HWHaiming Wang

Haiming Wang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Haiming Wang has a total of 225 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • FARFLAG: Fully Automatic Radio Frequency Layout Generation and Its Application to Design of CMOS Millimeter-Wave Low-Noise Amplifiers

    by Yifan Wang, Jing Gao, Hanyu Liu, Guqiao Chen, Yuzhen Song, Qi Wu, Guangyi Lu, Haiming Wang

    The design of millimeter-wave (mmWave) integrated circuits (ICs) is traditionally a slow process that requires extensive iterations and computationally intensive full-wave electromagnetic (EM) simulations. This paper presents a fully automatic radio-frequency layout generation (FARFLAG) method for the end-to-end synthesis of mmWave ICs, from device generation to final layout. The proposed method utilizes a mixture of machine-learning-assisted optimization, lookup tables, and fixed parameters for device generation. We introduce a hybrid placement and routing algorithm tailored for mmWave ICs, which is enhanced with a two-restart mechanism to improve placement robustness and exploration. The synthesis is driven by an improved Bayesian optimization approach with constrained multi-acquisition functions to navigate the design space efficiently. We demonstrate the effectiveness of the proposed framework by synthesizing two mmWave low-noise amplifiers. By integrating high-fidelity EM simulations, the FARFLAG method significantly reduces the search space compared to conventional methods, enabling the rapid synthesis of complex mmWave ICs with superior performance.


  • Machine Learning-Assisted Circuit-Network-Field Co-Optimization for Metal-Bezel Smartphone Antenna Design

    by Wenjian Gong, Qi Wu, Weiqi Chen, Libin Sun, Hanyang Wang, Haiming Wang, Wei Hong

    To meet advanced performance requirements, modern smartphone antennas frequently adopt circuit-network-field (CNF) co-design, which introduces challenges such as strong variable coupling, high sensitivity, and vast search spaces. Traditional machine learning-assisted optimization (MLAO) methods, typically focused on geometric structures, exhibit limited capability in addressing these complexities. This paper proposes a machine learning-assisted circuit-network-field co-optimization (MLA-CNFCO) method incorporating microwave network analysis. First, logarithmic sampling is applied to circuit variables to reshape the sample space, alleviating issues related to nonlinear impedance and high sensitivity. A proportion-modeling approach is then introduced to enhance the prediction accuracy of radiation pattern surrogate models. To accelerate the process, a microwave network analytical solution (MNAS) is employed to calculate reflection coefficients directly, enabling the rapid identification of superior initial solutions. These coefficients are integrated as constraints, significantly narrowing the search space. Finally, the co-design problem is effectively decoupled by embedding microwave network-based circuit optimization within potential-based geometry optimization. The effectiveness of the MLA-CNFCO framework is validated through the design, fabrication, and measurement of a high-performance smartphone antenna.

  • Amplitude-Phase-Sphere Block Modulation and Demodulation for Resisting Phase Noise in Millimeter-Wave Single-Carrier Wireless Communications

    by Guoxing Duan, Min Fan, Wei Xu, Haiming Wang, Xiaohu You

    Phase noise (PN) significantly degrades the performance of millimeter-wave (mmWave) wireless communication systems. To mitigate this challenge, we propose the amplitude-phase-sphere block modulation (APSBM) scheme, which integrates amplitude-shift keying, phase-shift keying, and sphere modulation. Exploiting the Wiener random walk characteristic of PN, APSBM carries information on the relative amplitude and phase between consecutive symbols, thereby counteracting the impact of common PN. Immunity to residual PN in the scheme is further enhanced through a three-dimensional spherical constellation and modulation configurations. We also propose flexible detection strategies (joint, parallel, and cascaded) at the receiver to accommodate diverse application scenarios, minimizing the influence of PN while maintaining a low computational burden. Simulation and experimental results demonstrate that in the presence of PN, APSBM outperforms quadrature amplitude modulation (QAM), circular QAM, and spiral modulation, achieving a lower peak-to-average power ratio, a reduced bit error rate, and a higher achievable information rate. The proposed modulation and demodulation scheme is particularly effective under high PN conditions, offering a robust solution for PN mitigation in mmWave wireless communication systems.

  • Two-Stage Signal Reconstruction for Amplitude-Phase-Time Block Modulation-Based Communications

    by Meidong Xia, Min Fan, Wei Xu, Haiming Wang, Xiaohu You

    Operating power amplifiers (PAs) at lower input back-off (IBO) levels is an effective way to improve PA efficiency, but often introduces severe nonlinear distortion that degrades transmission performance. Amplitude-phase-time block modulation (APTBM) has recently emerged as an effective solution to this problem. The intrinsic amplitude and phase constraints of each APTBM block can be leveraged to mitigate PA-induced nonlinear distortion via constraint-guided signal reconstruction. However, existing reconstruction methods apply these constraints only heuristically and statistically, limiting the achievable IBO reduction and PA efficiency improvement. This paper addresses this limitation by decomposing the nonlinear distortion into dominant and residual components, and accordingly develops a novel two-stage signal reconstruction algorithm consisting of coarse and fine reconstruction stages. The coarse reconstruction stage eliminates the dominant distortion by jointly exploiting the APTBM block structure and PA nonlinear characteristics. Subsequently, the fine reconstruction stage minimizes the residual distortion by casting it as a nonconvex optimization problem subject to explicit APTBM constraints, for which a closed-form solution is derived. The proposed algorithm is validated through comprehensive numerical simulations and testbed experiments. Results show that, without compromising transmission quality, the proposed algorithm enables an additional IBO reduction of approximately 5 dB in simulations and 2 dB in experiments over baseline methods, yielding relative PA efficiency improvements of 77.8% and 30.9%, respectively.

  • Full-Chain Automated ESD Power Clamp Design: From Specification to Silicon Validation With Diverse Performance Priorities

    by Zelong Huang, Guangyi Lu, Dengke Chen, Haoyu Xia, Qi Wu, Haiming Wang

    Electrostatic discharge (ESD) power clamps are critical for ensuring the reliability of modern integrated circuits (ICs). This paper presents an intelligent and automated design methodology that features application-driven topology selection tailored to diverse performance priorities and full-chain automation, from design specifications to silicon validation. The proposed approach begins with selecting representative circuit topologies, each tailored to specific application scenarios with distinct performance priorities. These topologies form a set of curated candidates defined by formalized design rules and established ESD protection principles. Subsequently, an optimization algorithm refines the dimensions of the device according to given specifications, while a professional floorplan and routing ensure high-current handling layout implementation. Compared with conventional manual design, the proposed methodology achieves superior design flexibility, comprehensive automation, and optimized performance, delivering balanced trade-offs among performance, power, and area (PPA). Post-layout simulations and silicon measurements in a 40-nm CMOS process validate the methodology, demonstrating the effectiveness of the intelligent algorithm in ESD protection design.


  • Machine Learning-Assisted Synthesis of Heterogeneous Arrays via Active Geometric Design

    by Wenzong Tao, Qi Wu, Bensheng Yang, Chen Yu, Guangyi Lu, Qian Zhang, Haiming Wang, Wei Hong

    Mutual coupling often leads to discrepancies between intended and synthesized array patterns due to the distortion of active element patterns (AEPs). In array synthesis, existing machine learning methods are predominantly confined to predicting AEPs for fixed structures, imposing a passive constraint that limits optimization flexibility. A holistic framework is introduced for the active geometric design of heterogeneous arrays, in which the geometric parameters of individual elements are treated as degrees of freedom alongside traditional ones. To manage the complexity associated with mutual coupling, a variational autoencoder-based surrogate model is developed to efficiently capture the mapping between diverse element structures and their radiation patterns in a latent space. Numerical comparisons indicate a 9.8× speedup in prediction over traditional methods. Furthermore, under rigorous uniform amplitude and 1-bit phase constraints, sidelobe suppression improvements of over 8 dB are achieved at scanning angles compared to homogeneous baselines, validating the efficacy of incorporating element-level geometric design into the global optimization process.

  • Sequential Hybrid Synthesis for Thinned Prephased Electronically Steered Phased Array Using Low-Resolution Phase Shifters and RF Switches

    by Ming Cheng, Qi Wu, Chen Yu, Haiming Wang, Wei Hong

    To reduce the complexity of large-scale electronically steered phased arrays (ESPAs) that use high-resolution attenuators and phase shifters, this letter proposes a sequential hybrid method. The method can synthesize a thinned prephased ESPA (TPESPA) that uses low-resolution phase shifters and 1-bit RF switches, while maintaining low sidelobe levels (SLLs) and high gains. The approach sequentially synthesizes the high-dimensional 0-1 amplitude vector, the prephase vector, and the one-dimensional reference phase. It thereby overcomes challenges such as slow computation, the unstable results of existing 0-1 amplitude algorithms, and the large storage demands of existing phasing methods. Simulations and measurements show that the synthesized $32\times 32$-unit TPESPA controlled by 2-bit phase shifters and 1-bit RF switches demonstrates SLLs between $-20.0$ dB and $-18.0$ dB, with an average increase in active gain of 1.1dB for a sparsity of 26.6% compared to a Taylor array using 6-bit phase-amplitude control.

  • Open Access

    Fully Automated Schematic and Layout Synthesis of Broadband LC Matching Networks with Arbitrary Topologies for RFIC Designs

    by Guqiao Chen, Yifan Wang, Yuzheng Song, Hanyu Liu, Qi Wu, Guangyi Lu, Wei Hong, Haiming Wang

    Broadband impedance matching is a cornerstone of radio frequency integrated circuit design. While LC matching networks are widely used, their implementation is often hindered by restricted design spaces, labor-intensive manual tuning, and high computational costs. This paper presents a fully automated framework for the synthesis and layout generation of arbitrary LC matching networks, addressing the inefficiencies of traditional workflows. The proposed framework integrates topology-agnostic network synthesis via the combinatorial optimization of formally encoded LC units. To eliminate manual layout drafting, an automated placement and routing engine is employed, followed by a layout sizing optimization process to bridge the performance gap between schematic and layout. Furthermore, to accelerate the electromagnetic-aware refinement, the framework utilizes variational auto-encoders to compress high-dimensional S-parameter data into low-dimensional latent representations, enabling highly efficient Bayesian optimization. The methodology is validated through the implementation of a K/Ka-band lownoise amplifier and a Ku-band single-pole double-throw switch in a 40-nm CMOS process. Post-layout simulation results demonstrate that this framework drastically reduces design cycle time and complexity while maintaining competitive high-frequency performance compared to conventional manual design flows.