Open RAN (O-RAN) has the potential for revolutionizing not only cellular communication but also spectrum sensing by carefully controlling uplink/downlink traffic in shared spectrum bands. In this paper, we present the design ofSenseORAN, which detects the presence of radar pulses within the Citizens Broadband Radio Service (CBRS) band. SenseORAN is especially useful for scenarios where these pulses (highest priority) are fully overlapping with interfering LTE signals (secondary priority licensee), requiring immediate detection of such an occurrence. This design paradigm of re-using existing cellular infrastructure with ORAN-compliant sensing and communication slices can potentially eliminate the need for dedicated spectrum sensors along the coastline as well as severe restrictions on the transmit power for the LTE operators that are enforced today. Our approach involves a machine learning module deployed as aRadar Detection xAppat the near-Real-Time (near-RT) Radio Access Network (RAN) Intelligent Controller, i.e., near-RT RIC. The base station or gNB (i) uses the you-only-look-once (YOLO)-based machine learning framework that is modified to detect radar signals present within spectrograms generated from I/Q samples collected during the regular uplink cellular operation, and (ii) maintains a list of ‘occupied’ channels in the 3.5 GHz CBRS band that indicate radar presence. Our design is validated with (i) an over the air collected dataset composed of Type 1 radar and standard-compliant LTE waveforms, and (ii) an experimental testbed of SDRs running a complete Open RAN stack with a near-RT RIC implementation integrated with our YOLO-based xApp. We show radar detection accuracy of 100% under SINR conditions ≥ 12 dB after combining 7 spectrograms into a single decision. Furthermore, using testbed results, we demonstrate that the gNB can be reconfigured to avoid radar interference within 866 ms, which represents a reduction of 85.5% over the 60 s response time mandated for pausing cellular operation in detecting radar presence in the CBRS band today.

This paper proposes and experimentally validates a new paradigm for computing with wireless signals over-the-air (OTA). It demonstrates the first fully connected (FC) neural network (NN) constructed entirely using channel propagation and signal interference principles. Our design is based on architecting the desired linear operation of an FC layer through the superposition of signals emitted from multiple transmitters and received at a single receiver, similar to multiple input single output (MISO) systems. Our design takes into account several practical considerations, such as the impact of multiple subcarriers, the number of transmit antennas, and the changing wireless channel. The key outcome of our work is developing a principled methodology that transforms a given trained digital FC NN into its OTA equivalent. This novel computational paradigm, which we call AirFC, allows us to run NN tasks without compute-specific hardware during tests. We validate our design using 9 time-synchronized software-defined radios (SDRs) available on the ORBIT testbed, emulating a 16 antenna array. We use the MNIST dataset as input to our wireless FC NN and demonstrate classification with 92.61% accuracy, which proves that our NN with OTA FC layers performs similar to the conventional, all-digital version with an accuracy decrease of only 0.73%.

Wireless communication between an unmanned aerial vehicle (UAV) and the ground base station is susceptible to adversarial jamming. In such situations, it is important for the UAV to indicate a new channel to the BS. This paper describes a method of creating spatial codes that map the chosen channel to the location of the UAVs in space, wherein the latter physically traverses the space from a given so called ”constellation points” to another. These points create patterns in the sky, analogous to modulation constellations in classical wireless communications, and are detected at the BS through a millimeter-wave radar sensor. A constellation point represents a distinct n-bit field mapped to a specific channel, allowing simultaneous frequency switching at both ends without any RF transmissions. The main contributions of this paper are: (i) We conduct experimental studies to demonstrate how such constellations may be formed using COTS UAVs and mmWave sensors, (ii) We develop a theoretical framework that maps a desired constellation design to error and band switching time, including multi-user scenario-specific challenges, (iii) We compare our approach against current FHSS technology and (iv) We experimentally demonstrate jamming resilient communications and validate system goodput for links formed by UAV-mounted software defined radios.

Over-the-air analog computation allows offloading computation to the wireless environment through carefully constructed transmitted signals. In this paper, we design and implement the first-of-its-kind convolution that uses over-the-air computation and demonstrate it for inference tasks in a convolutional neural network (CNN). We engineer the ambient wireless propagation environment through reconfigurable intelligent surfaces (RIS) to design such an architecture, which we call ’AirNN’. AirNN leverages the physics of wave reflection to represent a digital convolution, an essential part of a CNN architecture, in the analog domain. In contrast to classical communication, where the receiver must react to the channel-induced transformation, generally represented as finite impulse response (FIR) filter, AirNN proactively creates the signal reflections to emulate specific FIR filters through RIS. AirNN involves two steps: first, the weights of the neurons in the CNN are drawn from a finite set of channel impulse responses (CIR) that correspond to realizable FIR filters. Second, each CIR is engineered through RIS, and reflected signals combine at the receiver to determine the output of the convolution. This paper presents a proof-of-concept of AirNN by experimentally demonstrating convolutions with over-the-air computation. We then validate the entire resulting CNN model accuracy via simulations for an example task of modulation classification.

Millimeter-Wave (mmWave) has great potential to provide high data dates given its large available bandwidth, but its severe path loss and high propagation sensitivity to different environmental conditions make deployment planning particularly challenging. Traditional slope-intercept models fall short in capturing large site-specific variations due to urban clutter, terrain tilt or foliage, and ray-tracing faces challenges in characterizing mmWave propagation accurately with reasonable complexity. In this work, we apply machine learning (ML) techniques to predict mmWave path loss on a link-to-link basis over an extensive set of 28 GHz field measurements collected in a major city of USA, with over 120,000 links from both urban and suburban scenarios, with over 40 dB variation for links at similar distances. Either raw environmental profile (terrain+clutter) of each link or 8 selected expert features are used to either directly predict path loss via regression-based approaches or predict the best performing option out of a pool of theoretical/empirical propagation models. Our evaluation shows that Lasso regression provides the best path loss prediction with a performance (RMSE 8.1 dB) comparable to the per-site slope-intercept fit (RMSE 8.0 dB), whereas model selection method achieves 8.6 dB RMSE, both are significantly better than the best a posteriori 3GPP model (UMa-NLOS, 10.0 dB).

Accurate beam alignment in the millimeter-wave (mmWave) band introduces considerable overheads involving brute-force exploration of multiple beam-pair combinations and beam retraining due to mobility. This cost becomes often intractable under high mobility scenarios, where fast beamforming algorithms that can quickly adapt the beam configurations are still under development for 5G and beyond. Besides, blockage prediction is a key capability in order to establish mmWave reliable links. In this paper, we propose a data fusion approach that takes inputs from visual edge devices and localization sensors to (i) reduce the beam selection overhead by narrowing down the search to a small set containing the best possible beam-pairs and (ii) detect blockage conditions between transmitters and receivers. We evaluate our approach through joint simulation of multi-modal data from vision and localization sensors and RF data. Additionally, we show how deep learning based fusion of images and Global Positioning System (GPS) data can play a key role in configuring vehicle-to-infrastructure (V2I) mmWave links. We show a 90% top-10 beam selection accuracy and a 92.86% blockage prediction accuracy. Furthermore, the proposed approach achieves a 99.7% reduction on the beam selection time while keeping a 94.86% of the maximum achievable throughput.

5G and open radio access networks (Open RANs) will result in vendor-neutral hardware deployment that will require additional diligence towards managing security risks. This new paradigm will allow the same network infrastructure to support virtual network slices for transmit different waveforms, such as 5G New Radio, LTE, WiFi, at different times. In this multivendor, multi-protocol/waveform setting, we propose an additional physical layer authentication method that detects a specific emitter through a technique called as RF fingerprinting. Our deep learning approach uses convolutional neural networks augmented with triplet loss, where examples of similar/dissimilar signal samples are shown to the classifier over the training duration. We demonstrate the feasibility of RF fingerprinting base stations over the large-scale over-the-air experimental POWDER platform in Salt Lake City, Utah, USA. Using real world datasets, we show how our approach overcomes the challenges posed by changing channel conditions and protocol choices with 99.86% detection accuracy for different training and testing days.

Wireless communication between an unmanned aerial vehicle (UAV) and the ground base station (BS) is susceptible to adversarial jamming. In such situations, it is important for the UAV to indicate a new channel to the BS. This paper describes a method of creating spatial codes that map the chosen channel to the motion and location of the UAVs in space, wherein the latter physically traverses the space from a given so called ''constellation point'' to another. These points create patterns in the sky, analogous to modulation constellations in classical wireless communications, and are detected at the BS through a millimeter-wave (mmWave) radar sensor. A constellation point represents a distinct n-bit field mapped to a specific channel, allowing simultaneous frequency switching at both ends without any RF transmissions. The main contributions of this paper are: (i) We conduct experimental studies to demonstrate how such constellations may be formed using COTS UAVs and mmWave sensors, given realistic sensing errors and hovering vibrations, (ii) We develop a theoretical framework that maps a desired constellation design to error and band switching time, considering again practical UAV movement limitations, and (iii) We experimentally demonstrate jamming resilient communications and validate system goodput for links formed by UAV-mounted software defined radios.