
Day2-12-Northeastern-Edited-University-Platform-Qualcomm-Conference-Wlogos
May 7, 2024 | 10:38
Spectrum Sensing and Beamforming
We use the AI Model Efficiency Toolkit (AIMET) library to address two wireless problems including spectrum sensing and multimodal beamforming. We present customized deep learning frameworks and use the quantization methods in the AIMET library to reduce the computation. Codebase will be open-sourced to expand the AIMET model-zoo for challenging real-world wireless problems. Moreover, we use the Qualcomm® Neural Processing SDK to deploy our models on edge devices and interface the Qualcomm® Innovators Development Kit with Colosseum – a wireless large-scale emulator performing scalable and repeatable experimentation in wireless systems.
