TL;DR
This paper proposes a "green" shift for 6G by replacing power-hungry phased arrays with Analog Multi-Beam Antenna Arrays (MBAAs) for Joint Communications and Sensing (JCAS). By leveraging fixed analog structures like lenses or Butler matrices, the authors demonstrate a framework that achieves high-resolution sensing and flexible communication with significantly lower energy consumption and simplified signal processing.
What Problem Does This Paper Address?
As we move toward 6G, Joint Communications and Sensing (JCAS)—the ability for a single system to both transmit data and "see" its environment—is becoming a cornerstone technology. However, current JCAS designs typically rely on massive MIMO arrays with thousands of tunable phase shifters.
These components are "power hogs" and introduce significant hardware complexity. Furthermore, the optimization math required to steer these beams is computationally expensive. The authors argue that for JCAS to be sustainable (or "green"), we need a hardware architecture that provides the benefits of massive MIMO without the massive energy bill.
Methods
The authors advocate for Multi-Beam Antenna Arrays (MBAAs). Unlike traditional phased arrays that use active electronics to steer a single beam, MBAAs use passive, pre-designed analog devices:
- Lens Antenna Arrays: Use quasi-optical lenses to focus waves onto specific focal points.
- Butler Matrices: Use a network of fixed couplers and phase shifters to perform a hardware-based Discrete Fourier Transform (DFT).
Key Technical Innovations: 1. Flexible Beam Synthesis: Instead of complex optimization, they use a simple "beam selection" network. By combining adjacent fixed beams with 1-bit phase shifters (0 or $\pi$), they can synthesize wider or specifically shaped beams for sensing. 2. Base-2 Wide Beams: A specific method of combining two adjacent beams to create a wider mainlobe. This improves beamforming gain by 6 dB at the "dip" between beams and enables a low-complexity Angle-of-Arrival (AoA) estimation. 3. Exploiting Beam Squint: In wideband systems, the beam direction shifts slightly with frequency (usually considered a "curse"). The authors propose using this frequency-dependent shift as a "spatial-frequency signature" to quickly locate targets without scanning.
Datasets
Not discussed in the paper. The authors use numerical simulations and theoretical modeling to validate their framework rather than empirical datasets.
Key Insights and Ideas
- Negligible Power Consumption: The primary "aha" moment is the hardware comparison. A standard phased array requires $N \times M$ multi-bit phase shifters (where $N$ is antennas and $M$ is RF chains). MBAAs replace these with passive structures and simple switches, drastically reducing the power floor.
- The "Blessing" of Beam Squint: While most engineers try to eliminate beam squint, this paper suggests it can be used for "fast localization." Because different frequencies point to slightly different angles, a single wideband transmission can "cover" an angular region, and the received power across frequencies tells you exactly where the target is.
- Type-1 vs. Type-2 JCAS Beams:
- Type-1 (Regular): Good for systematic scanning and high-gain communication.
- Type-2 (Random): Randomly selects sensing beams in each time slot. This "scrambles" the signal in non-user directions, providing built-in Physical Layer Secrecy against eavesdroppers.
Key Results
- Power Efficiency: The authors calculate that for a 64-antenna array with 8 RF chains, an MBAA-based system can save approximately 10,240 mW (10.24 Watts) compared to a traditional phase-shifter-based massive MIMO array.
- Gain Improvement: Using the "base-2 wide beam" synthesis, the system achieves a 6 dB improvement in beamforming gain at the intersection of two beams compared to using a single DFT beam.
- AoA Accuracy: The proposed ratio-based estimation (using the difference and sum of two synthesized beams) allows for monotonic, high-accuracy angle estimation that is computationally lighter than traditional super-resolution algorithms.
Limitations and Future Work
- Discrete States: MBAAs have a finite number of fixed beam directions. While synthesis helps, they lack the infinite "steering resolution" of fully digital arrays.
- Beam Selection Complexity: Selecting the optimal subset of beams for multiple users and sensing targets is an NP-hard problem. The authors note that more efficient, sub-optimal search algorithms are needed.
- Waveform Design: Most existing JCAS waveforms are designed for digital arrays. New research is required to design waveforms that specifically exploit the spatial-frequency patterns unique to MBAAs.
- Hardware Constraints: The physical size of lenses or Butler matrices can be a challenge for integration into small mobile devices, making this more suitable for base stations or large vehicles.
Summary updated