Frequency-selective beamforming and single-shot beam training with dynamic metasurface antennas
Nitish Vikas Deshpande, Student Member, IEEE, Joseph Carlson, Miguel R. Castellanos, Member, IEEE, and Robert W. Heath Jr., Fellow, IEEE · 2024
Read Frequency-selective beamforming and single-shot beam training with dynamic metasurface antennas by Nitish Vikas Deshpande, Student Member, IEEE, Joseph Carlson, Miguel R. Castellanos, Member, IEEE, and Robert W. Heath Jr., Fellow, IEEE (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-00215.txt.
Introduction
Deshpande and co-authors turn a quirk of dynamic metasurface antennas into an advantage for wireless beamforming, in this 2024 paper. These antennas steer beams with cheap, reconfigurable elements. Each element, though, responds strongly only near its resonant frequency, so beams weaken off that mark.
Instead of fixing one frequency, the team tunes the operating frequency and element settings together. They also design a beam training routine that finds a receiver's direction from a single pilot signal, using a sub-array layout. A modest set of frequency states proved enough for both training and beamforming. The results promise simpler, lower-power beam control for future wireless links. Tests used the direction estimates directly in the beamforming step.
Catalog record
| CREATOR | Nitish Vikas Deshpande, Student Member, IEEE, Joseph Carlson, Miguel R. Castellanos, Member, IEEE, and Robert W. Heath Jr., Fellow, IEEE |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-00215
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Frequency-selective beamforming and single-shot beam training with dynamic metasurface antennas Thanks: Nitish Vikas Deshpande, Joseph Carlson, and Robert W. Heath Jr. are with the Department of Electrical and Computer Engineering, University of California San Diego, La Jolla, CA 92093 USA (e-mail: [email protected]; [email protected]; [email protected]). Miguel R. Castellanos is with the Department of Electrical Engineering and Computer Science at the University of Tennessee, Knoxville, TN 37996 USA (e-mail: [email protected]). This material is based upon work supported by the National Science Foundation under grant nos. NSF-ECCS-2435261, NSF ECCS-2414678, the Army Research Office under Grant W911NF2410107, and the Qualcomm Innovation Fellowship. Nitish Vikas Deshpande, Student Member, IEEE, Joseph Carlson, Miguel R. Castellanos, Member, IEEE, and Robert W. Heath Jr., Fellow, IEEE Abstract Dynamic metasurface antennas (DMAs) beamform through low-powered components that enable reconfiguration of each radiating element. Previous research on a single-user multiple-input-single-output (MISO) system with a dynamic metasurface antenna at the transmitter has focused on maximizing the beamforming gain at a fixed operating frequency. The DMA, however, has a frequency-selective response that leads to magnitude degradation for frequencies away from the resonant frequency of each element. This causes reduction in beamforming gain if the DMA only operates at a fixed frequency. We exploit the frequency reconfigurability of the DMA to dynamically optimize both the operating frequency and the element configuration, maximizing the beamforming gain. We leverage this approach to develop a single-shot beam training procedure using a DMA sub-array architecture that estimates the receiver’s angular direction with a single OFDM pilot signal. We evaluate the beamforming gain performance of the DMA array using the receiver’s angular direction estimate obtained from beam training. Our results show that it is sufficient to use a limited number of resonant frequency states to do both beam training and beamforming instead of using an infinite resolution DMA beamformer. I Introduction While most prior work on DMAs assumes frequency-flat operation, their inherent ability to reconfigure beamforming properties over a wide bandwidth makes them promising candidates for frequency-selective wireless architectures [ 1 , 2 ] . A DMA consists of a waveguide with multiple radiating slots that can be independently tuned to resonate at different frequencies [ 3 , 4 ] . Unlike conventional antenna arrays that require external components for beamforming, DMAs have the innate capability to beamform through low-powered varactor or PIN diodes embedded into the antennas [ 4 ] . Conventional analog components for frequency-selective beamforming such as true time delay (TTD) elements [ 5 ] and phase-shifters [ 6 ] consume more power than varactor or PIN diodes [ 7 ] . The low cost and low power consumption of DMAs make them highly versatile, with potential applications such as large array implementations in gNBs or use in battery-limited UE handsets. DMAs present a unique set of challenges for beamforming compared to a traditional phased-array architecture because of the differences in their underlying operation mechanism. The magnitude and phase response of each radiating slot in the DMA is coupled. The DMA architecture also has an intrinsic phase-advance in the waveguide. Furthermore, the underlying mechanism for beamforming in fact makes the beamforming frequency-selective. The operating bandwidth and extent of frequency selectivity are tightly coupled together in the design [ 2 ] . As a result, even a task like configuring a DMA to produce a directive beam for a line-of-sight channel is deceptively complicated. Prior work on DMA signal processing is limited to optimizing narrowband metrics. The design of beamforming algorithms with DMAs is non-trivial because of the coupled magnitude and phase constraint, also referred to as the Lorentzian constraint in [ 4 ] . One approach to configure DMA beamforming weights is to relax the Lorentzian constraint, solve for the desired phase-shift, and project the obtained weight on the set of feasible weights on the Lorentzian circle [ 8 ] . This relaxation and projection approach was shown to be sub-optimal in [ 9 ] which proposed another solution based on an optimal common phase rotation to each slot on the DMA. The single-user solution was generalized to a multi-user downlink system in [ 10 ] , uplink system in [ 11 ] and [ 12 ] , and hybrid beamforming with the DMA in [ 13 ] and [ 14 ] . The DMA beamforming approaches in [ 8 , 9 , 10 , 11 , 12 , 13 , 14 ] treat the DMA beamforming weights as frequency-flat, i.e., the phase and magnitude response as a function of frequency is not explicitly modeled. Hence, this prior work on DMA is not directly applicable for wideband scenarios where DMA frequency-selectivity and beam squint is significant. Although frequency-reconfigurable antennas have been proposed for wireless systems with multi-band capabilities [ 15 , 16 ] , there is limited prior work on analyzing frequency-selective beamforming with DMAs. In [ 17 ] , a DMA-based MIMO-OFDM system was analyzed by modeling the frequency-selectivity profile of the DMAs. Their approach, however, does not compare frequency-selective DMA beamforming with traditional TTD or phased-array architectures. In this paper, we bridge this gap by proposing a method to configure DMAs to achieve maximum beamforming gain comparable to TTD or phased-array systems. In our paper, we leverage the ability of the DMA to be dynamically reconfigured over a wide frequency range to make the DMA performance comparable to the conventional architectures. We envision a wireless system with DMA that has the ability to be reconfigured to different subcarrier frequencies. In current wireless standards, there exist methods that can dynamically adapt the operating frequency for several applications. For example, in 5G NR, techniques like dynamic spectrum sharing allows the base station to dynamically switch between different frequency bands depending on the network load and user demand [ 18 ] . Adaptive frequency hopping is used in Bluetooth to avoid interference, with the system able to operate at different frequency bands [ 19 ] . Taking motivation from the current standards, our proposed method for DMA beamforming allows the system operator to dynamically choose the DMA operating frequency within a feasible band. This enables configuring the DMA resonant frequencies based on the selected operating frequency. We propose a two-stage optimization framework where the operating frequency is also an optimization variable. The first stage of the algorithm configures the DMA resonant frequencies to maximize the beamforming gain at a specific operating frequency, expressing the optimal resonant frequencies as a closed-form parametric function of the operating frequency. The second stage of the algorithm optimizes over all operating frequencies in the feasible operating band. This approach is suitable for systems with spectrum availability similar to the DMA operating range, allowing for the selection of operating frequencies for communication. Our novel contribution is the second stage of optimizing the operating frequency, which mitigates the impact of magnitude degradation on the DMA beamforming weight. With this frequency-selective strategy, the beamforming gain and hence achievable rate improves compared to prior work [ 4 , 9 , 10 , 12 ] . We further show that by optimal design choice for the DMA waveguide refractive index and inter-element spacing, the maximum beamforming gain for the DMA architecture equals that of a TTD array within the desired angular range. To achieve the desired