TR2017-163

Stochastically Approximated Multiobjective Optimization of Dual Input Digital Doherty Power Amplifier


    •  Niu, S., Ma, R., Manjunatha, K., "Stochastically Approximated Multiobjective Optimization of Dual Input Digital Doherty Power Amplifier", IEEE International Workshop on Computational Intelligence and Applications, DOI: 10.1109/​IWCIA.2017.8203576, November 2017.
      BibTeX TR2017-163 PDF Video
      • @inproceedings{Niu2017nov,
      • author = {Niu, Sufeng and Ma, Rui and Manjunatha, Koushik},
      • title = {Stochastically Approximated Multiobjective Optimization of Dual Input Digital Doherty Power Amplifier},
      • booktitle = {IEEE International Workshop on Computational Intelligence and Applications},
      • year = 2017,
      • month = nov,
      • doi = {10.1109/IWCIA.2017.8203576},
      • url = {https://www.merl.com/publications/TR2017-163}
      • }
  • Research Area:

    Electric Systems

Abstract:

In this work, we propose a novel adaptive control on digital Doherty Power Amplifier (DPA) for efficiency andgain enhancement. Unlike traditional DPA design, we propose Simultaneously Perturbated Stochastic Approximation (SPSA) optimization algorithm by considering phase difference, power distribution and gate voltage parameters of main and peak amplifiers in order to achieve optimal performance. The optimal performance is defined as good linearity, Power Added Efficiency (PAE) and gain. We mainly investigate the efficiency and gain improvements since the linearity can be separately addressed by traditional DPD techniques. Even though we found that the cost function of optimization exists in several local optimums, they are very close, and thereby local minimum is not of great concern in this optimization. In addition, we approximate the DPA circuit model for the fast verification using regression model. Finally, we exam the algorithm in SystemVue and ADS co-simulation environment.

 

  • Related News & Events

  • Related Video