Two-Stage Hybrid Transceiver Design Relying on Low-Resolution ADCs in Partially Connected MU Terahertz (THz) MIMO Systems

This paper proposes a two-stage hybrid transceiver design for partially connected multi-user THz MIMO systems that utilizes low-resolution ADCs and a novel beamforming technique with few true time delay lines to mitigate the dual-wideband effect, achieving a 13% improvement in spectral efficiency over existing methods.

The Gist

Here is an explanation of the paper using simple language and creative analogies.

The Big Picture: The "Terahertz Highway" Problem

Imagine you are trying to send a massive amount of data (like streaming 1,000 movies at once) over a new, super-fast highway called the Terahertz (THz) band. This highway is incredibly wide, allowing for huge speeds, but it has two major traffic jams:

1. The "Foggy Lens" Problem (Frequency-Wideband Effect): Because the signal is so wide, different colors of the "light" (frequencies) travel at slightly different speeds. It's like trying to run a race where the runners start at the same time but finish at different times, causing the group to scatter.
2. The "Squinting Eye" Problem (Spatial-Wideband Effect / Beam-Split): This is the main villain. Imagine you have a giant flashlight (your antenna array) trying to shine a single, tight beam at a specific person. In normal conditions, the beam stays straight. But on this THz highway, the beam gets "squinty." The red part of the beam points left, and the blue part points right. Instead of one tight beam hitting your target, the signal splits into a messy fan shape, missing the target and wasting energy.

The Goal: The authors want to build a system that keeps the beam tight and focused on the target, even with these messy effects, while using cheap, low-quality hardware (low-resolution sensors) to save money and power.

---

The Solution: A Two-Stage "Smart Flashlight"

The paper proposes a Two-Stage Hybrid Transceiver. Think of this as a smart flashlight system at a base station (the transmitter) and a receiver.

Stage 1: Finding the Right Direction (The "Map")

First, the system needs to figure out exactly where the users are.

• The Analogy: Imagine you are in a dark room with 16 different flashlights (RF chains). You need to point them at 4 different people in the room.
• The Trick: Instead of trying to calculate the perfect angle for every single color of light (which is computationally impossible), the system uses a "dictionary" of pre-made angles. It picks the best angle from this list for each flashlight.
• The Hardware: To keep costs down, they don't connect every single antenna to every wire. They use a Partially Connected architecture. Think of it like a team of 16 captains, where each captain only talks to a small group of 6 soldiers (antennas), rather than one giant general talking to all 96 soldiers. This is cheaper and easier to build.

Stage 2: Fixing the "Squint" (The "True Time Delay")

This is the paper's biggest innovation. Once they know the direction, they need to stop the beam from splitting.

• The Problem: Traditional flashlights use "Phase Shifters." These are like gears that rotate the signal slightly. But gears work the same way for all colors of light. In a wideband system, this causes the beam to split (the "squint").
• The Solution: The authors introduce True Time Delay (TTD) lines.
• The Analogy: Imagine a marching band.
  • Phase Shifters are like telling everyone to just "step forward" at the same time. If they are far apart, the sound waves get out of sync.
  • True Time Delays are like telling the person at the back of the line to wait a tiny fraction of a second before stepping, and the person in the middle to wait a tiny bit less.
  • By adding these tiny, precise delays, the system ensures that all the different "colors" (frequencies) of the signal arrive at the target at the exact same moment. This forces the split beam to snap back together into a single, tight laser beam.

The "Cheap Sensor" Bonus (Low-Resolution ADCs)

Usually, to get high-quality data, you need expensive, high-precision sensors (like a 16-bit camera sensor). These are power-hungry and expensive.

• The Innovation: The authors show that you can use Low-Resolution ADCs (like a cheap 3-bit camera sensor) and still get great results.
• The Analogy: It's like taking a photo with a low-quality phone camera but using a very smart software filter (called Bussgang Decomposition) to clean up the grainy picture. They prove that even with this "grainy" hardware, the system performs almost as well as the expensive, perfect hardware.

---

The Results: Why Does This Matter?

The authors tested their system and found:

1. Beam Splitting is Fixed: The "squinting" beam is now a straight, tight laser.
2. Better Speed: They achieved about 13% more data speed (spectral efficiency) compared to the current best methods.
3. Cost Effective: They proved you don't need expensive, perfect sensors. A "3-bit" sensor works almost as well as a perfect one, saving huge amounts of power and money.

Summary Metaphor

Imagine you are trying to throw a ball to a friend across a windy field.

• Old Methods: You throw the ball, but the wind (THz effects) blows it sideways, and it splits into a cloud of dust. You miss your friend.
• This Paper's Method: You use a special "wind-correcting" arm (True Time Delays) that adjusts the throw in real-time so the ball stays on a straight line. Plus, you are using a cheap, plastic ball (Low-Resolution ADC) instead of a heavy, expensive leather one, but thanks to your perfect aim, the cheap ball hits the target just as well as the expensive one would have.

In short: They figured out how to build a super-fast, wideband communication system that stays focused, works with cheap hardware, and doesn't waste energy.

Technical Summary

Here is a detailed technical summary of the paper "Two-Stage Hybrid Transceiver Design Relying on Low-Resolution ADCs in Partially Connected MU Terahertz (THz) MIMO Systems."

1. Problem Statement

The paper addresses critical challenges in Terahertz (THz) massive MIMO systems, specifically within the 0.3–10 THz frequency range. While THz offers extreme data rates, it suffers from two detrimental phenomena collectively termed the "dual-wideband effect":

• Frequency-Wideband Effect: Caused by multipath-induced delay spread, leading to frequency-selective fading.
• Spatial-Wideband Effect: Caused by the short wavelength and large antenna arrays, resulting in a progressive phase shift across the array.

The Core Issue: The spatial-wideband effect manifests in the frequency domain as the beam-split effect, where the angle of arrival/departure (AoA/AoD) varies across subcarriers. This causes the beam to split and misalign with the physical target direction, severely degrading spectral efficiency.

Additional Constraints:

• Existing solutions often ignore low-resolution Analog-to-Digital Converters (ADCs), which are essential for reducing power consumption and hardware cost in THz systems.
• Most designs assume fully connected architectures, whereas partially connected architectures are more practical for large-scale THz arrays but exacerbate the beam-split effect due to limited connectivity.
• There is a lack of practical wideband precoding solutions that combine Single-Carrier Frequency Domain Equalization (SC-FDE), partially connected hybrid architectures, and low-resolution ADCs while compensating for dual-wideband effects.

2. Methodology

The authors propose a novel two-stage hybrid transceiver design for a partially connected Multi-User (MU) THz MIMO system.

System Model

• Architecture: The Base Station (BS) uses a partially connected architecture with $N_{RF}$ RF chains, where each chain connects to a subarray of antennas.
• Quantization: Low-resolution ADCs are employed at the receiver. The non-linear quantization is modeled using Bussgang decomposition, which linearizes the system model to approximate capacity.
• Channel: The THz channel is modeled incorporating absorption, reflection, and free-space losses, accounting for both Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS) paths.

Two-Stage Design Framework

Stage 1: Frequency-Independent Beamformer Design

• Goal: Determine optimal beam-steering angles for each RF chain.
• Algorithm: A spatially sparse algorithm (similar to Simultaneous Orthogonal Matching Pursuit - SOMP) is used.
  • It exploits the angular sparsity of the THz channel.
  • It selects optimal beamforming vectors from a dictionary matrix corresponding to each subarray.
  • This stage computes the optimal analog RF combiners and precoders based on the carrier frequency ($f_c$).

Stage 2: Frequency-Dependent Compensation via True Time Delays (TTD)

• Goal: Eliminate the beam-split effect by making the beamforming frequency-dependent.
• Mechanism: The system introduces True Time Delay (TTD) lines. Unlike standard Phase Shifters (PS) which are frequency-invariant, TTDs introduce a time delay that creates a phase shift proportional to frequency.
• Implementation:
  • Each RF chain is connected to $M$ TTD elements, which are further connected to groups of phase shifters.
  • The TTD elements convert the traditional frequency-invariant phase shifters into frequency-dependent equivalents.
  • By aligning the optimal beamforming direction with the physical ray direction using time delays, the beam-split effect is substantially mitigated.
  • The paper derives specific mathematical expressions to calculate the required time delays ($t_{sl}$) to align the array gain across all subcarriers.

3. Key Contributions

1. Novel System Model: The paper is the first (to the authors' knowledge) to propose a SC-FDE-based system with low-resolution ADCs in a partially connected THz MU-MIMO architecture, specifically addressing the dual-wideband effect.
2. Bussgang Decomposition Integration: The framework utilizes Bussgang decomposition to linearize the non-linear quantized system model, enabling the design of optimal precoders/combiners that approach channel capacity despite low-resolution hardware.
3. Two-Stage Hybrid Beamforming:
   • Stage 1: Uses a spatially sparse algorithm to select optimal beam-steering angles from a dictionary.
   • Stage 2: Employs a minimal number of True Time Delay (TTD) lines to generate frequency-dependent beamformers, effectively canceling the beam-split effect.
4. Pulse Shaping Analysis: The study compares Root Raised Cosine (RRC) and Rectangular (Rect) pulse shaping filters, revealing a trade-off between spectral leakage (Rect) and edge attenuation (RRC) in dual-wideband channels.

4. Simulation Results

The proposed framework was evaluated against state-of-the-art techniques (DPP, SOMP without TTD, Interference Cancellation) under various conditions:

• Beam-Split Mitigation: Figure 2(a) demonstrates that while traditional precoding causes beams to split away from the target direction across subcarriers, the proposed TTD-based technique aligns the beams perfectly with the physical target direction for all subcarriers.
• Spectral Efficiency (SE):
  • The proposed method achieves a ~13% improvement in spectral efficiency compared to existing state-of-the-art techniques.
  • It performs significantly closer to the theoretical capacity limit.
• ADC Resolution:
  • Simulations show that a 3-bit ADC yields performance nearly identical to an ideal infinite-bit quantizer, with only a 2% loss at high SNR. This validates the feasibility of using low-resolution hardware in THz systems.
• Pulse Shaping: The results highlight the trade-offs between RRC-PSF and Rect-PSF, showing that the proposed method outperforms others in both scenarios.

5. Significance and Impact

• Hardware Efficiency: By proving that high performance can be achieved with low-resolution ADCs and a partially connected architecture, the paper offers a practical pathway to reduce the power consumption and cost of THz transceivers, which are otherwise prohibitive due to high-frequency hardware requirements.
• Solving the Beam-Split Effect: The introduction of a TTD-based two-stage design provides a robust solution to the beam-split effect, a major bottleneck in wideband THz communications.
• Practical Implementation: The use of SC-FDE avoids the need for IFFT at the transmitter, simplifying the transmitter architecture. The proposed method requires only a few TTD lines, making it feasible for large-scale antenna arrays.
• Future Direction: This work lays the foundation for future research into hybrid beamforming designs that incorporate realistic channel estimation and more complex hardware constraints in 6G THz networks.