Artificial techniques applied to the improvement of the previous signals in the power amplifiers

Amelec Viloria, Nelson Alberto Lizardo Zelaya, Nohora Mercado-Caruzo
2020 Procedia Computer Science  
A rapid evolution in electronic systems has been experienced in recent years, and one of the fields where this development has been notorious is the telecommunication systems in which users demand more and better services and with higher data transfer speeds. This has generated the need to develop new devices, algorithms and systems that manage to satisfy the requirements demanded y new technologies. An example of the above is the front-end of telecommunication systems. Systems need to be more
more » ... fficient, but some elements of the systems, as the power amplifier, present nonlinearity when operating in its most efficient region, causing that it has to make a commitment between efficiency and linearity. This paper presents a comparison of different artificial neural network architectures, as a behavioral modeling method, to perform digital predistortion of power amplifiers. Abstract A rapid evolution in electronic systems has been experienced in recent years, and one of the fields where this development has been notorious is the telecommunication systems in which users demand more and better services and with higher data transfer speeds. This has generated the need to develop new devices, algorithms and systems that manage to satisfy the requirements demanded y new technologies. An example of the above is the front-end of telecommunication systems. Systems need to be more efficient, but some elements of the systems, as the power amplifier, present nonlinearity when operating in its most efficient region, causing that it has to make a commitment between efficiency and linearity. This paper presents a comparison of different artificial neural network architectures, as a behavioral modeling method, to perform digital predistortion of power amplifiers.
doi:10.1016/j.procs.2020.07.091 fatcat:walniqlbobhero7epa6z5mmd5m