Data-Driven Bayesian Optimization Framework for Rapidly Developing Novel Wideband, Low-Profile Dipole Antenna With 3-D-Printed Technology

Ying Sun, Jin Zhang, Alessandro Tibo, Peyman Aghabeyki, Zhaohui Wei, Shengyuan Luo, Shuai ZhangIEEE Transactions on Antennas and Propagation 2024 · DOI

Abstract

This article presents an innovative framework for the rapid development of broadband, low-profile, dual-polarized cross-dipole antennas. By integrating a petal-shaped cross dipole with a 3D-printed high-dielectric constant loading (HDL) structure, the researchers achieved significant performance gains in both bandwidth and profile reduction. The design process utilizes an enhanced tree-structured Parzen estimator Bayesian optimization (TPE-BO) algorithm to navigate a complex 20-parameter design space efficiently. The resulting antenna achieves a 79.3% overlap bandwidth and a profile height of only $0.1\lambda_L$, demonstrating a simple yet effective strategy for high-performance antenna realization.

Introduction

Achieving ultra-wideband performance in a low-profile package remains a persistent challenge in antenna engineering. Modern wireless communication and radar systems require antennas that are both wideband and compact. However, reducing the profile by moving an antenna closer to a metal ground plane typically causes destructive interference from reflected waves, which severely limits bandwidth and efficiency. Traditional solutions often involve complex parasitic elements or tedious manual design iterations. This work introduces a "black-box" optimization framework that leverages 3D-printed dielectric structures and machine learning to overcome these fundamental limitations.

Hardware Configuration: High Dielectric constant Loading (HDL)

The proposed antenna architecture consists of a standard petal-shaped cross-dipole mounted above a 3D-printed HDL structure. This loading structure is composed of 10 concentric rings of a high-permittivity ABS-based compound ($\epsilon_r \approx 14.2$).

The proposed antenna configuration featuring the petal-shaped radiation patch and the multilayer high dielectric constant loading (HDL) structure.

The HDL structure functions as a phase compensation layer. By slowing the phase velocity of electromagnetic waves, it allows the antenna to satisfy radiation conditions at a physical height of just $0.1\lambda_L$, significantly lower than the traditional $0.25\lambda_L$ limit. Additionally, the optimized ring configuration enables the HDL to act as a Dielectric Resonator Antenna (DRA), exciting new resonance modes that merge with the dipole's intrinsic resonances to create a single, massive operating bandwidth.

Optimization Framework: TPE-BO

To optimize the complex geometry of the HDL structure, the researchers defined 20 independent parameters representing the radii ($r_1$ to $r_{10}$) and heights ($h_1$ to $h_{10}$) of the concentric rings. Navigating this high-dimensional, non-convex design space is challenging for traditional gradient-based or standard Bayesian methods.

Detailed geometry of the irregular rhombic dipole (a) and the 20-parameter HDL structure (b, c) optimized via the TPE-BO algorithm.

The authors employed the Tree-structured Parzen Estimator Bayesian Optimization (TPE-BO) algorithm. Unlike Gaussian Process-based Bayesian optimization, TPE-BO models the probability distributions of successful and unsuccessful parameter sets separately. The design was validated through a co-simulation environment linking Python-based optimization scripts with CST Studio Suite. This approach proved highly efficient, reaching a superior global optimum in approximately 700 simulations, whereas industry-standard methods like Particle Swarm Optimization (PSO) and Trust Region Framework (TRF) failed to reach target performance even after 1,200 iterations.

Results and Discussion

The prototype validation yielded several key performance metrics:

  • Bandwidth: A measured 79.3% overlap bandwidth (1.62 to 3.75 GHz) with a Voltage Standing Wave Ratio (VSWR) less than 1.6.
  • Profile Reduction: The antenna height was halved from $0.2\lambda_L$ to $0.1\lambda_L$ compared to a dipole without the HDL structure.
  • Isolation and Gain: Port-to-port isolation exceeded 20.5 dB, and the realized gain ranged from 5 to 10.2 dBi.

While the results are promising, the use of ABS-based filament introduced some dielectric loss (loss tangent $\approx 0.018$), slightly reducing efficiency at higher frequencies. Future work will focus on utilizing lower-loss materials and refining the framework for more complex multi-functional antenna arrays.

Conclusion

This research demonstrates a repeatable and effective design strategy for wideband, low-profile antennas. By combining the physical flexibility of 3D printing with the computational efficiency of TPE-BO, the framework enables the rapid development of high-performance components for modern communication systems.

Summary updated

Abstract

This article presents an innovative framework for fast-developing broadband, low-profile, dual-polarized cross-dipole antennas. This approach integrates a broadband, petal-shaped cross dipole with a 3-D-printed structure comprising multilayers of high-permittivity material with varying heights and radii. In the absence of a theoretical model and a satisfactory initial design, the design process employs the enhanced tree-structured Parzen estimator Bayesian optimization (TPE-BO) algorithm, a significant advancement over traditional Bayesian methods, for effective hyperparameter tuning to streamline the design. The proposed method only adjusts the high dielectric constant loading (HDL) structure without changing the original dimensions of the dipole. Compared with conventional optimization algorithms integrated within CST, this method efficiently achieves the desired broadband and low-profile performance. Meanwhile, the mechanism of the HDL structure in increasing the bandwidth (BW) while enabling the dipole’s low-profile performance is explained. The proposed approach has been validated by fabricating and measuring a hybrid antenna prototype. The measured results show that the antenna achieves an overlap BW between the impedance matching BW and the 3-dB gain BW of 79.3% (from 1.62 to 3.75 GHz) with a voltage standing wave ratio (VSWR) of less than 1.6. In addition, the isolation between the two ports exceeds 20.5 dB. Significantly, this HDL technique notably enhances the BW and reduces the antenna’s profile from <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.2\lambda _{L}$ </tex-math></inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.1\lambda _{L}$ </tex-math></inline-formula> compared with a single-petal-shaped cross dipole. Thus, this study provides a simple, effective, and repeatable design strategy for realizing broadband and low-profile performance in dipole antennas.
BibTeX
@article{sun2024data,
  author = {Ying Sun and Jin Zhang and Alessandro Tibo and Peyman Aghabeyki and Zhaohui Wei and Shengyuan Luo and Shuai Zhang},
  title = {{Data-Driven Bayesian Optimization Framework for Rapidly Developing Novel Wideband, Low-Profile Dipole Antenna With 3-D-Printed Technology}},
  journal = {IEEE Transactions on Antennas and Propagation},
  year = {2024},
  doi = {10.1109/tap.2024.3490836},
}
Figure