Yizhuo Wu;Yi Zhu;Kun Qian;Qinyu Chen;Anding Zhu;John Gajadharsing;Leo C. N. de Vreede;Chang Gao
{"title":"DeltaDPD: Exploiting Dynamic Temporal Sparsity in Recurrent Neural Networks for Energy-Efficient Wideband Digital Predistortion","authors":"Yizhuo Wu;Yi Zhu;Kun Qian;Qinyu Chen;Anding Zhu;John Gajadharsing;Leo C. N. de Vreede;Chang Gao","doi":"10.1109/LMWT.2025.3565004","DOIUrl":null,"url":null,"abstract":"Digital predistortion (DPD) is a popular technique to enhance signal quality in wideband radio frequency (RF) power amplifiers (PAs). With increasing bandwidth and data rates, DPD faces significant energy consumption challenges during deployment, contrasting with its efficiency goals. State-of-the-art DPD models rely on recurrent neural networks (RNNs), whose computational complexity hinders system efficiency. This letter introduces DeltaDPD, exploring the dynamic temporal sparsity of input signals and neuronal hidden states in RNNs for energy-efficient DPD, reducing arithmetic operations and memory accesses while preserving satisfactory linearization performance. Applying a TM3.1a 200 MHz-BW 256-QAM OFDM signal to a 3.5-GHz GaN Doherty RF PA, DeltaDPD achieves −50.03 dBc in adjacent channel power ratio (ACPR), −37.22dB in normalized mean square error (NMSE) and −38.52 dB in error vector magnitude (EVM) with 52% temporal sparsity, leading to a <inline-formula> <tex-math>$1.8\\times $ </tex-math></inline-formula> reduction in estimated inference power. The DeltaDPD code is available at <uri>https://www.opendpd.com</uri>.","PeriodicalId":73297,"journal":{"name":"IEEE microwave and wireless technology letters","volume":"35 6","pages":"772-775"},"PeriodicalIF":3.4000,"publicationDate":"2025-03-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE microwave and wireless technology letters","FirstCategoryId":"1085","ListUrlMain":"https://ieeexplore.ieee.org/document/11006082/","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"0","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0
Abstract
Digital predistortion (DPD) is a popular technique to enhance signal quality in wideband radio frequency (RF) power amplifiers (PAs). With increasing bandwidth and data rates, DPD faces significant energy consumption challenges during deployment, contrasting with its efficiency goals. State-of-the-art DPD models rely on recurrent neural networks (RNNs), whose computational complexity hinders system efficiency. This letter introduces DeltaDPD, exploring the dynamic temporal sparsity of input signals and neuronal hidden states in RNNs for energy-efficient DPD, reducing arithmetic operations and memory accesses while preserving satisfactory linearization performance. Applying a TM3.1a 200 MHz-BW 256-QAM OFDM signal to a 3.5-GHz GaN Doherty RF PA, DeltaDPD achieves −50.03 dBc in adjacent channel power ratio (ACPR), −37.22dB in normalized mean square error (NMSE) and −38.52 dB in error vector magnitude (EVM) with 52% temporal sparsity, leading to a $1.8\times $ reduction in estimated inference power. The DeltaDPD code is available at https://www.opendpd.com.