Presentation | 2019-11-21 Deep Neural Network Based Distortion Compensation for Doherty Power Amplifier Yoshimasa Egashira, Reina Hongyo, Keiichi Yamaguchi, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | The problem with Doherty amplifiers, which are high-efficiency power amplifiers, is the occurrence of complex memory distortion. In this paper, we focus on digital pre-distortion (DPD) using deep neural networks (DNN) as a technology to compensate for memory distortion of Doherty amplifiers with high accuracy and investigate the effects of the number of learning parameters and neuron activation functions on the compensation performance of DNN-DPD. As a result of the evaluation using actual GaN (Gallium Nitride) Doherty amplifier, it is shown that DNN-DPD can achieve suppression performance that exceeds Volterra series based DPD by increasing the number of learning parameters. Futhermore, in order to improve the distortion compensation performance of DNN-DPD, it is important to select an appropriate activation function according to the number of learning parameters and it is shown that the ReLU function is more suitable for the activation function of DNN-DPD with more than 2000 learning parameters compared to the conventional sigmoid function. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Digital predistortion / Deep neural network / Doherty amplifiers |
Paper # | CQ2019-89 |
Date of Issue | 2019-11-14 (CQ) |
Conference Information | |
Committee | NS / ICM / CQ |
---|---|
Conference Date | 2019/11/21(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Rokkodai 2nd Campus, Kobe Univ. |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Network quality, Network measurement/management, Network virtualization, Network service, Blockchain, Security, Network intelligence, etc. |
Chair | Yoshikatsu Okazaki(NTT) / Kiyohito Yoshihara(KDDI Research) / Hideyuki Shimonishi(NEC) |
Vice Chair | Akihiro Nakao(Univ. of Tokyo) / Takumi Miyoshi(Shibaura Inst. of Tech.) / Yoichi Sato(NEC) / Jun Okamoto(NTT) / Takefumi Hiraguri(Nippon Inst. of Tech.) |
Secretary | Akihiro Nakao(Osaka Pref Univ.) / Takumi Miyoshi(NTT) / Yoichi Sato(NTT) / Jun Okamoto(NTT) / Takefumi Hiraguri(Nippon Inst. of Tech.) |
Assistant | Shinya Kawano(NTT) / Hiroki Nakayama(Bosco) / Chikara Sasaki(KDDI Research) / Yoshiaki Nishikawa(NEC) / Takuto Kimura(NTT) |
Paper Information | |
Registration To | Technical Committee on Network Systems / Technical Committee on Information and Communication Management / Technical Committee on Communication Quality |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Deep Neural Network Based Distortion Compensation for Doherty Power Amplifier |
Sub Title (in English) | |
Keyword(1) | Digital predistortion |
Keyword(2) | Deep neural network |
Keyword(3) | Doherty amplifiers |
1st Author's Name | Yoshimasa Egashira |
1st Author's Affiliation | Wireless System Laboratory, Corporate Research & Development Center, Research & Development Div., Toshiba Corp.(Toshiba) |
2nd Author's Name | Reina Hongyo |
2nd Author's Affiliation | Wireless System Laboratory, Corporate Research & Development Center, Research & Development Div., Toshiba Corp.(Toshiba) |
3rd Author's Name | Keiichi Yamaguchi |
3rd Author's Affiliation | Wireless System Laboratory, Corporate Research & Development Center, Research & Development Div., Toshiba Corp.(Toshiba) |
Date | 2019-11-21 |
Paper # | CQ2019-89 |
Volume (vol) | vol.119 |
Number (no) | CQ-298 |
Page | pp.pp.7-12(CQ), |
#Pages | 6 |
Date of Issue | 2019-11-14 (CQ) |