I/Q-based Beam Classification with the DeepBeam Dataset
* Register (or log in) to the AI4G Neural Network to add this session to your agenda or watch the replay
This talk introduces the problem statement “I/Q-based Beam Classification with the DeepBeam Dataset” for the 2022 ITU AI/ML in 5G Challenge. Learn how to leverage the DeepBeam dataset and (if needed) the DeepBeam codebase to develop ML/DL models that improve the classification accuracy, while reducing the number of samples that are required for the classification.
Highly directional millimeter wave (mmWave) radios need to perform beam management to establish and maintain reliable links. To achieve this objective, existing solutions mostly rely on explicit coordination between the transmitter (TX) and the receiver (RX), which significantly reduces the airtime available for communication and further complicates the network protocol design. In previous work, we proposed a new approach based on convolutional neural networks (CNNs) for the detection of the transmit beam used by the TX, and the angle of arrival at the RX side. In this way, the RX can associate Signal-to-Noise-Ratio (SNR) levels to beams without explicit coordination with the TX. DeepBeam thus does not require pilot sequences from the TX, nor any beam sweeping or synchronization from the RX. This is possible because different beam patterns introduce different “impairments” to the waveform, which can be subsequently learned by a CNN.