A New Frontier For Power Amplifiers Enabled By Machine Learning

The reported applications show that compact data-driven Al techniques can assist in unlocking the full potential of new high performance PAs for flexible and wideband wireless applications. Even after deployment in the field, these devices can adapt to changing operating conditions. Integrating cutting edge GaN semiconductor device technology, circuit design innovation and Al (digital assisted auto-tuning together with digital front-end including digital predistortion(14)) will facilitate a solution of agile and superior performance RF front-ends. It is worth pointing out that the proposed methodology is not only applicable for cellular transmitters, but also for mobile handset and general RF applications (such as microwave industrial heating), in which RF hardware/amplifiers are the key device dominating the system-level performance. Its adaptability and intelligence make Al-assisted RF front-end modules a highly promising solution for future radios.