Design and Measurement of a High-Efficiency W-Band Microstrip Antenna with Enhanced Matching for 6G Automotive Radar and ADAS Systems

Abstract

A compact, single-layer W-band microstrip antenna for forward-looking ADAS radar in the 77–79 GHz band is presented. The 16.5 × 22 mm² PCB element integrates a linear microstrip taper, two shorting vias, and a slot-loaded cavity to stabilize input reactance and broaden the in-band match. Full-wave simulations and launcher-based measurements using WR-12 TRL de-embedding and anechoic-chamber substitution confirm S₁₁ ≤ −10 dB across 77–79 GHz. At 77/79 GHz, the antenna achieves end-fire realized gains of ≈9.9/≈11.2 dBi. The main beam is end-fire (peak near θ ≈ 90°), with −3 dB beamwidths of ≈36° in the θ-cut at φ = 0 (pointing ≈ 61°/56°) and ≈11.6° in the φ-cut at θ = 90°. First sidelobes are about −2.3/−2.5 dB (θ-cut) and −3.1/−3.4 dB (φ-cut). Cross-polarization is ≥18 dB below co-polarization, and the simulated radiation efficiency reaches ≈85% at 77 GHz and ≈80% at 79 GHz. A controlled thermal sweep (25–105 °C) yields < 100 MHz resonance drift while maintaining ≥ 10 dB return loss. Due to its planar architecture and clean feed integration, compact module packaging in short- to medium-range automotive radars.

1. Introduction

1.1. Background and Motivation

Advanced driver-assistance systems and emerging 6G/JCAS vehicular platforms require compact, manufacturable antennas around 77–79 GHz that deliver clean end-fire beams, suppressed sidelobes, and stable impedance under operational temperature swings. With the continuous advancement of automotive technologies, wireless communication systems have become essential for improving vehicle safety, efficiency, and overall user experience [1,2,3]. Among the most transformative innovations are Advanced Driver Assistance Systems (ADAS), which play a crucial role in reducing traffic accidents, an urgent concern, given that approximately 44,450 traffic-related fatalities were recently reported in the United States alone [4]. ADAS are reshaping modern transportation by integrating intelligent sensing and control technologies that enhance situational awareness and vehicle responsiveness [5].

As illustrated in Figure 1, radar-based sensing systems mounted on the vehicle body enable ADAS to detect and track surrounding objects, supporting critical functionalities such as automatic emergency braking, lane departure warning, and adaptive cruise control [6,7]. These capabilities are fundamental to enhancing road safety, particularly considering increasing regulatory mandates and consumer demand for advanced safety features. Moreover, the integration of ADAS with V2X communication networks further amplifies these benefits by enabling more accurate object detection and seamless V2X and V2I [7,8].

A key enabler of radar-based applications is the utilization of millimeter-wave frequency bands, particularly the W-band (75–110 GHz), defined by IEEE standards [9]. Within this spectrum, the 77–79 GHz range is especially important for short- and medium-range automotive radar. Microstrip antennas play a central role in achieving high-resolution detection due to their compact size, ease of integration, and high-frequency performance [10,11].

Ref. [12] proposed a compact dual-band mi-crostrip antenna operating at 24 GHz and 77 GHz for automotive radar and intelligent sensing of ADAS applications. While the design achieved high gain values exceeding 12 dBi in both bands, it suffered from potential inter-band interference, fabrication complexity, and a lack of validation under realistic automotive conditions. In [13], a magneto-electric dipole antenna array was developed specifically for 77 GHz radar applications. It demonstrated high gain (~16.5 dBi), narrow beamwidth, and excellent impedance matching. However, the multilayered structure increased fabrication difficulty and cost. Similarly, Ref. [14] introduced a dual-layer transmit-array antenna at 77 GHz with a peak gain of approximately 21 dBi and beam-steering capability of ±15°, providing improved angular resolution and ~75% radiation efficiency. Yet, the dual-layer configuration introduced significant alignment and fabrication challenges. Furthermore, Ref. [15] reported a 79 GHz wide-beam patch antenna and array for short-range radar, achieving a peak gain of 10.74 dBi. Its angular resolution, however, remained limited, and real-world performance validation was insufficient. In [16], a 77 GHz gap waveguide slot array was proposed, achieving a measured gain of 17.5 dBi with excellent polarization purity and isolation (>30 dB). Despite its high performance, the design required extremely precise fabrication, raising concerns about scalability for automotive mass production.

Although this work focuses on the regulated 77–79 GHz automotive band, numerous recent W-band (75–110 GHz) studies report planar, PCB-compatible antennas and packaging consistent with our approach [17]. Ref. [18] presents a single-layer PCB array with measured ≈14 dBi gain maintained across 100–110 GHz, validating low-cost planar implementations at the upper W-band, as well as a glass-substrate W-band microstrip array that mitigates dielectric and feed losses while sustaining >10 dBi over ~76–84 GHz. Also, Ref. [19] explains an LTCC/SIW W-band 4 × 4 array demonstrating high integration and high gain. Collectively, these findings support extending our element toward the higher end of 75–110 GHz.

1.2. Application Scope and Specifications

These prior works highlight the ongoing research efforts aimed at meeting the stringent requirements of modern ADAS [20,21]. 6G roadmaps target sub-THz/W-band JCAS/ISAC for high-resolution automotive sensing and short-range, high-capacity V2X links [22]. Antennas with clean end-fire beams, suppressed sidelobes, and easy arrayability are essential. The proposed single-layer PCB element at 77–79 GHz provides low mismatch and ≈−9 dB first-sidelobe levels, making it a practical tile for 6G-ready ADAS/JCAS arrays. However, challenges persist, particularly in balancing performance, integration feasibility, and fabrication complexity [23].

Table 1. Antenna specifications for ADAS (77–79 GHz) versus achieved performance of the proposed design.

Parameter	ADAS Target (Spec)	Achieved 77 GHz	Achieved 79 GHz
Operating band	77–79 GHz	Met (within band)	Met (within band)
Input match	S11 ≤ −10 dB (VSWR ≤ 2) across band	S11 ≤ −10 dB; VSWR = 1.03; RL ≈ −36.6 dB	S11 ≤ −10 dB; VSWR = 1.12; RL ≈ −25 dB
Realized gain (end fire)	≥9.5 dBi @77 GHz; ≥10.5 dBi @79 GHz	9.9 dBi	11.2 dBi
Radiation efficiency (sim.)	≥75% over band	≈85%	≈79.7%
Main lobe pointing θmax	θmax ≤ 5° (end fire θ = 90°)	+3° (φ = 0°), +2° (φ = 90°)	+2° (φ = 0°), +1° (φ = 90°)
3 dB beamwidth	15–38°	15.8° (φ = 0°), 13.8° (φ = 90°)	14.6° (φ = 0°), 14.7° (φ = 90°)
First sidelobe level (SLL)	φ = 0° (≈≤−10 dB) θ = 90° (≈≤−10 dB)	−9.3 dB (φ = 0°), −9.1 dB (θ = 90°)	−8.5 dB (φ = 0°), −9.4 dB (θ = 90°)
Polarization	Linear directional	Linear directional	Linear directional
Ground plane	Continuous, coextensive with the substrate	Lg × Wg = 16.5 × 22 mm2	Same
Footprint (board)	As compact as possible	L × W = 16.5 × 22 mm2	—

compares the ADAS requirements in the 77–79 GHz band with the results obtained in this study. All mandatory specifications are satisfied, confirming the proposed antenna’s suitability for automotive radar. The table summarizes the target specifications, achieved values at 77 GHz and 79 GHz, and the corresponding figure references.

In this paper, we present the design, simulation, and fabrication of an ADAS antenna with a compact footprint of 22 × 16.5 mm² (5.7 λ₀ × 4.3 λ₀). The antenna is specifically tailored for forward-looking automotive radar operating in the 77–79 GHz band (AEB/ACC/FCW). For completeness, wideband S-parameter and gain responses are reported across 75–110 GHz, while detailed radiation performance is evaluated at the critical band anchors, 77 and 79 GHz.

2. The Proposed ADAS Antenna Design

ADAS antennas play a critical role in multiple functions, including radar systems, Vehicle-to-Everything (V2X) communication, and navigation. Among microwave transmission lines, the microstrip line is one of the most widely used due to its compactness and ease of integration. The design process requires careful consideration of materials, antenna geometry, and placement within the vehicle to optimize performance and comply with regulatory standards [24]. In radar systems, ADAS antennas must operate at high frequencies, such as 77 GHz, to provide high-resolution data for object detection and collision avoidance. The proposed microstrip antenna (MSA) is implemented on Roger’s substrate with a dielectric constant ${\epsilon }_r$ of 2.2, loss tangent (tan δ) of 0.0009, and a thickness of 0.762 mm. The antenna geometry was optimized through extensive simulations using CST Microwave Studio [25]. Figure 2 presents the top view and bottom view of the proposed design, while the optimum dimensions are summarized in

Table 2. Optimized dimensions of the proposed ADAS antenna.

Dimensions	Value (mm)	Dimensions	Value (mm)
Lsub = Lg	16.5	Wsub = Wg	22
l1	9.22	w1	14.54
L2	3.1	w2	4.17
L3	8.6	w3	5.18
L4	2	w4	0.4
dvia	0.35	W5	5.58

.

The antenna employs a rectangular patch precisely dimensioned to resonate at the target frequencies. It is fabricated on a substrate incorporating two shorting vias and is excited through a tapered transition. Figure 3 illustrates the simulated scattering parameters across 75–110 GHz, with particular emphasis on the 77–79 GHz band where S₁₁ < −10 dB and excellent return loss (~−50 dB) is achieved at both 77 and 79 GHz, which is critical for ADAS applications. The antenna demonstrates multiple resonances, as indicated by the recurring dips in the S-parameter curve, confirming its ability to operate effectively at several frequencies. The 77–79 GHz response originates from coupled cavity-slot/patch interactions: the etched rectangular slot supports an approximately λg/2 resonance along its length, while the shorting vias introduce a λg/4 shunt return that tunes and slightly splits the fundamental mode, producing the two in-band minima

2.1. Tapered Transition Design

To achieve optimal impedance matching between the feeding network and the antenna structure, a tapered transition is strategically employed [26]. This transition is critical in minimizing reflection losses and ensuring efficient energy transfer, particularly in high-frequency antenna systems where abrupt impedance discontinuities can severely degrade performance [27]. Two configurations are commonly used to transform from microstrip to Dielectric-Filled Rectangular Waveguide (DFRW) [28]. The first approach uses a multi-section Chebyshev impedance transformer. The second one consists of a linear taper that matches the microstrip line impedance with the DFRW impedance. In microstrip antenna (MSA) design, tapered transitions are widely adopted to enhance bandwidth, impedance matching, and radiation efficiency. In this work, both approaches were considered. While Chebyshev sections offer excellent narrowband matching, their stepping discontinuities at W-band introduce fabrication sensitivity and additional parasitic reflections. In contrast, a linear taper provides a monotonic impedance profile and smoother broadband matching across 75–110 GHz with lower tolerance sensitivity. Therefore, the linear taper was selected for the proposed design. The topology of the implemented transition follows the second approach, as illustrated in Figure 4.

Figure 5 presents the input impedance (Rin and Xin) as a function of frequency, with arrows marking 77 and 79 GHz, where Xin = 0 coincides with the S₁₁ minima and realized-gain peaks. The plot confirms the broad banding mechanism: the reactance (Xin) crosses zero at 77 and 79 GHz while Rin ≈ 50 Ω, validating the presence of true resonances aligned with the S₁₁ dips. Between these resonances, the linear taper behaves as a quasi-continuous transformer, while the shorting vias introduce shunt inductance that reduces the reactance slope, maintaining |Γ| at a low level across 77–79 GHz. Beyond this band, Xin increases, leading to a mismatch and explaining the observed multi-resonance behavior that results in the achieved −10 dB bandwidth.

The tapered transition of the proposed design is characterized by three main parameters: the transition length $w_4$, the width $l_3$, and the initial width $l_4$. Consequently, $l_4$ must be calculated to obtain the desired characteristic impedance [29]. This width is selected to achieve a characteristic impedance of 50 Ω, and by calculating the ratio (${\displaystyle \raisebox{1ex}{$l_4$}\!\left/ \!\raisebox{-1ex}{$h$}\right.}$) as “$h$” is the substrate height, it becomes easy to evaluate the value of $l_4$ by the following formula:

$${\displaystyle \frac{l_4}{h}}=\left\{\begin{array}{c}{\displaystyle \frac{8e^M}{e^{2M-2}}} for {\displaystyle \frac{l_4}{h}}<2 \hfill \\ {\displaystyle \frac{2}{\pi }}\left[B-1-\ln \left(2B-1\right)+{\displaystyle \frac{{\epsilon }_r-1}{2{\epsilon }_r}}\left\{\ln \left(B-1\right)+0.39-{\displaystyle \frac{0.61}{{\epsilon }_r}}\right\}\right] for {\displaystyle \frac{l_4}{h}}>2\hfill \end{array}\right.$$

(1)

In standard closed-form models for microstrip lines, two auxiliary parameters, M and B, are introduced, defined as

$$M={\displaystyle \frac{z_0}{60}}\sqrt{{\displaystyle \frac{{\epsilon }_r+1}{2}}}+{\displaystyle \frac{{\epsilon }_r-1}{{\epsilon }_r+1}}(0.23+{\displaystyle \frac{0.11}{{\epsilon }_r}}),$$

(2)

$$B={\displaystyle \frac{377\pi }{2Z_0\sqrt{{\epsilon }_r}}}$$

(3)

These parameters are used to compute the effective width and impedance of the tapered section. The equivalent waveguide width $a_e$ can then be obtained by equating the right-hand sides of Equations (1) and (4), and solving it ($l_e$ is the width of an equivalent waveguide) [30].

Here $a_e$ denotes the effective characteristic-admittance parameter of the microstrip section. It is defined from the local characteristic impedance Z₀ (W/h, ε_eff) as

$${\displaystyle \frac{1}{a_e}}={\displaystyle \frac{Z_0}{{\eta }_{0}h}}$$

$${\displaystyle \frac{1}{a_e}}=\left\{\begin{array}{c}{\displaystyle \frac{60}{{\eta }_{0}h}}\ln \left(8{\displaystyle \frac{h}{l_3}}+0.25{\displaystyle \frac{l_3}{h}}\right) for {\displaystyle \frac{1}{l_e}}<2\hfill \\ {\displaystyle \frac{120\pi }{{\eta }_{0}h(\frac{l_3}{h}+1.393+0.667 \mathrm{l}\mathrm{n}(\frac{l_3}{h}+1.444)}}-0.627{\displaystyle \frac{{\epsilon }_r}{\frac{{\epsilon }_r+1}{2}+\frac{{\epsilon }_r-1}{2}\frac{1}{\sqrt{1+12\frac{h}{l_3}}}}}\hfill \end{array}\right. for {\displaystyle \frac{1}{l_e}}<2$$

(4)

where ${\eta }_0=\sqrt{{\displaystyle \raisebox{1ex}{${\mu }_0$}\!\left/ \!\raisebox{-1ex}{${\epsilon }_0$}\right.}}\approx 377$ Ω is the free-space wave impedance.

$${\displaystyle \frac{1}{l_e}}={\displaystyle \frac{4.38}{l_2}}e$$

(5)

where $l_2$ is the width of the substrate, which is connected to the cavity as illustrated in Figure 4 and listed in Table 2. The effective permittivity of a line of width $w$ is approximated by the Hammerstad Jensen model [31]:

$${\epsilon }_{eff}\approx {\displaystyle \frac{2{\epsilon }_r+1}{2}}+{\displaystyle \frac{2{\epsilon }_r-1}{2}}\left({\displaystyle \frac{1}{\sqrt{1+12h/\omega }}}+0.04 {(1-{\displaystyle \frac{\omega }{h}})}^2\right)$$

(6)

At 77 GHz, the free-space wavelength is

$${\lambda }_0={\displaystyle \frac{c}{f}}\approx {\displaystyle \frac{3\times {10}^8}{77\times {10}^9}}\approx 3.90 \mathrm{m}\mathrm{m},$$

and the guided wavelength is ${\lambda }_g={\displaystyle \raisebox{1ex}{${\lambda }_0$}\!\left/ \!\raisebox{-1ex}{$\sqrt{{\epsilon }_{eff}}$}\right.}.$ For ${\epsilon }_r=2.2$ and practical $(\omega /h)$ in our stack, ${\epsilon }_{eff}$ lies near 1.6–1.8; taking ${\epsilon }_{eff}\approx 1.7$ gives ${\lambda }_g\approx 3.90/\sqrt{1.7}\approx 2.99 \mathrm{mm}$ and thus a quarter-wave taper length of:

$${\mathcal{l}}_t\approx {\lambda }_g/4\approx 0.75 \mathrm{mm} (\mathrm{at} 77 \mathrm{GHz}).$$

2.2. Impact of Short Pins on Antenna Performance and Design

Shorting vias (or pins) are essential in MSA designs, particularly for high-frequency applications in the 75–110 GHz band [32]. They improve performance by minimizing parasitic effects, enhancing electrical connectivity, and creating vertical conductive paths within the substrate. This reduces radiation losses and increases antenna efficiency and gain [33]. The placement and diameter of these vias strongly influence wave propagation and overall electromagnetic performance, making their optimization critical for high-frequency operation. Figure 6 illustrates the impact of via spacing on the reflection coefficient. Introducing shorting vias flattens the reactance curve Xin(f), producing a continuous (~−15 dB) matching region across 77–79 GHz. In contrast, the case without shorting vias exhibits a more reactive (capacitive/inductive) input impedance and poor low-frequency matching. The optimum via spacing results in the deepest S₁₁ minima at both band anchors while preserving the overall bandwidth.

Figure 7 compares the surface current distributions with and without shorting vias. Without vias, the current flows more uniformly across the patch, producing broad but weakly defined regions of high and low current density. In contrast, when vias are introduced, the current becomes strongly concentrated near the via locations, forming localized high-density regions. Each shorting via behaves as a shunt inductive path to ground, which compensates for the distributed slot/microstrip capacitance. A widely used approximation for the partial inductance of a vertical via of length $h$ and barrel diameter $d$ as in Equation (7):

$$L_{via}\approx {\displaystyle \frac{{\mu }_{0}h}{2\pi }}[\ln \left({\displaystyle \frac{4h}{d}}\right)+1] \mathrm{H},$$

(7)

with ${\mu }_0=4\pi \times {10}^{-7} \mathrm{H}/\mathrm{m}$, $h$ in meters, $d$ in meters. With two identical vias in parallel, the effective shunt inductance is $L_{\mathrm{s}\mathrm{h}}\approx L_{\mathrm{v}\mathrm{i}\mathrm{a}}/2$ (neglecting minor spreading/pad effects). The AC resistance is small at W-band (copper, $\sigma \approx 5.8\times {10}^7$ S/m), with skin depth $\delta =\sqrt{2/\left(\omega {\mu }_0\sigma \right)}\approx 0.76$ μm at 77 GHz; $R_{\mathrm{v}\mathrm{i}\mathrm{a}}\approx \mathcal{l}/(\sigma 2\pi (d/2) \delta )$ is $(10 \mathrm{m}\Omega )$ and negligible versus $\left|X_L\right|=\omega L\sim {10}^2 \Omega$.

This results in a stable −10 dB matching band. The presence of vias therefore produces a more complex and spatially varied current distribution, directly contributing to improved impedance matching and bandwidth control.

The etched slot is necessary to confine and steer the field under the patch, flatten the input reactance slope, and together with the shorting vias realize the dual-resonance match across 77–79 GHz. The |E| field maps reveal slot-induced cavity confinement beneath the patch, which helps moderate Xin. The slight antinode shifts observed between 77 and 79 GHz correspond to the S₁₁ minima and Xin ≈ 0 crossings, explaining the stable matching in this band.

A slot is integrated into the patch to engineer the surface-current path and phase, reshaping the active aperture toward the end-fire direction and thereby tightening the main beam and increasing the realized gain while maintaining the required S₁₁ match.

3. Results and Discussion

The simulated radiation efficiency is calculated as follows [34]:

$${\eta }_{rad}=1-{Loss}_{copper}-{Loss}_{dielectric}-{Loss}_{mismatch}$$

(8)

where ${Loss}_{copper}$, ${Loss}_{\mathrm{d}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{c}}$, and ${Loss}_{\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{c}\mathrm{h}}$ are losses in the copper conductors, dielectric, and mismatch. The simulated radiation efficiency accounting for conductor, dielectric, and mismatch losses exceeded 85% at 77 GHz owing to the low-loss PCB stack-up. Figure 8 plots radiation efficiency and end-fire realized gain at θ = 90°, φ = 0° over the 75 to 110 GHz range, with efficiency mostly ~78–92% and a smooth, quasi-periodic gain profile. At the ADAS band edges, performance is strong: ≈9.9 dBi/≈85% at 77 GHz and ≈11.2 dBi/≈80% at 79 GHz. The stable traces around 77–79 GHz indicate consistent aperture excitation rather than loss-driven variation, confirming robust forward directivity with low dissipation.

Figure 9a–d present the normalized total electric field $\left|E\right|$ (dB) in IEEE spherical coordinates: θ-cut (φ = 0°) and φ-cut (θ = 90°) at 77/79 GHz. The patterns exhibit a stable, forward-pointing end-fire main beam with the radiation maximum near θ ≈ 90°, and a consistent −3 dB beamwidth (~40–60°) across frequency. Sidelobes are well suppressed, and back radiation is low, yielding a high F/B ratio. Together with the angular reference in Figure 9e, these results confirm wideband stability suitable for automotive radar/ADAS.

Figure 10 documents the fabricated PCB and the WR-12 launcher-based metrology used for both S₁₁ and far-field characterization, including TRL de-embedding to the microstrip feed plane. Far-field realized gain was measured in a fully anechoic chamber using the comparison (substitution) method with a calibrated WR-12 standard-gain horn. The AUT was placed beyond the Fraunhofer distance, polarization was aligned, and the VNA reference plane was TRL de-embedded to the microstrip feed. Realized gain was obtained from $G_{AUT}\left(\theta ,f\right)=G_{ref}\left(f\right)+\left[S_{21,AUT}\left(\theta ,f\right)-S_{21,ref}\left(f\right)\right]$, with measurement uncertainty of ±0.5 dB (amplitude) and ±1° (pointing). Prototype fabricated on εr = 2.2, h = 0.762 mm laminate with copper thickness (35 μm). Shorting vias: finished drill 0.25 mm, drilling precision ±0.05 mm, plating thickness ≥ 25 μm (hole wall). Slot/patch etch tolerance ≤ ±0.02 mm. Surface finish ENIG (Ni 3–6 μm, Au 0.05–0.1 μm).

The measured S₁₁ in Figure 11 confirms robust matching centered at 77/79 GHz and higher frequencies on band, with minor simulation-to-measurement deviations attributed to fabrication tolerances and material dispersion. Figure 12 overlays the measured and simulated co-polar realized gain at 77 GHz and 79 GHz plotted versus φ ∈ [−180°, +180°] at θ = 90°. The measured main-lobe peaks (≈9.9 dBi and ≈11.2 dBi) and ≥18 dB cross-polar suppression corroborate the simulations; pointing uncertainty is ±1° for both measurement and simulation. Figure 9 presents the normalized total field $\left|E\right|$ (dB) cuts, each normalized to its own peak: (a, b) θ-plane ($\varphi =0°$) at 77/79 GHz, and (c, d) φ-plane ($\theta =90°$) at 77/79 GHz. Unless stated otherwise, φ-plane patterns are plotted as $\left|E\right|\left(\varphi \right)$ at $\theta =90°$, ensuring consistency with Figure 9c,d and symmetry about $\varphi =0°$. Unless stated otherwise, φ-plane patterns are plotted as $\left|\mathrm{E}\right|(\phi )$ at θ = 90°, ensuring consistency with Figure 9c,d and symmetry about φ = 0°.

Table 3. Comparisons of performance with related research publications.

Ref.	Antenna Type	Size mm2	${\mathbf{f}}_{\mathbf{c}}$ GHz with RL	Peak Gain (dBi)	Efficiency %	−10 dB BW (GHz; %FBW)	Fabrication
[13]	(ME-dipole) antenna array	$(11.8\times 8.5)$	77 GHz, RL < −10 dBi	12.338	90	NR	PCB
[15]	Patch loaded with I-shaped elements	$(23\times 12.82)$	79 GHz RL < −10 dBi	10.74	NR	(operating range 77–81 reported)	PCB
[14]	SIW antenna	$(20\times 20)$	76.5 GHz RL < −10 dBi	18.5	68.72	NA	PCB
[36]	Horn antenna	$(10.16\times 22.86)$	77 GHz RL < −15 dBi	12.2	NR	NA	WR−12 waveguide
[37]	Horn and slotted waveguide	$(61.67\times 9.5)$	77 GHz, RL < −10 dBi	17.5	79.4	NA	3D printing
[38]	Wideband cavity-slotted waveguide antenna	NA	79 GHz RL > −10 dBi	12.36	62	NA	PCB
[39]	Cavity-slotted waveguide, arrayable	NA	(76–81) GHz	14.5	NA	NA	PCB
The Proposed Design	SIW-tapered MSA	$(22\times 16.5$)	77, 79 GHz RL < −10 dBi	11.5	85	(2.6%) covers 77–79	PCB

benchmarks the design against related work, highlighting the compact PCB-compatible footprint, high efficiency, and competitive gain without multilayer or waveguide-only complexity. These results validate the antenna’s suitability for ADAS applications requiring focused and reliable forward-looking signal transmission [35].

4. Thermal Impact on Antenna Geometry and Resonance in ADAS Applications

Radar applications operating in the 77 GHz,79 GHz band, particularly for automotive ADAS systems, are often deployed in environments subject to wide temperature variations (e.g., –40 °C to +105 °C), depending on weather, engine proximity, and mounting location [40]. These thermal fluctuations induce dimensional changes in the antenna structure due to the coefficient of thermal expansion (CTE) of the conductive and dielectric materials, which in turn affect the antenna’s electrical performance, most notably the resonance frequency [41].

To analytically estimate the thermal impact, the standard linear expansion model was applied as in Equation (9) [42]:

$$\left({\displaystyle \frac{∆L}{L_0}}-1\right).f_0\approx f_r\left(T\right) ∆T.L_0.\propto = ∆L$$

(9)

where α (alpha) is the thermal expansion coefficient (for copper, =$7\times {10}^{-6} {°\mathrm{C}}^{-1})$, L₀ is the initial patch length, ∆T is the change in temperature, and f₀ is the design frequency at nominal conditions (77.08 GHz at 25 °C).

Assuming a rectangular patch antenna with dimensions L₀ = 23.6 mm and W₀ = 15.0 mm,

Table 4. Effect of temperature on antenna dimensions and resonance frequency.

Temperature (°C)	25	45	65	85	105
Patch Length (mm)	23.6	23.608	23.616	23.624	23.632
Patch Width (mm)	15	15.005	15.01	15.015	15.02
Fr1 (GHz)	77.08	77.054	77.028	77.001	76.975
Fr2 (GHz)	79.06	79.036	79.019	79.01	79.08
Gain at Fr1 (dBi)	10	9.4	8.7	8.1	7.5
Gain Fr2 (dBi)	11.5	11	10.64	10.21	9.85

summarizes the dimensional changes and corresponding resonance frequency shift and gain across various operating temperatures. Resonant drift is fitted linearly, $f\left(T\right)=f_0+kT$(25–105 °C), yielding $R^2\ge 0.98$. Pre-tuning from $T_{\mathrm{c}\mathrm{a}\mathrm{l}}$ to $T_{\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}\mathrm{t}}$ uses $f_{\mathrm{c}\mathrm{a}\mathrm{l}}^{\star }=f_{\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}\mathrm{t}}-k (T_{\mathrm{t}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}\mathrm{t}}-T_{\mathrm{c}\mathrm{a}\mathrm{l}})$; applied at 77.79 GHz for 85/105 °C, this aligns the $\left|{\mathrm{S}}_{11}\right|$ minima within fit tolerance and ≤±0.5 dB gain change as in Figure 13.

For completeness, Figure 14 reports the temperature dependence of the match on both band anchors. Panels (a) and (b) sweep S₁₁ around 77 GHz and 79 GHz, respectively, for 25–105 °C. In both cases, the resonance remains within ≈±0.05–0.10 GHz of the anchor, and the return loss is ≥10 dB at/near the anchor across the sweep, confirming robust matching at both 77 and 79 GHz.

To counteract the thermally induced resonance shift, the antenna was pre-tuned to 77.08 GHz and 79.06 GHz at 25 °C to ensure precise alignment with the target 77.00 GHz and 79 GHz under elevated temperature conditions (~85 °C). This thermal pre-tuning strategy effectively mitigates performance degradation caused by dimensional expansion, allowing the antenna to maintain optimal impedance matching and gain stability across the intended thermal range. Such an approach is critical for ensuring robust and reliable operation in ADAS radar applications, where high thermal variability is common.

5. Integration and Simulation of the Proposed Antenna in ADAS-Enabled Environments

5.1. Initial Evaluation of Vehicle-Mounted Antenna in ADAS

After designing the 77, 79 GHz microstrip antenna and obtaining promising simulation results, a realistic automotive environment was constructed to assess its practical performance. The primary objective was to evaluate the antenna’s effectiveness within ADAS and V2X communication systems [43]. As illustrated in Figure 15, electromagnetic simulations were conducted using the WIPL-D ProCAD 2024 Demo [44], where the antenna was integrated onto a detailed vehicle body model to reflect real-world installation conditions.

Effect of the roof plate at location D. The metallic roof forms a nearby conducting surface that is excited by the antenna’s near field, producing image currents. The resulting field can be interpreted as $E_{dir}\left(\theta ,\phi \right)+\rho \left(\phi \right)E_{dir}\left(\theta ,\phi \right)e^{j2khcos\theta }$, which biases radiation toward the forward hemisphere and can tilt the main lobe by a few degrees while slightly modifying beamwidth/directivity.

The simulations highlighted the influence of the car’s physical structure on radiation characteristics such as beam direction, realized gain, and the presence of side lobes. This in situ evaluation provided critical insights into the antenna’s real-world behavior, underlining the necessity of environment-aware design approaches in automotive radar systems.

Moreover, the positioning of antennas for applications like Adaptive Cruise Control (ACC) and Automatic Emergency Braking (AEB) must follow specific guidelines to ensure optimal performance and safety [45]. An unobstructed field of view is essential, typically covering horizontal angles of 20–40° and vertical angles of 10–15° to reliably detect surrounding vehicles and obstacles [46]. Standard installation height ranges from 0.4 to 0.8 m above ground level (usually near the front bumper or grille), with a slight downward tilt of 1–2 degrees to enhance ground-level object detection and forward-looking radar coverage.

Tx–Rx isolation is ultimately set by the module/array (spacing, filtering/duplexing, shielding); the proposed single-pol element, co-pol dominated with low sidelobes, readily supports high isolation when orthogonally paired or modestly spaced, with absolute isolation determined by the final module architecture.

Outside 77–79 GHz, higher-order slot/taper modes provide ~3–6 dB antenna-level discrimination; with standard 20–40 dB front-end filtering, this exceeds 23–46 dB isolation suitable for Tx–Rx isolation.

5.2. Virtual Scenario for Smart Road Deployment

The system architecture illustrated in Figure 16 serves as a representative model of an ADAS-enabled vehicular communication network. It captures the dynamic interaction between vehicles and infrastructure through integrated Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication links. These connections are facilitated by On-Board Units (OBUs), Roadside Units (RSUs), a centralized control node, and Global Navigation Satellite System (GNSS)-based positioning mechanisms [47].

In the illustrated scenario, vehicles A, B, and C actively participate in both V2V and V2I communications. Their OBUs continuously exchange real-time data with nearby RSUs, enabling the dissemination of traffic conditions, hazard alerts, and other critical situational awareness information. Vehicle D, as it approaches a signalized intersection, benefits from coordinated data inputs received from both the RSU and the central control node, thereby enhancing its decision-making processes for safer navigation.

Conversely, vehicle E exemplifies the implementation of high-speed V2V communication, leveraging GNSS-based localization to maintain safe inter-vehicle spacing and accurate lane tracking. This interconnected communication ecosystem enables a high level of cooperative awareness, forming the operational basis for key ADAS features such as adaptive cruise control, automatic emergency braking, and real-time collision prediction.

5.3. Impact of Non-Embedded Antennas on Smart Road Performance

Without embedded antenna integration within smart road infrastructure, advanced ADAS encounter significant limitations in terms of coverage precision and communication latency. In such scenarios, vehicles must rely solely on their OBUs and internal sensors, which reduces their situational awareness, especially in complex urban areas or locations with weak signal coverage.

Figure 17 illustrates a common failure scenario caused by antenna system defects or misconfigurations. These vulnerabilities lead to communication breakdowns and delayed responses in safety-critical functions.

For instance, Vehicle D approaching an intersection may fail to receive timely alerts from the RSU, impairing its ability to anticipate hidden hazards or react to cross-traffic. Similarly, Vehicle E, which depends primarily on V2V communication, might suffer from data delays due to interference or temporary signal loss. Such latencies compromise features like automatic braking and proactive lane shifting.

Furthermore, the absence of embedded antennas results in discontinuous and uncoordinated coverage across the road network, leading to awareness gaps between vehicles and diminishing the reliability of cooperative perception and semi-autonomous functions. Therefore, integrating high-performance microstrip antennas into road infrastructure is essential for robust ADAS operation. This underlines the urgent need for integrated solutions that merge advanced antenna technologies with smart roadway systems.

6. Conclusions

This work demonstrated a compact, single-layer W-band PCB antenna tailored to the 77–79 GHz ADAS band and validated via a launcher-based measurement workflow (WR-12 TRL de-embedding with anechoic-chamber substitution). The prototype achieves a robust match (S₁₁ ≤ −10 dB across 77–79 GHz) and end-fire radiation with realized gains of ≈9.9/≈11.2 dBi at 77/79 GHz, ≥18 dB cross-polar suppression in the main beam, and well-controlled sidelobes. The measured −3 dB beamwidths are ≈36° in the θ-cut (φ = 0) and ≈11.6° in the φ-cut (θ = 90°), consistent with full-wave simulations. A controlled thermal sweep from 25 to 105 °C yields < 100 MHz resonance drift while maintaining ≥ 10 dB return loss, confirming thermal robustness. Combined with high simulated radiation efficiency (≈85% at 77 GHz; ≈80% at 79 GHz) and gain flatness of ~1.3 dB across 77–79 GHz, the element is well-suited for short- and medium-range automotive radar. Future research will extend this design to higher mmWave bands (up to 220 GHz) for next-generation 6G, IoT, and V2X applications.