Let’s envision a busy airport where unauthorized drone activity and unauthorized users of RF spectrum can pose serious risks to flight safety and network performance. Booz Allen’s R.AI.DIO AI-enabled spectrum sensing system can detect and classify this in-band interference, even if it is overlapping normal commercial and aviation transmissions. Critical capabilities such as video feeds or ISACs can be degraded by jammers or unauthorized users' interference. R.AI.DIO Spectrum Sensing is a GPU-enabled, distributed application or dApp that continuously monitors the RF environment, allowing mobile network operators and private 5G providers to respond to threats in real-time by alerting security teams or by initiating automated mitigations like physical resource block, or PRB, blanking to preserve network integrity. Here we have a laptop that is using a software-defined radio, SDR, to transmit signals which interfere with the 5G UE uplink inside the shielded box or test chamber. In this setup, one or more users send uplink data via a UE, and an SDR creates in-band interference, which R.AI.DIO senses. When an uplink message transmission fails, R.AI.DIO then detects whether interference is present, and the output is shared with the spectrum sensing capability. Next, we will look at the R.AI.DIO dApp view. This visualization helps us to quickly identify and address any issues. The R.AI.DIO AI model provides details about an interfering signal, such as the interfering waveform type,  percentage confidence level for that prediction, and more. This view indicates interference detection. You will see a grid of network resources where green signifies normal operation, gray is unused spectrum,  and orange, red, or purple indicates the presence of a jammer. This visual representation helps us quickly identify and address any issues. When a user activates the jammer by selecting a waveform type and transmit duration, a command is sent to an SDR connected to the RF test chamber with the radio unit and UE inside to inject  interference into the spectrum. When the interference is detected by the R.AI.DIO spectrum sensing application, the time and frequency cells on the plot will be color coded orange, red, or purple, depending on the type of interference detected. This allows network operators to quickly determine if degraded network performance is due to unauthorized users, jammers, or normal conditions. Here we can see the R.AI.DIO spectrum sensing application correctly identifying the different types of interference that were transmitted at 3.75GHz, with DSSS in orange, QPSK in red, and 64QAM in purple. The R.AI.DIO spectrum sensing application also has the ability to visualize the real-time fronthaul IQ data from NVIDIA's aerial framework, which is used by the dApp to identify interference. The top plots show the spectrogram view of data at each antenna. The middle plots show the time series IQ data. In the bottom, plots show the PSD, or power spectral density, where you can see the interfering signal in the center of the spectrum, which is stronger than the 5G uplink. R.AI.DIO enables mobile network operators to understand if users are experiencing degraded performance on a network due to normal conditions, or if malicious actors are causing degraded network performance so that MNOs can appropriately respond. 

Imagine we are at an intersection we want to monitor for transportation and pedestrian safety. We will use combined vision and integrated sensing and communications, or ISAC capabilities to sense pedestrian traffic in order to meet this need. Booz Allen's Vision AI is a platform for edge inference that can ingest various data types such as imagery, video, and radio frequency signals data. To run AI on RAN, we are  implementing vision AI output for detection and classification of pedestrians. For this demo, NVIDIA is implementing their ISAC capability. This system showcases how existing wireless signals can be used to sense and understand the environment, utilizing available 5G radio waves. Camera feeds provide excellent sensing in clear, non-obstructed conditions. However, the ISAC sensing capability provides additional information where vision is weaker, such as at night or conditions with low visibility like rain, snow, and physical obstructions. ISAC additionally provides information about the detected pedestrians, such as distance and location. For the setup, we have an RU acting as the RF receiver, along with a camera positioned at the same location as the RU. We also use a commercial UE for uplink transmission. With this setup, we are able to sense a pedestrian through two modes,  Vision and ISAC, and view the fused results overlaid. As we switch to conditions with reduced visibility the camera no longer detects the pedestrian, while ISAC continues to detect it. Having multi-modal information produces stronger and more accurate outcomes. By bringing these capabilities to our AI-RAN stack Booz Allen can improve sensing capabilities at the edge, meeting the challenge to use our resources effectively in complex environments. 

Maintaining continuous awareness of the environment is critical for many industries such as public safety and emergency response. Visible light cameras, also known as electro optical cameras, are valuable sensing tools. But just like human eyes on their own, they have limitations – especially in low visibility conditions such as fog, obstructions, or darkness. The answer to this is sensor fusion, combining multiple data types to enhance certainty. Utilizing multimodal information produces stronger outcomes compared to relying on a single sensing modality. The integration of vision,  IR, and ISAC capabilities enhances overall scene comprehension, as each modality compensates for the other's limitations. Deploying multiple cameras expands coverage and reduces blind spots while integrating infrared, or IR, and integrated sensing and communications, or ISAC, significantly enhances detection confidence when visibility is poor. Consider a scenario where emergency responders need to perform search and rescue in an environment with poor lighting conditions at night and partially hidden victims. By performing sensor fusion and combining multiple electro optical, or EO, cameras with IR and ISAC sensors, the system can reliably detect and track the individual, even in challenging conditions. Booz Allen's Vision AI platform provides artificial intelligence model training and deployment  that ingests various data types, or modalities, such as imagery, video, and radio frequency signals data. In this scenario, Vision AI  brings AI to RAN through the use of these trained machine learning models for detection and classification of people for search and rescue. Here, EO and IR data are streamed with results displayed in real time. Edge fusion builds on our existing ISAC capabilities by introducing IR for resiliency. A combined spatial view allows for rapid user understanding of the scene. Here we can sense a person with multiple modes, with fused results displayed as a unified detection. Additionally, sensor fusion combines information from multiple modalities, increasing confidence and providing reliable detections even in low visibility or complex conditions. As we switch to darkened conditions where vision alone no longer produces accurate detections, we see we are still able to detect the person due to the added resiliency of IR. The integration of multiple data modalities such as vision, IR, ISAC, and others  showcases how sensor fusion  can transform response efforts, ensuring that even in challenging conditions, critical missions are executed with confidence.