Page 341 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

• Interference: Co-channel interference and multipath reflections affect clarity; AI models
help mitigate these through adaptive filtering.

AI is integrated into multiple layers of 5G-NR baseband processing to ensure real-time RF-
based object detection as 4.3 - 5G

• Preprocessing Layer: The programmable baseband receives communication data and
multicore processor is configured w.r.t the data. FPGA-based AI models filter noise,
mitigate interference, enhance clarity, and update AI configurations for optimized
processing in the next communication cycle.
• Feature Extraction Layer: AI model extracts delay, AoA, and signal strength and converts
them into structured feature maps.
• Inference Layer: AI-based CNNs classify objects in real-time using Tensor processor at
edge computation.
• AI Edge Compute (Tensor Processor): Processes the extracted features using CNN-based
object classification models.

The proposed solution thus performs real-time inference for RF-based object detection and
classification.

Partners

Potential partner: IIT Delhi

2�2 Benefits of the Use Case

The integration of AI-driven RF-based object detection into 5G-Advanced and beyond enhances
connectivity, automation, and security across next-generation networks. It also supports smart
urban development by improving public safety, mobility, and infrastructure efficiency.

Smart Infrastructure & Cities

AI-powered RF sensing enables intelligent traffic control, vehicle-to-everything (V2X)
communication, and real-time surveillance, reducing the need for additional sensor
installations. These capabilities improve transportation systems, monitor infrastructure health,
and strengthen disaster response. The solution also supports environmental monitoring and
promotes optimized resource use in urban environments.

Industry 4�0 & Smart Manufacturing

RF-based AI enables automated quality inspection, predictive maintenance, and robotic
operations in low-visibility conditions where optical systems are ineffective. This reduces
manufacturing waste and energy consumption, driving more sustainable industrial growth.

Next-Generation Telecom Networks

AI enhances Integrated Sensing and Communication (ISAC), beamforming, and spectrum
efficiency, resulting in lower power usage and expanded broadband access. These
improvements support digital inclusion and smarter connectivity infrastructure for urban and
rural communities alike.

305