Microwave Journal
www.microwavejournal.com/articles/35790-a-new-frontier-for-power-amplifiers-enabled-by-machine-learning

A New Frontier for Power Amplifiers Enabled by Machine Learning

April 14, 2021

Artificial intelligence (AI) and machine learning (ML) technologies are pervasive in our daily life empowering devices ranging from smart speakers to thermostats, self-driving cars to robots and social networks to banking systems. In wireless communications, ML has been recently applied across all layers including network planning, spectrum sensing, channel modeling, security and even the smart applications running on our mobile devices. Meanwhile, some are envisioning a future communication system that brings the hyper-connected experience to every corner of life in beyond 5G and 6G.1 Application and deployment of AI technology for next-generation wireless communications have the profound potential to improve the end-to-end experience and reduce both the CAPEX and OPEX of networks.2 AI becomes a necessary tool for delivering reliable and versatile services to connect hundreds of billions of machines and humans.

Improving radio hardware performance of radio access network, particularly, RF power amplifiers (PAs), has been a long-lasting challenge with ever-increasing system demands. In the past decades, RF engineers have spent numerous efforts to enhance PAs figure of merits such as power efficiency, gain, bandwidth and linearity. They came up with many brilliant solutions. Nevertheless, as the complexities of advanced PA circuits, modules and systems keep increasing, it becomes even more challenging and time consuming to design, operate and optimize PAs for highly dynamic signals with fast varying envelopes, dynamic network traffic and beam dependent radio environments such as massive-MIMO. However, such challenging use cases are becoming very common for modern mobile communications.

This article focuses on the recent studies of introducing ML for radio frequency PAs’ online operational conditions optimization, primarily at sub-6 GHz frequency of 5G. Two demonstrators of advanced PA architectures are designed with cutting edge 0.15 μm GaN high electron mobility transistor (HEMT) technology, namely: a digital Doherty power amplifier (DDPA) and an innovative digitally assisted ultra-wideband mixed mode dual-input PA based on frequency-periodic load modulation (FPLM). For both examples, compact data-driven ML techniques are applied to significantly boost PAs performance. Combined with innovative hardware design, AI and ML can become a powerful tool to assist RF engineers dealing with sophisticated PAs design and operating challenges.

DIGITAL TO INTELLIGENT DOHERTY PA

Doherty PAs have been the workhorse for cellular base station radio transmitters3 thanks to its relatively simple topology and attractive average power efficiency for amplifying signals with high peak-to-average power ratio (PAPR > 6 dB). Due to its active load pulling principles and analog nature, Doherty PAs still suffer from several key limitations such as non-optimal power splitting ratio, phase alignment and peaking amplifier turning-ON, especially over wide RF bands and input power levels.

Figure 1

Figure 1 A wideband Doherty power amplifier8 and its modified version for dual-input Doherty power amplifier. The modified input power divider is outlined.

To overcome such difficulties, various modified design methods and architectures including Advanced Doherty Alignment module and DDPA were proposed by eliminating the conventional analog-based power splitting circuitry (i.e., Wilkinson divider). Instead, these designs are feeding dual-input RF signals directly to the Carrier and Peaking amplifiers of the Doherty PA, respectively.4,5 Hence, the circuit can independently control input signals amplitudes and phases with better results. Figure 1 provides a comparison of a Doherty PA and its modified version as dual-input DDPA. The input network change is highlighted in the figure.

Multi-input Doherty PA can be digitally controlled by following a set of derived closed-form equations, which approximate a pre-determined static power splitting ratio and phase imbalance between Carrier and Peaking amplifiers. Alternatively, it can be done by offline brute force search, finding an optimum input signal condition for high efficiency or high output power.5-7 However, these two approaches have several limitations in practice: (1) derived mathematical equations only provide an approximation of highly non-linear relationship within PA (i.e., using arctan function), (2) bias voltages optimization is not included but critical and (3) open-loop implementation does not capture the device-to-device variation or operating condition changes (i.e., ambient temperature). Consequently, manual tuning is still required to account for the dynamics of real systems and condition variations. Because of the large searching space of variables, brute force searching is inefficient for practical implementations.

Very recently, there have been new ML data-driven online optimization methods proposed and demonstrated. In an initial study shown by simulation,9 a Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm was applied to optimize the input power splitting ratio, phase offset and gate bias voltages at the same time for the Carrier and Peaking amplifiers of a dual-input Doherty PA using ADS and SystemVue software. The algorithm is here:

Algorithm

 

Figure 2

Figure 2 Online optimization of dual-input Doherty PA.

It formulates digital DDPA real-time optimization as an adaptive online control problem by searching for an optimum solution for a user defined cost function consisting of several PAs figure of merits (a weighted sum of power, gain, efficiency and linearity etc.), as depicted by Figure 2. Different hyper-parameters and initial conditions of optimization were tested. As a result, optimal points of power added efficiency can be found between 60 to 70 percent with many closely spaced local minimum points. A further development with a lab test bench, shown in Figure 3, is a proof-of-concept and engineering demonstration that was implemented.10

Figure 3

Figure 3 Testbench of dual-input digital Doherty PA with machine learning online optimization.

One setup implemented a model-free optimization method with simulated annealing (SA) and extremum seeking (ES), as shown in Figure 4.10 The combination of SA and ES makes the system optimization efficient. SA captures the random and abrupt variation in the system mainly due to frequency and input power level variations, where ES captures slow variation in the model such as temperature.

Figure 4

Figure 4 Model-free machine learning algorithms used for DDPA optimization.10

Figure 5

Figure 5 DDPA performance with online auto-tuning of control parameter: Pout and defined cost function (a), gate bias voltages for main and peaking amplifier (b), PAE and Gain (c) and input phase imbalance and power splitting ratio (d).

 

The compactness of ML algorithm adopted here is Quite different from the general deep learning ML category, such as deep neural network, in the sense that it neither requires massive training data nor powerful computation power and memory. This is an important feature for efficient implementation of RF front-end applications. Figure 5 shows DDPA online auto-tuning of performance including output power, gain, power added efficiency via adaptive control of gate bias voltages (Vg_main, Vg_peak) and input power splitting ratio (α: how much power distributed to Peaking amplifier from total input) and phase imbalance (ΔΦ) using SA and ES. The optimization goal is to search for an optimal control parameters θ* maximizing cost function Q(θ), which is expressed as the weighted sum of PA performance of interest: θ* = argmaxQ(θ), θ∈U, where θ is a vector of the amplifier tuning parameters defined as θ = [Vg_main, Vg_peak, ΔΦ, α].

As shown in Figure 5, it takes approximately 40 iterations for SA to perform random exploration with Quick convergence, limited mainly by the interface communication of the test instruments. SA is then followed by ES algorithm for a fine tuning to account for effects such as temperature changes. The program is written in MATLAB and running on a PC controlling the measurement setup depicted in Figure 3. Significant performance enhancement in DDPA over a wider frequency range and different input power range (in particular lower input power range) has been observed compared with single input conventional Doherty PA thanks to the auto-tuning procedure. Over a 15 percent efficiency boost and 2 to 3 dB gain is realized without using digital predistortion (DPD). The algorithm is also able to figure out a reasonable tradeoff among these conflicting PA performance targets by assigning different weights in Q(θ). It must be mentioned that dedicated DPD schemes were not used.10



DIGITALLY ASSISTED FREQUENCY-PERIODIC LOAD MODULATION PA

Doherty PAs in practice are limited in terms of RF bandwidth, due to many factors such as device patristics, power combiner and phase alignment challenges. An innovative mixed mode ultra-wideband FPLM PA has been proposed to achieve high power efficiency over multiple contiguous frequency bands, enabled by a digitally assisted dual-input AI module. It provides automatic optimum signal combination, magnitude and phase of dual-input signals. Figure 6 illustrates several types of load modulations, such as a virtual open stub Doherty, Outphasing, general Doherty and anti-phasing Outphasing spanned over a 3x RF frequency range (0.5f0~1.5f0, f0 denotes the design center frequency). Very distinct and proper input signal’s amplitude and phase relationships are necessary for this amplifier to behave as Doherty and Outphasing modes over five different frequency ranges.

Figure 6

Figure 6 Concept of a frequency-periodic load modulated (FPLM) PA with controlled dual-input signal over frequency, and f0 being the center frequency.

 

A new output combiner was also proposed by absorbing devices capacitances into part of the equivalent transmissions and providing differently desired output power combining functions for above mentioned five freQuency ranges, respectively. The design details can be found in Y. Komatsuzaki et al.11

Figure 7 shows the FPLM GaN PA prototype consisting of two bare die chips with 0.15 μm HEMTs. Similar AI algorithms shown in Figure 4 have been adopted to auto-tuning the two RF input signals amplitude and phases based on a user defined cost function. The bias voltages of these two HEMTs are at pinch-off and not tuned during the optimizations. Without manual interaction and specifying the specific PA operation modes, the AI module is able to autotune the dual-channel transceivers parameters on the fly and achieves the desired PA modes with high efficiency. Figure 8 shows the measured FPLM PA average efficiency (6 dB back-off) over the whole band.

Figure 7

Figure 7 Prototype of a dual-input FPLM GaN PA under AI digitally assisted operation.

Figure 8

Figure 8 Measurement of the prototype dual-input FPLM PA.

 

A detailed analysis of the operating mode at each frequency range is omitted here but can be found in Komatsuzaki et al.11 Digital assistance is able to fully use the FPLM PA’s design potential and handle its sophisticated control and optimization thus offering the state-of-the-art efficiency performance over 110 percent fractional bandwidths, as compared in Table 1.

Table 1

 

SUMMARY

The reported applications show that compact data-driven AI 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 AI (digital assisted auto-tuning together with digital front-end including digital predistortion14) 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 AI-assisted RF front-end modules a highly promising solution for future radios.

 

Related Video: Machine Learning Power Amplifier

 

References

  1. 6G Vision Whitepaper, The Next Hyper Connected Experience for All, Samsung Research, July 2020.
  2. X. Zhou et al., “Intelligent Wireless Communications Enabled by Cognitive Radio and ML,” China Communications, December 2018, pp. 16–48.
  3. F. H. Raab et al., “Power Amplifiers and Transmitters for RF and Microwave,” IEEE Transactions on Microwave Theory and Techniques, Vol. 50, No. 3, March 2002, pp. 814–826.
  4. A. Ahmed et al., “Line-up Efficiency Improvement Using Dual Path Doherty Power Amplifier,” 2013 European Microwave Conference, 2013, pp. 275–278.
  5. R. Darraji et al., “Doherty Goes Digital: Digitally Enhanced Doherty Power Amplifiers,” IEEE Microwave Magazine, Vol. 17, No. 8, August 2016. pp. 41–51.
  6. R. Darraji and F. M. Ghannouchi, “Digital Doherty Amplifier with Enhanced Efficiency and Extended Range,” IEEE Transactions on Microwave Theory and Techniques, Vol. 59, No. 11, November 2011, pp. 2898–2909.
  7. C. M. Andersson et al., “A 1–3 GHz Digitally Controlled Dual-RF Input Power-Amplifier Design Based on a Doherty-Outphasing Continuum Analysis,” IEEE Transactions on Microwave Theory and Techniques, Vol. 61, No. 10, Oct. 2013, pp. 3743–3752.
  8. Y. Komatsuzaki et al., “3.0–3.6 GHz Wideband, Over 46% Average Efficiency GaN Doherty Power Amplifier with Frequency Dependency Compensating Circuits,” IEEE Topical Conference on RF/Microwave Power Amplifiers for Radio and Wireless Applications (PAWR), 2017, pp. 22–24.
  9. S. Niu et al., “Stochastically Approximated Multi Objective Optimization of Dual Input Digital Doherty Power Amplifier,” IEEE 10th International Workshop on Computational Intelligence and Applications (IWCIA), 2017, pp. 147–152.
  10. R. Ma et al., “Machine-Learning Based Digital Doherty Power Amplifier,” IEEE International Symposium on Radio-Frequency Integration Technology (RFIT), 2018, pp. 1–3.
  11. Y. Komatsuzaki et al., “A Novel 1.4-4.8 GHz Ultra-Wideband, Over 45% High Efficiency Digitally Assisted Frequency-Periodic Load Modulated Amplifier,” IEEE MTT-S International Microwave Symposium (IMS), 2019, pp. 706–709.
  12. J. M. Rubio et al., “3–3.6-GHz Wideband GaN Doherty Power Amplifier Exploiting Output Compensation Stages,” IEEE Transactions on Microwave Theory and Techniques, Vol. 60, No. 8, Aug. 2012, pp. 2543–2548.
  13. S. Sakata et al., “An 80 MHz Modulation Bandwidth High Efficiency Multi-Band Envelope-Tracking Power Amplifier Using GaN Single-Phase Buck-Converter,” IEEE MTT-S International Microwave Symposium (IMS), 2017, pp. 1854–1857.
  14. M. Mengozzi et al., “A. GaN Power Amplifier Digital Predistortion by Multi-Objective Optimization for Maximum RF Output Power,” Electronics, 2021, https://doi.org/10.3390/electronics10030244.

     

AUTHOR BIOS

Rui Ma received his Dr.-Ing. degree from University of Kassel, Germany in 2009. From 2010 to 2012, Dr. Ma was a Senior RF Research Engineer with Nokia Siemens Networks. Since 2012, he has been with Mitsubishi Electric Research Labs (MERL) in Cambridge, USA, where he is a Senior Principal Scientist of RF Research, leading technologies development of advanced radio including applications of GaN RF power devices. He holds more than 30 US patents and patent applications. In 2013, he received the specification award by MIPI Alliance for the development of Analog Reference Interface (ACI) for Envelope Tracking eTrak specification. He was a visiting scientist with THz Integrated Electronics Group at Massachusetts Institute of Technology (MIT) from 2016 to 2021.

Dr. Ma is currently an Associate Editor of IEEE Transactions on Microwave Theory and Techniques.

Yuji Komatsuzaki received the B.Sc., M.Sc. and Ph.D. degrees in electrical engineering from Waseda University, Tokyo, Japan, in 2007, 2009 and 2012, respectively. 

Since 2012, he has been with the Information Technology Research and Development Center, Mitsubishi Electric Corporation, Kamakura, Japan, where he has been involved with the research and development of microwave amplifiers for telecommunication systems. From 2016 to 2017, he was a visiting scholar with the Center for Wireless Communication, University of California, San Diego.

Mouhacine Benosman is a Senior Research Scientist at Mitsubishi Electric Research Labs (MERL) in Cambridge, USA. He has received his Ph.D. in 2002 from Ecole Centrale de Nantes, France. Before joining MERL he worked at Reims University, France, Strathclyde University, Scotland, and National University of Singapore. His research interests include multi-agent distributed control with applications to robotics and smart-grids, control of infinite dimensional systems with applications to fluid dynamics and learning and adaptive control with applications to analog and RF circuits auto-tuning. He is an Associate Editor for IEEE Control Systems Letters, and for the Journal of Opt. Theory and Applications. He is a Senior Editor of the Inter. Journal of Adaptive Control and Signal Processing.

Koji Yamanaka received his B.S. degree in electric engineering and M.S. and Ph.D. degrees in electronic engineering from the University of Tokyo, Japan, in 1993, 1995, and 1998, respectively. In 1998, he joined the Information Technology Research and Development Center, Mitsubishi Electric Corporation, Kamakura, Japan, where he engaged in development of GaAs low-noise monolithic microwave integrated circuit amplifiers and GaN high-power amplifiers. From 2012 to 2018, he has managed the Amplifier Group, in Mitsubishi Electric Corporation. He was in charge of the civil application GaN device business section in from 2018 to 2020. He is the recipient of the Best Paper Prize of GAAS2005.

Shintaro Shinjo received the B.S. and M.S. degrees in physics and Ph.D. degree in engineering from Keio University, Tokyo, Japan, in 1996, 1998, and 2011, respectively.  In 1998, he joined Mitsubishi Electric Corporation, Kamakura, Japan, where he has been involved in the research and development of microwave monolithic integrated circuits and solid-state power amplifiers. From 2011 to 2012, he was a visiting scholar with the University of California at San Diego, San Diego, CA, USA. He is a senior member of IEEE. He was a recipient of the Prize for Science and Technology (Development Category) of the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology in 2009.