Teng Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Teng Wang has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Pere Ll. Gilabert
Department of Signal Theory and Communications, Universitat Politcnica de Catalunya, Castelldefels, Spain
4 shared publications - Gabriel Montoro
Universitat Politcnica de Catalunya, Castelldefels, Spain
3 shared publications - David Lopez-Bueno
Centre Tecnologic de Telecomunicacions de Catalunya (CTTC), Castelldefels, Spain
1 shared publication - Gelu Liu
School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen, China
1 shared publication - Jose Angel Garcia
Department of Communications Engineering, University of Cantabria, 39005 Santander, Spain
1 shared publication - Junliang Wu
School of Electronics and Communication Engineering and Shenzhen Key Laboratory of Navigation and Communication Integration, Sun Yat-sen University, Shenzhen, China
1 shared publication
Latest Publications
- 1 August 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (8) , 9343-9350
CycleRL: Sim-to-Real Deep Reinforcement Learning for Robust Autonomous Bicycle Control
- 1 August 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 11 (8) , 9343-9350
CycleRL: Sim-to-Real Deep Reinforcement Learning for Robust Autonomous Bicycle Control
- Open Access
Machine-Learning Assisted Optimisation of Free-Parameters of a Dual-Input Power Amplifier for Wideband Applications
This paper presents an auto-tuning approach for dual-input power amplifiers using a combination of global optimisation search algorithms and adaptive linearisation in the optimisation of a multiple-input power amplifier. The objective is to exploit the extra degrees of freedom provided by dual-input topologies to enhance the power efficiency figures along wide signal bandwidths and high peak-to-average power ratio values, while being compliant with the linearity requirements. By using heuristic search global optimisation algorithms, such as the simulated annealing or the adaptive Lipschitz Optimisation, it is possible to find the best parameter configuration for PA biasing, signal calibration, and digital predistortion linearisation to help mitigating the inherent trade-off between linearity and power efficiency. Experimental results using a load-modulated balanced amplifier as device-under-test showed that after properly tuning the selected free-parameters it was possible to maximise the power efficiency when considering long-term evolution signals with different bandwidths. For example, a carrier aggregated a long-term evolution signal with up to 200 MHz instantaneous bandwidth and a peak-to-average power ratio greater than 10 dB, and was amplified with a mean output power around 33 dBm and 22.2% of mean power efficiency while meeting the in-band (error vector magnitude lower than 1%) and out-of-band (adjacent channel leakage ratio lower than −45 dBc) linearity requirements.
Open AccessMachine-Learning Assisted Optimisation of Free-Parameters of a Dual-Input Power Amplifier for Wideband Applications
This paper presents an auto-tuning approach for dual-input power amplifiers using a combination of global optimisation search algorithms and adaptive linearisation in the optimisation of a multiple-input power amplifier. The objective is to exploit the extra degrees of freedom provided by dual-input topologies to enhance the power efficiency figures along wide signal bandwidths and high peak-to-average power ratio values, while being compliant with the linearity requirements. By using heuristic search global optimisation algorithms, such as the simulated annealing or the adaptive Lipschitz Optimisation, it is possible to find the best parameter configuration for PA biasing, signal calibration, and digital predistortion linearisation to help mitigating the inherent trade-off between linearity and power efficiency. Experimental results using a load-modulated balanced amplifier as device-under-test showed that after properly tuning the selected free-parameters it was possible to maximise the power efficiency when considering long-term evolution signals with different bandwidths. For example, a carrier aggregated a long-term evolution signal with up to 200 MHz instantaneous bandwidth and a peak-to-average power ratio greater than 10 dB, and was amplified with a mean output power around 33 dBm and 22.2% of mean power efficiency while meeting the in-band (error vector magnitude lower than 1%) and out-of-band (adjacent channel leakage ratio lower than −45 dBc) linearity requirements.
- 1 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Microwave and Wireless Components Letters
- Vol. 31 (1) , 37-40
Mesh-Selecting for Computational Efficient PA Behavioral Modeling and DPD Linearization
This letter proposes a mesh-selecting (MeS) method for complex-valued signals oriented at significantly reducing the training data required to extract the parameters of mathematical models for characterizing the nonlinear behavior of power amplifiers or digital predistortion linearizers. Experimental results will show the advantages of the proposed MeS method when properly combined with dimensionality reduction techniques. A reduction of the parameters' identification computational complexity by a factor of 65 can be achieved with respect to training with consecutive samples and employing the commonly used QR least-squares solution.
- 1 January 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Microwave and Wireless Components Letters
- Vol. 31 (1) , 37-40
Mesh-Selecting for Computational Efficient PA Behavioral Modeling and DPD Linearization
This letter proposes a mesh-selecting (MeS) method for complex-valued signals oriented at significantly reducing the training data required to extract the parameters of mathematical models for characterizing the nonlinear behavior of power amplifiers or digital predistortion linearizers. Experimental results will show the advantages of the proposed MeS method when properly combined with dimensionality reduction techniques. A reduction of the parameters' identification computational complexity by a factor of 65 can be achieved with respect to training with consecutive samples and employing the commonly used QR least-squares solution.
- Open Access
- 1 October 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 182-185
Comparison of model order reduction techniques for digital predistortion of power amplifiers
This paper compares and discusses four techniques for model order reduction based on compressed sensing (CS), less relevant basis removal (LRBR), principal component analysis (PCA) and partial least squares (PLS). CS and PCA have already been used for reducing the order of power amplifier (PA) behavioral models for digital predistortion (DPD) purposes. While PLS, despite being popular in some signal processing areas, to the best author's knowledge, still has not been used in the PA linearization field. Finally, the LRBR is an iterative search algorithm proposed by the authors in this paper for the sake of comparison. Experimental results are presented and the advantages and drawbacks of each method discussed.
- 1 October 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 182-185
Open AccessComparison of model order reduction techniques for digital predistortion of power amplifiers
This paper compares and discusses four techniques for model order reduction based on compressed sensing (CS), less relevant basis removal (LRBR), principal component analysis (PCA) and partial least squares (PLS). CS and PCA have already been used for reducing the order of power amplifier (PA) behavioral models for digital predistortion (DPD) purposes. While PLS, despite being popular in some signal processing areas, to the best author's knowledge, still has not been used in the PA linearization field. Finally, the LRBR is an iterative search algorithm proposed by the authors in this paper for the sake of comparison. Experimental results are presented and the advantages and drawbacks of each method discussed.
- Open Access
- 7 December 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Microwave Magazine
- Vol. 17 (1) , 79-87
Amping Up, Saving Power: Digital Predistortion Linearization Strategies for Power Amplifiers Under Wideband 4G\/5G Burst-Like Waveform Operation
Fourth-generation (4G) communication systems based on orthogonal frequency division multiplexing (OFDM) and the proposed backwards compatible fifth-generation (5G) variants, like filter-bank multicarrier (FBMC), are based on modulation techniques that allow significantly increased spectral efficiency and capacity in mobile radio access networks (RANs). However, the use of these modulation techniques impacts the requirements of the radio base stations, which have been traditionally the most energy-consuming element of mobile networks, accounting for up to 80% of the energy consumption of RANs [1]. Non-constant envelope-modulation techniques with high peak-to-average power ratios (PAPRs) require highly linear power amplifier (PA) amplitude and phase responses to fulfill stringent spectral mask and modulation accuracy requirements. This is often achieved with significant PA back-off, which considerably reduces PA efficiency because the PA's maximum efficiency is achieved near the saturation point.
- 7 December 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Microwave Magazine
- Vol. 17 (1) , 79-87
Open AccessAmping Up, Saving Power: Digital Predistortion Linearization Strategies for Power Amplifiers Under Wideband 4G\/5G Burst-Like Waveform Operation
Fourth-generation (4G) communication systems based on orthogonal frequency division multiplexing (OFDM) and the proposed backwards compatible fifth-generation (5G) variants, like filter-bank multicarrier (FBMC), are based on modulation techniques that allow significantly increased spectral efficiency and capacity in mobile radio access networks (RANs). However, the use of these modulation techniques impacts the requirements of the radio base stations, which have been traditionally the most energy-consuming element of mobile networks, accounting for up to 80% of the energy consumption of RANs [1]. Non-constant envelope-modulation techniques with high peak-to-average power ratios (PAPRs) require highly linear power amplifier (PA) amplitude and phase responses to fulfill stringent spectral mask and modulation accuracy requirements. This is often achieved with significant PA back-off, which considerably reduces PA efficiency because the PA's maximum efficiency is achieved near the saturation point.
- 1 September 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 57-60
Under-sampling effects and computational cost reduction in RF power amplifier behavioral modeling
This paper presents a study oriented at reducing the computational complexity of least squares (LS) identification of the parameters describing power amplifier's (PA) behavioral models. To reduce the dimensions of the input data matrix, two strategies are proposed: i) model order reduction based on the principal component analysis (PCA) theory; and ii) apply a mesh-selecting method to reduce the number of required equations. In this context, the effect of using under-sampling ADCs for the LS parameter extraction aiming at reducing the costs of PA identification is also discussed. Finally, the trade-off between the cost/complexity reduction and quality (or identification accuracy) loss is evaluated. The proposed strategies can also be considered for low-computational cost digital predistortion implementations.
- 1 September 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 57-60
Under-sampling effects and computational cost reduction in RF power amplifier behavioral modeling
This paper presents a study oriented at reducing the computational complexity of least squares (LS) identification of the parameters describing power amplifier's (PA) behavioral models. To reduce the dimensions of the input data matrix, two strategies are proposed: i) model order reduction based on the principal component analysis (PCA) theory; and ii) apply a mesh-selecting method to reduce the number of required equations. In this context, the effect of using under-sampling ADCs for the LS parameter extraction aiming at reducing the costs of PA identification is also discussed. Finally, the trade-off between the cost/complexity reduction and quality (or identification accuracy) loss is evaluated. The proposed strategies can also be considered for low-computational cost digital predistortion implementations.