveguide-enabled pinching-antenna systems (SWANs), establishes the existence of an optimal segment activation level, and proposes hybrid segment selection and aggregation (HSS/A) schemes with low-complexity greedy algorithms that outperform conventional full-segment aggregation.

Original authors: Songnan Gu, Zhenqiao Cheng, Hao Jiang, Chongjun Ouyang, Yuanwei Liu, Arumugam Nallanathan

Published 2026-05-14

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Songnan Gu, Zhenqiao Cheng, Hao Jiang, Chongjun Ouyang, Yuanwei Liu, Arumugam Nallanathan

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 listen to a group of friends (the users) talking to you (the base station) in a very noisy, echoey hallway. To help you hear them better, you've installed a special, low-loss "sound tube" (a dielectric waveguide) running along the ceiling. Attached to this tube are many small, flexible microphones (pinching antennas) that can pick up the voices and send them down the tube to your ears.

This setup is called a Pinching-Antenna System (PASS). The idea is that by placing these microphones close to the speakers, you avoid the signal getting lost in the air.

The Problem: Too Many Microphones Can Be a Bad Thing

In the old way of doing this, engineers thought: "The more microphones we turn on, the louder and clearer the signal will be!" So, they turned on every single microphone along the tube.

However, the authors of this paper discovered a surprising twist: Turning on every single microphone actually makes the signal worse after a certain point.

Here is why, using a simple analogy:
Imagine each microphone is a person shouting a message down a long hallway.

The Good: The people standing right next to the speakers shout clearly. Their voices add up to make a loud, clear message.
The Bad: The people standing far away at the end of the hallway are shouting too. Their voices are very faint and distorted by the time they reach you.
The Noise: Every time you add a new microphone, you also add a tiny bit of "hiss" (static noise) from the equipment itself.

If you turn on too many microphones, the faint, distant voices don't help much, but the "hiss" from all those extra microphones adds up. Eventually, the static noise drowns out the helpful signal. It's like trying to listen to a conversation in a room where you keep adding more and more people who are whispering incoherently; the room just gets louder with noise, not clearer with speech.

The Solution: The "Smart Selection" Strategy

The paper proposes a new system called SWAN (Segmented Waveguide-enabled Antenna Network). Instead of using one giant tube with everyone shouting, they divide the tube into separate, independent short sections.

In each section, they only turn on one microphone. This stops the microphones from interfering with each other.

But the real breakthrough is their Hybrid Segment Selection and Aggregation (HSS/A) strategy. Instead of blindly turning on all sections, they use a "smart greedy" algorithm to figure out exactly how many sections to use.

Think of it like a Tasting Panel:

You have 100 different ingredients (segments).
You don't just dump all 100 into a pot.
You taste them one by one. You add the first few, and the flavor gets better. You add a few more, and it gets even better.
But then, you add one more, and it starts tasting salty and ruined.
The smart strategy says: "Stop right there! We found the perfect number of ingredients. Let's use that specific group and ignore the rest."

How They Did It

The authors did two main things:

The Math Proof: They used math to prove that there is indeed a "sweet spot." They showed that while adding more microphones gives you a little bit more signal (like logarithmic growth), the noise grows faster (linearly). Therefore, there is a mathematically perfect number of microphones to use, and using all of them is usually a mistake.
The Algorithm: They created a simple, fast computer recipe (a greedy algorithm) that tests different combinations. It starts with no microphones, adds the best one, checks the result, adds the next best one, and keeps going until adding another one starts to hurt the performance. It then picks the "best" group it found during the test.

They tested two versions:

Type-I: Just turning the right microphones on and off.
Type-II: Turning the right microphones on and adjusting their timing (phase) so their voices arrive at your ear perfectly in sync, like a choir singing in harmony.

The Result

Their simulations showed that this "smart selection" method works much better than the old "turn everything on" method.

In some cases, using fewer microphones actually gave a clearer signal than using all of them.
The "Type-II" version (with timing adjustments) was even better, but the main win came from simply knowing when to stop adding microphones.

In short: The paper teaches us that in high-tech wireless systems, "more" isn't always "better." Sometimes, the best way to get a clear signal is to be picky and only use the best few helpers, rather than inviting the whole noisy crowd.

Technical Summary: Joint Segment Activation and Antenna Placement for Uplink SWAN Systems

Problem Statement
The paper addresses the uplink performance of Segmented Waveguide-enabled Pinching-Antenna Systems (SWANs). While Pinching-Antenna Systems (PASS) offer advantages in high-frequency communications by mitigating path loss and blockage via dielectric waveguides, conventional uplink PASS architectures suffer from Inter-Antenna Radiation (IAR). In standard PASS, multiple pinching antennas (PAs) on the same waveguide cause signal re-radiation and strong coupling, complicating receiver design. SWANs were proposed to eliminate IAR by dividing the waveguide into independent segments, each connected to the base station (BS) via a separate feed point, allowing at most one active PA per segment.

However, a fundamental design challenge remains: the relationship between the number of activated segments and the achievable sum-rate is non-monotonic. Intuitively, activating more segments should increase array gain; however, numerical observations suggest that activating too many segments can degrade performance. The paper identifies that this phenomenon has not been theoretically characterized, nor has a transmission strategy been developed to explicitly exploit this tradeoff.

Methodology
The authors analyze a multiuser uplink SWAN system where a BS serves $K$ users distributed in a rectangular area. The system model considers two segment aggregation (SA) architectures:

Type-I SA: All segment feed points connect to a single RF chain via an analog combiner.
Type-II SA: Each segment includes an analog phase shifter before RF combining to enable coherent signal alignment.

The core methodology involves three steps:

Theoretical Analysis: The authors derive an upper bound on the per-user channel gain under ideal coherent combining. By approximating the summation of channel contributions using midpoint Riemann integrals, they analyze the asymptotic behavior of the sum-rate as the number of segments ( $M$ ) approaches infinity.
Optimization Framework: Motivated by the theoretical finding that full-segment activation is suboptimal, the paper proposes a Hybrid Segment Selection and Aggregation (HSS/A) architecture. In this framework, only a subset of segments is activated to maximize the sum-rate.
Algorithm Development: To solve the joint optimization of segment activation (discrete) and PA placement (continuous), the authors develop low-complexity greedy algorithms for both Type-I and Type-II HSS/A.
- The algorithms iteratively add segments one by one.
- For each candidate segment, the optimal PA position is found via one-dimensional search.
- For Type-II, phase shifts are optimized using an element-wise alternating optimization approach.
- Crucially, unlike standard greedy methods that stop when performance drops, these algorithms evaluate all activation levels generated during the search and select the configuration yielding the maximum sum-rate.

Key Contributions

Theoretical Proof of Optimal Activation Level: The paper theoretically establishes the existence of an optimal segment activation level. It proves that while nearby segments provide significant coherent combining gain, distant segments contribute weak signal components that are outweighed by the linearly accumulated receiver noise. Consequently, the achievable sum-rate approaches zero as the number of activated segments tends to infinity.
HSS/A Architecture: The authors propose a hybrid architecture that dynamically selects a subset of segments for activation, explicitly avoiding the performance degradation associated with full-segment aggregation.
Low-Complexity Greedy Algorithms: Practical algorithms are developed to jointly optimize segment activation and PA placement (and phase shifts for Type-II) with computational complexity scaling as $O(M^2KQ)$ for Type-I and $O(M^2KQ + I_\theta M^3)$ for Type-II, where $Q$ is the grid resolution and $I_\theta$ is the number of phase optimization iterations.

Numerical Results
Simulations validate the theoretical analysis and the proposed schemes:

Rate Bound Verification: The derived upper bound closely matches exact summation results, confirming the non-monotonic trend where the sum-rate initially increases with the number of segments but eventually decreases.
Performance Gain: The proposed Greedy HSS/A schemes consistently outperform conventional full-segment aggregation (Type-I and Type-II SA) across various transmit power levels and user counts.
Type-II vs. Type-I: Type-II HSS/A achieves higher sum-rates than Type-I due to analog phase alignment. However, the improvement is moderate because a single phase shift per segment must serve all users, preventing simultaneous perfect alignment for every user.
Robustness: The performance advantage of HSS/A persists as the number of users increases, confirming the strategy's efficacy in multiuser NOMA scenarios.

Significance
The paper claims that its primary significance lies in filling the theoretical gap regarding the non-monotonic behavior of SWAN sum-rates. By proving that full-segment activation is generally suboptimal due to the tradeoff between coherent signal aggregation and accumulated receiver noise, the work provides a fundamental design insight for large-scale SWAN systems. The proposed HSS/A schemes offer a practical, low-complexity solution to exploit this insight, demonstrating that selective activation yields superior performance compared to activating all available segments.

The Problem: Too Many Microphones Can Be a Bad Thing

The Solution: The "Smart Selection" Strategy

How They Did It

The Result

More like this