Researchers have developed a novel deep learning methodology for designing Doherty power amplifiers (PAs) that incorporate pixelated output combiner networks. This approach utilizes a deep convolutional neural network (CNN) as an electromagnetic surrogate model to rapidly predict the performance of these complex combiners. By integrating this CNN within a black-box framework and employing a genetic algorithm, the team successfully synthesized Doherty combiners that enhance efficiency across a wider range of power levels. Prototypes demonstrated impressive efficiency and output power, maintaining over 52% drain efficiency at a 9-dB back-off level and achieving over 51% power added efficiency with a strong adjacent channel leakage ratio when tested with a 5G NR-like waveform. AI
IMPACT This research could lead to more efficient wireless communication systems by improving power amplifier performance.
RANK_REASON The cluster contains an academic paper detailing a new research methodology and experimental results. [lever_c_demoted from research: ic=1 ai=1.0]
- deep convolutional neural network
- deep learning
- GaN HEMT transistors
- genetic algorithm
- pixelated output combiner networks
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
