The SpectrumX research center is focused on the development of next generation tools and techniques for understanding, managing, and dynamically utilizing the radio spectrum. Efforts focus on enabling improved spectrum awareness, approaches to enable the coexistence of different users and applications, and the education of a next generation workforce. As part of Center research efforts our team has developed platforms that provide observational, data management, and visualization capabilities for the radio spectrum. Using these tools we have executed experiments relevant to radio frequency interference (RFI) measurement, modeling, and coordination.

Radio frequency spectrum awareness requires the ability to detect, localize, and characterize emitters in dense and contested wireless environments. In this work, we propose a task-oriented distributed compression framework for joint multi-emitter localization and characterization using spatially distributed receivers. Each receiver observes a short window of complex IQ samples, converts the observation to a time–frequency representation, and encodes it into a compact latent vector. A central fusion decoder combines the receiver latents to estimate an unordered set of active emitters, including their locations, center-frequency offsets, occupied bandwidths, and waveform families. A permutation-invariant training objective is used to handle the arbitrary ordering of emitters and predictions. Experiments on synthetic multi-emitter scenes with spectral overlap show that even extremely compact receiver-side representations can preserve useful information for emitter counting and waveform-family estimation. However, accurate localization and spectral-parameter regression require larger latent dimensions. Increasing the receiver latent dimension from d_rx=1 to d_rx=16 provides the largest improvement, while further increasing to d_rx=64 gives smaller gains. These results demonstrate the potential of learned task-oriented compression for communication-efficient distributed spectrum awareness.

The radio frequency spectrum is an increasingly scarce resource, emphasizing the importance of efficient utilization. Although spectrum efficiency has been extensively studied in communication systems, its definition and evaluation in radar systems remain inadequate. To bridge this gap, this paper introduces a novel approach for defining spectrum efficiency in radar systems based upon the Cramer-Rao Lower Bound (CRLB). Analogous to the Shannon capacity theorem in communication systems, the CRLB serves as a fundamental limit, enabling the determination of the idealized reference system. Spectrum efficiency for a radar system is then defined by comparing its performance to that of the idealized reference system. We illustrate the proposed approach with the example of a Multiple-Input Multiple-Output (MIMO) radar system for target localization through a combination of analysis and numerical evaluation.

Spectrum policy documents, such as notices and comments in Federal Communications Commission (FCC) and the National Telecommunications and Information Administration (NTIA) proceedings, are often lengthy and complex, making them difficult to access and understand. Combining large language models (LLMs) with retrieval-augmented generation (RAG) techniques offers a promising way to address these challenges by improving information retrieval and processing. This paper evaluates the use of RAG in spectrum policy analysis. We establish an open-source knowledge dataset of policy comments that has been cleaned and annotated to work effectively with RAG systems. We also develop a corresponding test dataset of question-answer pairs to evaluate system performance. By benchmarking four RAG-based systems, we show that RAG techniques significantly enhance the ability of LLMs to interpret and respond to complex spectrum policy queries, demonstrating their potential to make these important documents more accessible.

Organizations such as NOAA and NASA develop and operate satellite radiometers to obtain various meteorological data for weather and climate forecasting. For example, to estimate water vapor concentrations in the atmosphere, weak signals resulting from natural radiation and molecular resonances in the 23.6-24 GHz band require the radiometer to be extremely sensitive, potentially making it vulnerable to interference from other radio emissions. In particular, the recent ITU regulation and 3GPP standardization of 5G cellular systems in the adjacent 24 GHz band (24.25-27.5 GHz) has raised concerns that comparatively much stronger 5G signals may cause harmful interference to the radiometer measurements. Some reports suggest that such interference could set the quality of meteorological predictions back by 3–4 decades. On the other hand, limiting the power of the 5G signals too much to avoid interference could reduce the utility of the 24 GHz band for mobile telecommunications.

We consider a continuous-time complex-valued (approximately) bandlimited additive white Gaussian noise channel with a fixed unknown carrier frequency offset. The frequency offset is assumed to lie within plus or minus of a maximum value around the origin. The transmitter and receiver are unaware of the exact value of the frequency offset but are aware of this maximum value. The lack of knowledge of the precise channel probability law, while knowing the family of transition probability laws to which it belongs, gives rise to a compound channel model. An expression for the capacity of this compound channel is derived and shown to be independent of the exact value of the frequency offset. We provide a proof of the achievability part of the coding theorem.

We present single-layer, direct digitally modulated, reconfigurable intelligent surface (RIS) unit-cell (UC) designs at 12GHz. These UCs find application in low-power, gigabits per second (Gbps) ON-OFF Keying (OOK) modulators where each UC of the surface can be individually programmed using Gbps general purpose input-output (GPIO) lines. The key challenge is to realize a large shift in response from a small voltage change (the direct digital drive). We propose that transmission efficiency $\eta_{T}$ should be maximized and modulation depth $m$ should be greater than 20 dB.

In 2015, the FCC established the Citizens Broadband Radio Service (CBRS) for sharing the 3.5 GHz Band (3550-3700 MHz) among federal and non-federal users in the United States. This rulemaking created an experiment in a novel three-tier rights structure: strong protections for incumbents, including government radar systems; Priority Access Licenses (PALs) granting exclusive rights to high bidders in an FCC auction, in part of the band and subject to avoiding interference with incumbents; and Generalized Authorized Access (GAA) for unlicensed users, subject to avoiding interference with both PALs and incumbents. The first commercial deployments in this band were approved in 2019 for GAA devices, and an auction of PALs completed in 2020 generated $4.5 billion in revenues.It is now timely to evaluate this experiment and glean lessons for applications to other spectrum bands, such as the neighboring 3.1-3.45 GHz band or portions of the upper mid-band spectrum 7-24 GHz. In fact, a number of perspectives on CBRS have been recently published. In this paper we review these developments and suggest related policy questions that should be considered when evaluating the use of CBRS-style allocation rules in future bands. The CBRS policy involves several different innovations, seeking to accomplish multiple objectives. One can evaluate this approach from a technical point of view, as an experiment to show that dynamic sharing can provide multiple tiers of commercial access to a band of spectrum while protecting incumbent users. The approach involves coordinated access via a cloud-based Spectrum Access System (SAS) and an Environmental Sensing Capability (ESC) to monitor incumbent users of the band, with requirements standardized through the Wireless Innovation Forum (WInnForum) and implementations certified by the FCC. From an economic and policy point of view, this type of dynamic sharing is asserted to reduce the costs and delays involved in making additional spectrum available for commercial use, as it seeks to avoid relocating incumbents. Of course, costs and benefits should be observed, not simply assumed, and the process undertaken should be compared to those associated with the relevant policy alternatives. CBRS adopted the PAL and GAA tiers for commercial access in an attempt to provide spectrum that could not only support deployments by traditional wireless providers, but also enable new uses of the spectrum by non-traditional entities. How well the spectrum can support these uses, and whether this type of approach leads to an economically efficient mix of uses, provides another economic and policy lens through which this system can be evaluated. This paper explores the technical implementation of the CBRS spectrum sharing approach, and then attempts to appraise the economic welfare results of the novel allocation policy.

This article presents a 39-GHz 800-Mb/s antenna-coupled ON–OFF-key (OOK) receiver with a baseband output capable of driving a 50- $\Omega $ load. The antenna-coupled receiver demonstrates a bit error rate (BER) of $10^{-3}$ over a range of 18 cm while dissipating only 0.71 mW for a record energy efficiency per distance metric of 0.049 pJ/bit/cm. A wireline version of the receiver achieves a record sensitivity level of −36 dBm without preamplification while dissipating only 1.15 mW, resulting in an energy efficiency of 1.44 pJ/bit at a BER of $10^{-3}$ . Including 11.5 dB of RF gain prior to the wireline receiver, 800-Mb/s communication is demonstrated at 3.67 m for a BER of $10^{-5}$ . The low power consumption and long range make this receiver suitable for scaling to hundreds or thousands of elements in massive multi-in–multi-output (MIMO) arrays for next-generation millimeter-wave wireless communications systems.

Spectrum monitoring could improve spectral management and efficiency by enabling spectrum sharing, strengthening policy enforcement, and facilitating data-driven modeling of RF environments. This paper considers spectrum monitoring sensor networks in three dimensions to extend previous work on two-dimensional models. We derive a closed-form expression for the probability of emitter detection, which acts as a metric for system design. We find optimal antenna half-power beamwidths for emitter detection. Additionally, we find that for a path loss exponent less than approximately 3.22, directional antennas are optimal. Otherwise, omnidirectional antennas are optimal. Further, despite potential sub-optimality, omnidirectional antennas are found to be robust in changing environments. Finally, a survey of antenna gain models is evaluated against real antennas to validate the choice of gain model in the paper.

We provide a mutual information lower bound that can be used to analyze the effect of training in models with unknown parameters. For large-scale systems, we show that this bound can be calculated using the difference between two derivatives of a conditional entropy function. We provide a step-by-step process for computing the bound, and apply the steps to a quantized large-scale multiple-antenna wireless communication system with an unknown channel. Numerical results demonstrate the interplay between quantization and training.

Low-resolution transceivers are being considered for millimeter-wave and higher frequency communications because of their simplicity and low power consumption. However, the non-linearities introduced by low-resolution digital-to-analog converters at the transmitters can cause significant out-of-band emissions since traditional bandwidth-limited pulse-shaping is not generally available. We model the performance of a low-resolution transmitter in terms of its spectral efficiency under out-of-band emission constraints. We show that the spectral efficiency can increase linearly with the symbol rate while satisfying out-of-band constraint. This implies that in order to achieve a given spectral efficiency under the bandwidth constraint, the symbol rate of the transmitter should be larger than a threshold. We derive an upper bound on this threshold.

Widespread RF spectrum monitoring could enable data-driven modeling of spectrum usage, enhance spectral utilization, and help automate policy enforcement. Previous works in wireless sensor networks offer design insights for RF sensors, but they assume emitters that radiate omnidirectionally. This paper develops a new framework for directional sensors and emitters, which are increasingly common with the growth of millimeter wave technologies. We focus on two-dimensional random sensor deployments modeled as Poisson point processes. Specifically, we determine the probability that a sensor network detects a single emitter for a channel model including path loss, fading, and the directivity of emitters and sensors with random orientations and locations. Our results suggest that with a path loss exponent of 4, quartering the emitter half-power beamwidth doubles the required average sensor density. We also conclude that omnidirectional sensors optimize detection probability. For multiple emitters, we develop a lower bound on the probability of multi-emitter detection and find the average number of undetected emitters. Finally, assuming higher sensor quality results in higher sensor cost, we consider a fixed-budget deployment and observe that decreasing the individual sensor cost by a decade and therefore increasing the quantity of sensors reduces the missed detection probability by about a decade.

In the past few years, unmanned aerial vehicles (UAVs) have drastically increased in popularity both from consumer and industry perspectives. A key component towards enabling the widespread usage of UAVs is the ability to stay in near-constant communication with the drone for command and control and conveying relevant instrumentation. The usage of cellular technology, namely LTE, seems to be a natural fit for addressing coverage and Line of Sight (LoS) issues. However, there is a relative dearth of data, specifically open source data that explores key performance aspects of cellular at altitudes typically envisioned for commercial UAV operation. The key contribution of this paper is to analyze data taken from numerous drone flights that include varying altitudes, locations, and multiple cellular carriers as recorded in a medium-sized Midwestern city. Further, we offer our data as an open-source repository for the community offering multiple vantage points for the various runs including the operating system, chipset (through MobileInsight), drone instrumentation, and server-side packet captures as part of the recorded data streams.

We show that a quantized large-scale system with unknown parameters and training signals can be analyzed by examining an equivalent system with known parameters by modifying the signal power and noise variance in a prescribed manner. Applications to training in wireless communications and signal processing are shown. In wireless communications, we show that the optimal number of training signals can be significantly smaller than the number of transmitting elements. Similar conclusions can be drawn when considering the symbol error rate in signal processing applications, as long as the number of receiving elements is large enough. We show that a linear analysis of training in a quantized system can be accurate when the thermal noise is high or the system is operating near its saturation rate.

Inter-cell interference (ICI) is one of the major performance-limiting factors in the context of modern cellular systems. To tackle ICI, coordinated multi-point (CoMP) schemes have been proposed as a key technology for next-generation mobile communication systems. Although CoMP schemes offer promising theoretical gains, their performance could degrade significantly because of practical issues such as limited backhaul. To address this issue, we explore a novel uplink interference management scheme called anywhere decoding, which requires exchanging just a few bits of information per coding interval among the base stations (BSs). Despite the low overhead of anywhere decoding, we observe considerable gains in the outage probability performance of cell-edge users, compared to no cooperation between BSs. Additionally, asymptotic results of the outage probability for high-SNR regimes demonstrate that anywhere decoding schemes achieve full spatial diversity through multiple decoding opportunities, and they are within 1.5 dB of full cooperation.

We present a high‐fidelity measurement‐based nonlinear model of low‐complexity millimeter‐wave transmit and receive circuits for design and analysis of 1‐bit on‐off‐key (OOK) massive MIMO wireless communications systems. The receive model is based upon a fabricated 38 GHz energy detector, representative of state‐of‐the‐art OOK millimeter‐wave receivers. The model is validated with measurements and includes nonlinear noise modeling. Performance of a large‐scale massive MIMO system is predicted with the model, and predictions are compared against a 4‐transmit‐element, N‐receive‐element testbed. Finally, we compute the channel capacity of several OOK massive MIMO systems, exploring tradeoffs in power consumption and number of transmit and receive cells. Results indicate a 1‐bit OOK array with low power pre‐amplifiers can achieve similar capacity to a classical linear receiver with less than one tenth the power consumption. The 1‐bit array compensates for the per‐cell simplicity by increasing the total number of cells while maintaining low overall power consumption.

Recent efforts to obtain high data rates in wireless systems have focused on what can be achieved in systems that have nonlinear or coarsely quantized transceiver architectures. Estimating the channel in such a system is challenging because the nonlinearities distort the channel estimation process. It is therefore of interest to determine how much training is needed to estimate the channel sufficiently well so that the channel estimate can be used during data communication. We provide a way to determine how much training is needed by deriving a lower bound on the achievable rate in a training-based scheme that can be computed and analyzed even when the number of antennas is very large. This lower bound can be tight, especially at high SNR. One conclusion is that the optimal number of training symbols may paradoxically be smaller than the number of transmitters for systems with coarsely-quantized transceivers. We show how the training time can be strongly dependent on the number of receivers, and give an example where doubling the number of receivers reduces the training time by about 37 percent.

One-bit transceivers with strongly nonlinear characteristics are being considered for wireless communication because of their low cost and low power consumption. Although each such transceiver can support only a low data rate, multiple such transceivers can be used to obtain an aggregate high data rate. An important part of many communication systems is the process of channel estimation, which is particularly challenging when the estimation process uses these transceivers. The standard analysis of estimation mean-square error versus training length that is available for linear transceivers does not apply with the nonlinearities inherent in one-bit transceivers. We analyze the training requirements in a large- scale system and show that the optimal number of training symbols strongly depends on the number of receivers, and the optimal number of training symbols can be significantly smaller than the number of transmitters. These results contrast sharply with classical results obtained with linear transceivers.