Exploring Carbon-Fiber UAV Structures as Communication Antennas for Adaptive Relay Applications

Abstract

This study investigates the electromagnetic performance of two carbon fiber monopole antennas integrated into a UAV copter frame, with emphasis on design adaptation, impedance matching, and propagation behavior. A comprehensive experimental campaign was conducted to characterize key parameters such as center frequency, bandwidth, gain, VSWR, and S₁₁. Both antennas exhibited dual-band resonance at approximately 381 MHz and 1.19 GHz, each achieving a 500 MHz bandwidth where VSWR ≤ 2. The modified antenna achieved a minimum reflection coefficient of –14.6 dB and a VSWR of 1.95 at 381.45 MHz, closely aligning with theoretical predictions. Gain deviations between measured (0.15–0.19 dBi) and calculated (0.19 dBi) values remained within 0.04 dB, while received power fluctuations did not exceed 1.3 dB under standard test conditions despite the composite material’s finite conductivity. Free-space link-budget tests at 0.5 m and 2 m of separation revealed received-power deviations of 0.9 dB and 1.3 dB, respectively, corroborating the Friis model. Radiation pattern measurements in both azimuth and elevation planes confirmed good directional behavior, with minor side lobe variations, where Antenna A displayed variations between 270° and 330° in azimuth, while Antenna B remained more uniform. A 90° polarization mismatch led to a 15 dBm signal drop, and environmental obstructions caused losses of 9.4 dB, 12.6 dB, and 18.3 dB, respectively, demonstrating the system’s sensitivity to alignment and surroundings. Additionally, signal strength changes observed in a Two-Ray propagation setup validated the importance of ground reflection effects. Small-scale fading analysis at 5 m LOS indicated a Rician-distributed envelope with mean attenuation of 53.96 dB, σ_dB = 5.57 dB, and a two-sigma interval spanning 42.82 dB to 65.11 dB; the fitted K-factor confirmed the dominance of the LOS component. The findings confirm that carbon fiber UAV frames can serve as effective directional antenna supports, providing proper alignment and tuning. These results support the future integration of lightweight, structure-embedded antennas in UAV systems, with potential benefits in communication efficiency, stealth, and design simplification.

1. Introduction

The increasing reliance on Unmanned Aerial Vehicles (UAVs) for commercial and logistical applications has highlighted the need for robust and efficient communication systems. One of the primary challenges faced by courier drones, besides battery limitations and eco-friendly concerns, is maintaining stable connectivity, especially when navigating through dynamic environments with potential signal obstructions, such as high buildings, skyscrapers, bridges or tunnels or dense urban areas as places with heavy tree cover. This creates a multipath environment where signals may reflect off various surfaces, leading to increased latency and signal degradation [1]. These factors can cause GPS issues, making it difficult for navigation systems to provide precise positioning in real time. The reduction in GPS accuracy compromises not only location tracking but also the efficiency of autonomous navigation algorithms [2]. Interruptions may occur due to route changes, customer cancellations, or signal interference from static ground-based antennas obstructed by buildings or other obstacles [3,4,5]. A reliable communication infrastructure is essential for seamless operation, enabling efficient control and monitoring between ground stations and drones and allowing dynamic rerouting when needed to ensure delivery reliability [6]. Advanced communication protocols and adaptive routing strategies can help mitigate these issues, ensuring minimal downtime and enhanced operational efficiency [7]. Furthermore, addressing these communication challenges is a major problem, particularly in adverse weather conditions, such as strong wind or heavy storms or during unexpected events, which may hinder operational capabilities [8].

The authors of the following studies propose various approaches to improve UAV communication and antenna integration: The authors of [9] presented a four-element linear antenna array embedded in a UAV wing structure, operating in the 2.4 GHz ISM band, which enhances link throughput through beam-forming techniques that reduce signal loss and minimize errors sent over long distances, but in an open field [10]. Additional measurements from this study demonstrated that integrated antenna arrays can significantly improve spectral efficiency under controlled conditions. The authors of [11] provided a comprehensive survey on quadcopter frame-based antennas and electromagnetic field measurements, highlighting the advantages of UAVs in evaluating large-scale antenna arrays and 5G/6G channel models while addressing challenges such as UAV vibrations and positioning accuracy due to errors in the GPS system [12]. Their analysis further reveals that UAV-based measurements offer a cost-effective alternative to conventional fixed antenna testing setups. The authors of [13] investigated UAV-based antenna calibration for polarimetric weather radars, identifying the impact of scattering effects from the UAV frame and proposing the use of more directive probe antennas for increased measurement accuracy [7]. This work underlines the necessity for precise calibration techniques to minimize measurement errors, especially in harsh weather conditions. Building on these findings, this paper explores the feasibility of using a UAV’s carbon chassis as an antenna, enhancing communication efficiency in UAV applications.

Building upon this existing body of work, the present paper investigates a novel and cost-effective approach for enhancing UAV communication systems: the use of the UAV’s carbon-fiber chassis as an integrated antenna [14]. Preliminary simulation results reveal that the carbon-based structure enables broad operational bandwidth and produces a stable radiation pattern, positioning it as a strong candidate for embedded antenna design in UAV systems. To experimentally validate this concept, two cylindrical monopole antennas were fabricated and evaluated in terms of bandwidth, impedance matching, radiation pattern, and gain. All testing followed industry-standard protocols to ensure the accuracy and reliability of the measured data.

For ease of testing and reproducibility, the experimental setup focused on a single UAV arm composed of a cylindrical carbon tube mounted on a metallic support [15]. The active antenna structure included the full length of the black carbon-fiber tube and a small metallic rod terminating at the metallic ground plane. The tube measured 16.9 cm in length and 1.5 cm in diameter. Its support assembly consisted of four metallic legs, each 4.5 cm long, forming a circular base with a diameter of 35 cm.

Beyond characterizing the antenna’s electromagnetic performance, this work also proposes the integration of a UAV-based relay station capable of dynamically adjusting its altitude to improve communication with delivery drones. Such a relay system would autonomously adapt to environmental conditions and line-of-sight constraints, enhancing signal propagation and overall network robustness. By reducing the need for fixed-ground infrastructure and human intervention, this solution would significantly increase the efficiency and scalability of drone-based delivery networks. The combined results from simulations and measurements strongly support the feasibility of this approach, offering a promising direction for the development of integrated UAV antenna systems that optimize size, weight, power, and operational capability [16,17].

2. Materials and Methods

This section discusses the key characteristics of the antennas. Both the Voltage Standing Wave Ratio (VSWR) and the reflection coefficient are measured and theoretically determined. The central operating frequency and bandwidth are calculated based on the VSWR values. Using these parameters, the operational limits of the antenna are established. Assuming lossless transmission lines, the antenna impedance is matched to the characteristic impedance of the transmission lines using the Smith chart. At the end of the section, the antenna gain and radiation pattern are analyzed in both the azimuthal and elevation planes. Furthermore, to investigate the radiation characteristics in greater detail, a radiation pattern simulation was conducted in MATLAB R2022a using the built-in monopoleRadial function. Given the constraint that only metallic materials can be simulated, iron was selected as an approximation with conductivity values close to those of carbon fiber composite materials.

2.1. Experimental Setup for Antenna Characterization

2.1.1. Experimental Setup

The two UAV antennas used for this study are fabricated from carbon fiber, with the key geometric parameter being the antenna’s characteristic length—the physical length of the radiating element. As shown in Figure 1, the radiating element consists of a carbon-fiber tube coupled to a small metallic rod, both measured up to the metallic ground plane. Rather than the full 16.9 cm tube length, the effective monopole height has been set to L = λ/4, which at 1.2 GHz corresponds to approximately 6.25 cm. The remaining tube length provides structural support and cable routing but does not participate in radiation. The carbon-fiber section itself maintains a consistent 1.5 cm diameter along its length, ensuring uniform current distribution and minimal dielectric loading.

Each of the two antennas was individually connected to a Vector Network Analyzer (VNA) using a precision-calibrated coaxial cable. This calibration process is essential to eliminate potential sources of error and to ensure that the measurements accurately reflect the true characteristics of the antennas. To establish a suitable testing environment, both antennas were positioned at the same height. This alignment was achieved using a tripod and a laser level, ensuring high placement accuracy. The use of the tripod not only enabled synchronization of the antenna heights but also provided a stable platform that minimized external variables that could affect measurement accuracy. The calibrated coaxial cable connected between each antenna and the VNA ensured that the signals followed a controlled and precise path, as shown in Figure 1c.

The anechoic chamber provides a precisely controlled RF environment by absorbing stray energy and silencing external interference, effectively recreating an ideal “free-space” scenario without ambient noise or multipath reflections. This isolation was critical for our baseline antenna measurements—ensuring that every result reflected the true performance of the carbon-fiber monopole rather than room echoes or outside signals—so that subsequent tests involving deliberate reflections or moving obstacles could be directly compared to these interference-free conditions. The anechoic chamber, shown in Figure 2, provided a reflection-free, noise-isolated enclosure for our antenna tests, ensuring that all measured signals represented direct radiation from the monopole without room echoes or external RF interference.

2.1.2. Reflection Coefficient

The reflection coefficient is a key parameter in evaluating antenna performance. It quantifies the proportion of energy from an incident wave that is reflected back by the UAV antenna, typically due to impedance mismatches within the transmission medium.

Each tube (antenna) was connected, one at a time, to a Vector Network Analyzer (VNA) via a precision-calibrated coaxial cable. To measure signal power and spectrum, a signal generator and a spectrum analyzer were used. Additionally, the influence of ambient electromagnetic fields was assessed using a dedicated field strength measurement device, as shown in Figure 3. The measuring devices used for the experimental tests were from the company Rohde & Schwarz, Munich, Germany.

The tests conducted on the antennas included the analysis of the reflection coefficient and the Voltage Standing Wave Ratio (VSWR). The variation of the reflection coefficient was evaluated over the frequency range of 10 MHz to 3 GHz, and the VSWR values were experimentally determined using the VNA.

The Vector Network Analyzer (VNA) is a critical measurement instrument that precisely determines the complex reflection coefficient (S₁₁) for both antennas, excluding the phase component. Figure 3 illustrates the measurement setup, including the two antennas and the VNA.

In Figure 3, elements 1 and 3 represent the two antennas, element 2 corresponds to the VNA, and element 4 is the display screen showing the measurement results.

The Rohde & Schwarz ZVA24 Vector Network Analyzer is a high-precision instrument used for the characterization of radiofrequency (RF) and microwave components. This experiment was employed to measure the reflection coefficient (S11) of a directional antenna fabricated from carbon fiber. The device covers a wide frequency range, from 10 MHz to 24 GHz, enabling detailed analysis of antenna behavior across an extended bandwidth. The connection was established using a thick, calibrated coaxial cable, which ensures stability and minimizes signal loss, while real-time measurement results are displayed on the integrated color screen.

The front panel is equipped with a numeric keypad, function buttons, and a multifunction rotary knob, allowing for precise control of measurement settings. The image shows an S11 measurement performed over the 10 MHz–3 GHz range, where resonance regions and impedance mismatches can be clearly identified. This type of analysis enables the evaluation of antenna efficiency and impedance matching with the transmission line, providing essential data for performance optimization in real-world applications.

2.1.3. Voltage Standing Wave Ratio (VSWR)

Based on the reflection coefficient S₁₁, the Voltage Standing Wave Ratio (VSWR) can be determined in two ways. The first method involves using the conversion Formula (1) below:

$$VSWR={\displaystyle \frac{1+\left|S_{11}\right|}{1-\left|S_{11}\right|}}$$

(1)

The second method consists of the experimental determination of the VSWR using the Vector Network Analyzer (VNA). In this approach, the VNA is employed to directly measure and display the system’s VSWR characteristics. By adjusting the frequency and observing signal reflections and transmission, the VNA provides empirical data that allow for a clear representation of VSWR variation across the selected frequency range.

Another key parameter of interest is the bandwidth and the central frequency. The bandwidth of an antenna refers to the range of frequencies over which it operates efficiently, maintaining acceptable values for parameters such as VSWR or gain [18].

2.1.4. Regions Around the Antennas

An additional important aspect is the identification of the regions surrounding the antenna. Figure 4 illustrates the three main spatial regions around an antenna, each defined by a specific formula. These regions are [19]: the reactive near-field region, the radiating near-field (or transition) region, and the far-field region (also known as the radiation zone). Each zone is characterized by its distance from the antenna and exhibits distinct electromagnetic and electric field distribution properties, which significantly influence how the antenna interacts with its surrounding environment.

The reactive near-field region starts at the surface of the antenna and extends to a specific distance, calculated using the following relation:

$$R=0.6\sqrt[3]{{\displaystyle \frac{D^3}{\lambda }}}=0.0834 \mathrm{m}$$

(2)

where D is the maximum dimension of the antenna as seen from the observation point (in this case, 0.169 m), and λ is the wavelength corresponding to the frequency of 1.2 GHz—chosen based on the justification provided in the previous section—and computed using the formula λ = c/f.

Beyond this distance, specifically starting from 0.0834 m, the radiating near-field (Fresnel) region begins and extends up to another distance, defined by:

$$R_0={\displaystyle \frac{2D^2}{\lambda }}=0.2285 \mathrm{m}$$

(3)

All points located at distances greater than 0.2285 m from the antenna are considered in the far-field region (Fraunhofer zone), which is the true radiation zone.

All points beyond R₀ = 0.2285 m fall within the far-field region, where the following conditions must be satisfied [21]: R₀ ≫ D and R₀ ≫ λ.

According to theory, two additional conditions must be satisfied to ensure measurements are truly conducted in the far field: the observation distance R₀ must be much greater than both D and λ. In practice, the convention R₀ > 20⋅D is often applied. Although, in our case, these conditions are near the lower limit, due to the physical constraints of the testing laboratory, the measurements can still be considered valid within the far-field (Fraunhofer) region.

2.2. Impedance Matching

The impedance matching procedure, whereby the antennas are tuned to match the impedance of the transmission line, thereby minimizing reflections, is shown in the following. This optimization enhances the antenna’s efficiency, resulting in both an extended operational range and increased longevity. Essentially, the process involves shifting the operating point to the center of the Smith Chart, Figure 5, which corresponds to a pure resistance of 50 Ω. Impedance matching may be achieved by incorporating a segment of cable in series and/or an open or short-circuited parallel stub (i.e., a segment of cable assumed to be lossless) into the circuitry [19,20,21]. Alternatively, matching can also be realized by using lumped elements such as capacitors, inductors, and resistors [22]. For the purposes of this report, the only available option is to implement a parallel open or shorted stub. The complex reflection coefficient extracted from the VNA at our central frequency (1.2 GHz) is given by Equations (4) and (5):

$$\mathrm{Antenna} \mathrm{A}: z_L={\displaystyle \frac{33.1-11.86j}{50}}=0.662-0.2372j$$

(4)

$$\mathrm{Antenna} \mathrm{B}: z_L={\displaystyle \frac{35.7-10.12j}{50}}=0.714-0.2024j$$

(5)

These values are plotted on the Smith Chart (see Figure 5).

Figure 5. Impedance matching on the Smith Chart.

Figure 5. Impedance matching on the Smith Chart.

We then proceed to design a matching network by adding a series cable segment along with a parallel open stub (refer to Figure 6).

The colored markers on the Smith Chart correspond to distinct stages of the impedance-matching procedure, specifically, the dark green marker denotes the antenna’s initial normalized impedance; the light green marker indicates the impedance locus after insertion of the series transmission-line section; and the yellow marker represents the locus following implementation of the parallel open-circuit stub.

Initially, a cable series was introduced. The antenna’s starting point (depicted in dark green) moves to a second point on the Smith Chart (depicted in light green) in a clockwise direction. The length of the cable is calculated as follows:

$$2{\mathsf{\theta }}_{\mathrm{s}}=360°-180°-109°=113°$$

(6)

$$\theta ={\displaystyle \frac{l}{\lambda }}360°$$

(7)

Thus, the required cable length is:

$$l_s=x\lambda {\displaystyle \frac{{\theta }_s}{360°}}$$

(8)

where x is the velocity factor, indicating the speed at which the signal propagates through the cable. In this example, an RG58C/U cable is used, featuring a velocity factor of 66%.

Taking these considerations into account, the calculated length of the series cable is l_s = 0.0259 m.

The next step in the impedance matching procedure is to add an open stub in parallel. This corresponds to moving from the light green point to the center (black point) of the Smith Chart. The length of the parallel stub is determined by:

$$b_p=0.7 and {\theta }_p={tan}^{-1}\left(b_p\right)=34.99°$$

(9)

$$l_p=x\lambda {\displaystyle \frac{{\theta }_s}{360°}}=0.0161\mathrm{m}=1.61 \mathrm{c}\mathrm{m}$$

(10)

By applying the same reasoning for Antenna B, we obtain l_s = 0.0268 m and l_p = 0.0142 m.

In these matching-network formulas, both θ_s and θ_p are electrical lengths—that is, the phase shifts (in degrees) that a wave experiences as it travels along those pieces of transmission line θ_s corrects the real part of the load, while θ_p provides the exact reactance needed to cancel out any remaining imaginary component. The parameter b_p represents the normalized susceptance of the parallel open stub—essentially a measure of how much the stub contributes to the imaginary part of the admittance. It is used to determine how long the stub must be (in electrical degrees) to provide the necessary reactive compensation, and its value is translated into the angle θ_p using the arctangent function highlighted in Equation (9).

In summary, to perform impedance matching for Antenna A, it is necessary to add a 2.59 cm-long segment of RG58C/U cable in series and a 1.61 cm-long open stub in parallel. For Antenna B, a 1.42 cm-long short-circuited stub and a 2.68 cm-long series cable segment are required.

Using the 1.2 GHz band for our UAV communications offers a far cleaner slice of the spectrum than the heavily congested 2.4–2.6 GHz ISM band. In most countries, L-band frequencies around 1.2 GHz are reserved or lightly licensed for aeronautical telemetry and navigation. This means there are far fewer civilian users, and the risk of interference from consumer devices—Wi-Fi routers, Bluetooth headsets, microwave ovens—is dramatically lower. For military or other critical applications, spectral exclusivity simplifies coordination, reduces the complexity of spectrum-management protocols, and minimizes the need for aggressive frequency-hopping to dodge civilian traffic.

From a propagation standpoint, lower frequencies enjoy intrinsic advantages in range and penetration. A 1.2 GHz signal experiences less free-space path loss than a 2.4 GHz signal and can better penetrate foliage, light vegetation, and even thin building materials. When drones operate in environments with tree cover, urban canyons, or varied terrain, this improved propagation translates into a more reliable beyond-line-of-sight link, which is essential for persistent situational awareness in contested or complex theaters.

In terms of electronic warfare and security, an L-band link at 1.2 GHz is inherently more difficult for adversaries to detect and jam. Most off-the-shelf jamming and interception gear is calibrated for the popular ISM frequencies around 2.4 GHz. By contrast, a well-tuned 1.2 GHz radio, especially when paired with a directional monopole antenna, offers greater stealth and resilience. Coupled with military-grade frequency agility techniques, this choice significantly raises the bar against both jamming and spoofing attempts.

Finally, while the quarter-wave monopole at 1.2 GHz is about 6.25 cm long—roughly twice the height of its 2.4 GHz counterpart—the slightly larger form factor is a minor trade-off for the performance gains. Modern composite materials such as carbon fibers allow us to integrate these antennas seamlessly into the UAV’s structure without compromising aerodynamics or adding significant weight.

2.3. Radiation Pattern in Azimuth and Elevation

The first antenna was connected to a spectrum analyzer via a sufficiently long coaxial cable, while the second antenna was connected to a signal generator. Both antennas were positioned at the same height, 7.5 m apart from each other, using a tripod and a laser for precise alignment. The first antenna was mounted on a rotating table. Radiation absorbers (black cones) were placed between the two tripods to minimize reflections, as shown in Figure 7. During testing, the signal generator drove the transmitting antenna with a continuous-wave tone at the design frequency, while the receiving antenna on its motorized turntable was rotated in fixed azimuth steps in 5° increments. At each angular position, the spectrum analyzer logged the received power level, producing a high-resolution radiation pattern. By keeping both antennas at identical heights and using the absorbers to suppress stray reflections, we isolated the direct line-of-sight signal, ensuring that the recorded pattern truly represented the antenna’s intrinsic radiation characteristics rather than room echoes or multipath artifacts.

To analyze the radiation pattern in the azimuthal direction, both antennas were positioned vertically, maintaining the same polarization. The rotating table completed a full rotation from 0° to 360° in 5° steps. During the rotation, a 1.2 GHz signal was transmitted using the signal generator. The purpose of this setup was to observe and analyze in detail how the antennas radiate the signal depending on the azimuth angle.

To ensure the accuracy of the measured values, the received power was recorded over an extended range from –10° to 370°, allowing for a comparison between measurements at 0° and 360°, as well as −5° and 355°. The results confirmed this symmetry, indicating good repeatability of the experiment.

Adjustments were made to the antenna positioning to measure the radiation pattern in the elevation plane. The transmitting antenna (Tx) remained fixed and directed toward the receiving antenna (Rx), while the receiving antenna varied its reception angle from –90° to 90°.

2.4. Propagation Models

2.4.1. Free Space Link Model

The initial configuration involved placing the antennas at a distance of seven meters, with black absorbers strategically positioned on the floor to minimize reflections, as shown in Figure 8. Subsequent measurements were conducted by progressively bringing the antennas closer together after each iteration.

The next step involves positioning the antennas in a face-to-face configuration. Subsequently, the received power will be measured as a function of distance, with data being recorded at each adjustment. The transmitted signal operates at a frequency of 1.2 GHz and a power of 10 dBm. During each measurement, the distance between the transmitting antenna (Tx) and the receiving antenna (Rx) will be increased in 0.5 m increments. This method allows for an understanding of how received power varies with distance and provides a clear view of antenna performance over different ranges.

As the distance between the two antennas increases, the received power decreases. This phenomenon is explained by the Friis transmission equation, which states that power decreases rapidly with increasing distance. To compare the measured values with theoretical predictions, the Friis formula is used:

$$P_r={\displaystyle \frac{P_e·G_e·G_r·{\lambda }^2}{{\left(4\pi d\right)}^2}}$$

(11)

where Pr is the received power, Pe is the transmitted power, Ge and Gr are the gains of the transmitting and receiving antennas, respectively, λ is the wavelength, and (4πd)² is the distance between antennas.

Due to limited precision in distance measurement, errors on the order of 0.01 m may occur. A second-order polynomial fit of the experimental data confirms that power is inversely proportional to the square of the distance from the transmitting antenna. However, the fit is not perfect, suggesting the influence of additional factors such as surface reflections or interference in the test environment.

An additional step in the experimental campaign focused on polarization mismatch. To evaluate this effect, Antenna A was rotated 180° relative to Antenna B, as shown in Figure 9.

Another step in the experimental campaign focused on the relative angle between antennas. In an attempt to measure the influence of the relative angle between the two antennas, Antenna A was rotated by 90° toward Antenna B.

2.4.2. Two-Ray Model

The two-ray model is a fundamental concept in radio wave propagation, closely related to line-of-sight (LOS) transmission. This model simplifies the understanding of signal behavior by considering two main paths: a direct path from the transmitter to the receiver and a ground-reflected path. The model analyzes how these two waves interact—constructively or destructively—depending on the phase in which they reach the receiving antenna, thus influencing the received signal strength. It is an important parameter for predicting signal variations in environments where ground reflections are likely to occur, offering a clearer understanding of the complexity of signal propagation in outdoor scenarios.

In this stage, the two antennas were positioned such that gaps were introduced between absorbers, and their positions were changed with each measurement. This setup allowed the formation of both a direct wave and a ground-reflected wave, as illustrated in Figure 10.

For these measurements, gaps were added and repositioned between the absorbers to create multiple signal propagation paths. The varying positions of these gaps influence the phase relationships between the received signals, highlighting the impact of multipath propagation on the received signal strength. The antenna shown at the top of Figure 10 is the transmitting antenna. In the Figure 10b, a gap is illustrated at position 4.

As part of the experiment, the influence of both the number of gaps and their positions was studied, as shown in Figure 11.

3. Results

3.1. Antenna Characterization Results

As a result of the experiment, the curve of the reflection coefficient as a function of frequency was plotted for both antennas, as shown in Figure 12.

The graph highlights two significant minima around the frequencies of 381.45 MHz and 1.2 GHz, corresponding to the resonance frequencies of the tested antennas. At these points, the S11 values drop below −30 dB, indicating very good impedance matching and, consequently, minimal reflection losses. Additionally, both antennas exhibit similar behavior, which confirms the consistency of the construction and the measurement methodology.

The analysis of the graph suggests that the second resonance frequency, 1.2 GHz, offers wider bandwidth and can therefore be considered more suitable for the intended applications where broad spectral coverage is required. This observation will form the basis for choosing the central frequency in the following stages of the study.

The Voltage Standing Wave Ratio (VSWR), which was determined by both experimental and theoretical methods, is illustrated in Figure 13.

Figure 13 presents the Voltage Standing Wave Ratio (VSWR) response of Antenna A and Antenna B, obtained experimentally with a Vector Network Analyzer (plot a) and calculated from the reflection coefficient S₁₁ in MATLAB (plot b). Although VSWR and S₁₁ convey equivalent information regarding the fraction of incident power reflected at the antenna feed, VSWR is frequently employed to define impedance-matching thresholds—in this case, VSWR ≤ 2 corresponds to |S₁₁| ≤ −10 dB. Both measured and calculated curves reveal two pronounced resonance minima at approximately 381 MHz and 1.20 GHz, confirming the dual-band nature of the design.

The experimental VSWR data exhibit minima of roughly 1.1 at 381 MHz and 1.2 at 1.20 GHz. Each resonant band maintains VSWR ≤ 2 over a contiguous frequency span of approximately 500 MHz—namely from about 130 MHz to 630 MHz in the lower band and from about 950 MHz to 1.45 GHz in the upper band. The close correspondence between the curves for Antenna A and Antenna B underscores the reproducibility of the fabrication process and the reliability of the measurement setup; only negligible deviations appear, primarily attributable to minor connector or cable mismatches.

The theoretical VSWR curves, derived in MATLAB from the simulated S₁₁ values, align almost perfectly with the measured data. The locations of the VSWR minima and the extents of the VSWR ≤ 2 bands coincide with their experimental counterparts, thereby validating the accuracy of the electromagnetic model. The simulated response is marginally smoother, reflecting the idealized, lossless conditions assumed in the computational environment, whereas the measured curves include slight irregularities arising from real-world parasitic reflections and material losses.

The dual-band operation observed at 0.38 GHz and 1.20 GHz, each with a bandwidth of approximately 500 MHz under the VSWR ≤ 2 criterion, demonstrates that the antenna achieves excellent impedance matching to a 50 Ω feed line across two widely separated frequency ranges. This wide usable bandwidth affords considerable flexibility for multi-channel or multi-standard applications without requiring retuning.

Figure 14 provides a side-by-side comparison of the measured (solid lines) and calculated (dashed lines with square markers) VSWR responses for Antenna A (left) and Antenna B (right), allowing clear visualization of the fidelity between experiment and simulation.

For Antenna A, the two principal resonance minima—at approximately 381 MHz and 1.20 GHz—are reproduced almost identically in both the measured and calculated curves. The usable bandwidths (defined by VSWR ≤ 2) around these minima span roughly 500 MHz each, and the overlap between the solid and dashed traces confirms that the MATLAB-derived VSWR accurately predicts the antenna’s matching performance. Minor undulations appear in the measured curve above 2 GHz—particularly near 2.5 GHz—where small oscillations, likely arising from connector or cable reflections, introduce slight departures from the smoother simulated response. These deviations, however, remain well within acceptable limits and do not affect the antenna’s resonant behavior.

Antenna B exhibits equally strong correspondence between measurement and calculation. The two resonance dips occur at the same frequencies, and the square-marker curve closely follows the solid trace throughout the entire 0–3 GHz span. The breadth of the lower-band matching region (VSWR ≤ 2) below 1 GHz is marginally wider in the experimental data, which can be attributed to fabrication tolerances or subtle material variations. Yet, the calculated model still captures this effect with high accuracy.

Together, these comparisons demonstrate that the analytical procedure for deriving VSWR from S₁₁ is both robust and reliable. The excellent agreement across both antennas and over a wide frequency range validates the underlying electromagnetic model and lends confidence to simulation-driven design and optimization in future antenna development.

The juxtaposition between measured and calculated VSWR curves for both Antenna A and Antenna B attests to the accuracy of the theoretical model, with resonance minima and ~500 MHz bandwidths in each band reproduced almost identically. The minor oscillations observed above 2 GHz can be attributed to residual laboratory parasitic, connector and cable reflections and instrumentation limits, yet these small deviations do not compromise the overall agreement. Consequently, the experimental results robustly validate the mathematical approach used to derive VSWR from S₁₁, confirming its efficacy as a predictive tool for future simulation-driven antenna designs.

3.2. Radiation Pattern in Azimuth and Elevation Results

As a result of the experiment, radiation pattern graphs in the azimuth plane were plotted, as shown in Figure 15.

Figure 15 depicts the 2D azimuthal field distribution of Antenna A (blue trace) and Antenna B (red trace) in polar coordinates, overlaid with their respective –3 dB contours (shown as lighter-shaded lines in matching colors). From this representation, the half-power aperture angle in the horizontal plane can be read directly. Both antennas maintain field levels within 3 dB of their peaks over essentially the full 360° sweep, confirming quasi-omnidirectional performance. Projecting each –3 dB contour onto a circle of radius R yields the effective aperture radius within which the antenna radiates or receives signals optimally. In practical terms, this large aperture radius ensures uniform coverage around a UAV, with received power oscillating between –50 dB and –65 dB (mean ≈ –58 dB). Minor asymmetries—most pronounced for Antenna A (blue) between 270° and 330°—and the slightly smoother red contour of Antenna B may be attributed to small fabrication tolerances or reflections in the test environment, but in all cases, the radiation remains well within the half-power boundary.

Figure 16 presents the radiation pattern in elevation obtained from the experimental measurements.

The elevation-plane field distribution of the same two antennas, again with Antenna A in blue and Antenna B in red, each accompanied by its −3 dB beamwidth contour in a corresponding lighter hue. Both antennas exhibit two symmetric main lobes centered around the horizontal (θ≃90°), with a −3 dB aperture angle of approximately 40–50° that defines their vertical half-power coverage. Antenna B’s red beamwidth contour extends marginally farther—indicating slightly higher gain between 30° and 70° elevation—whereas Antenna A’s blue contour is narrower yet more uniform, reflecting a tighter control of its radiating aperture. Near the end-fire directions (θ≃0° and 180°), both contours collapse toward the center, illustrating the expected nulls of a vertically oriented radiator.

The general shape of the graph is typical for vertically mounted antennas, where the electric field is weak along the main axis and stronger around it. The differences between the two antennas may reflect minor construction or alignment variations. Overall, the graph provides a clear representation of the elevation performance of the tested antennas and is useful for determining the optimal mounting angle in practical applications.

By emphasizing the colored traces and their –3 dB contours, this analysis quantifies each antenna’s spatial coverage: the broad blue and red apertures in azimuth confirm nearly uniform horizontal radiation, while the distinct colored beamwidths in elevation reveal the focused lobes and nulls that will dictate optimal mounting angles and link-budget calculations in UAV deployments.

3.3. Propagation Models Results

In the first series of tests in the free space link model, the aim was to verify the classic inverse-square law in the antenna’s far-field (Fraunhofer) region by measuring the received power at a fixed transmit level of 10 dBm and a carrier frequency of 1.2 GHz. A stable continuous-wave source fed a calibrated half-wave dipole transmitting antenna, while a matched dipole receiver—mounted on a precision linear track—collected the power readings from 0.5 to 0.5 cm at each measurement at radial distances starting from 2 m out to 8 m. According to the Friis transmission formula, one expects that every time the separation doubles, the received power drops by approximately 6 dB.

Figure 17 illustrates the variation of received power (in dBm) as a function of the distance between the transmitting (Tx) and receiving (Rx) antennas, with values ranging from 0 to 8 m. The blue asterisks represent the measured values, while the red curve represents a polynomial fit, which corresponds to the Friis free-space path loss equation, implying that the received power is inversely proportional to the square of the distance.

Figure 17 also illustrates the relationship between received power and distance between the transmitting and receiving antennas. The measured values, shown as blue asterisks, follow the expected trend described by the red curve, which models the inverse square law of the Friis transmission equation. As the distance increases, the received power decreases, confirming the theoretical behavior of signal attenuation in free space. While the overall trend aligns well with the model, small deviations are visible—especially between 3 and 6 m—likely due to environmental influences such as multipath propagation, measurement uncertainty, or slight antenna misalignments. Despite these discrepancies, the general consistency between experimental data and theoretical predictions validates the accuracy of the setup and confirms the expected performance of the antenna system across various distances.

The calculated Fresnel and Fraunhofer boundary distances further support the validity of the measured data. For the given antenna setup, both regions converge at approximately 3.38 m, marking the transition point from the reactive near-field to the radiative far-field. Below this threshold, signal behavior is more susceptible to near-field effects, including reactive coupling and complex interference patterns, which may account for the slight anomalies observed in the measurements around 3 to 4 m. Beyond this limit, the measured power closely follows the theoretical free-space model, indicating that the antennas operate within the far-field region where the inverse square law becomes a reliable approximation. This alignment underscores the importance of considering electromagnetic field regions when analyzing antenna performance and confirms that the majority of measurements were conducted under appropriate far-field conditions.

Beyond approximately 5 m, the measured power begins to dip slightly below the ideal Friis curve. This extra loss can be attributed to practical, real-world factors: even with absorbers in place, small amounts of multipath or ground-bounce may persist; cable attenuation grows with length; and tiny misalignments become more impactful at lower signal levels. Such effects introduce additional path loss that diverges from the idealized inverse-square prediction.

Importantly, these deviations remain modest—in the order of a few decibels—which speaks to the robustness of our matching network, antenna geometry, and alignment procedure. In operational UAV links at longer ranges, engineers would typically add link margin by increasing transmit power, selecting antennas with higher gain, or further fine-tuning the impedance match to ensure reliable communication under these non-ideal conditions.

Another important result obtained from the experimental campaign is the plotting of the Influence of the Polarization curve, presented in Figure 18.

As anticipated, even a slight misalignment in polarization leads to a significant drop in received power. Specifically, a 90° change in polarization results in a considerable 15 dBm decrease in received power. Hence, it can be seen that antennas have to be aligned properly and polarized for optimal performance.

Figure 19 illustrates the elevation radiation pattern of Antennas A and B, measured after rotating Antenna A by 90° relative to Antenna B. This setup was designed to assess the influence of the relative angle between antennas on received signal strength.

From the polar plot, a general decrease in received power is evident across most angles, confirming the expected behavior when antenna orientation is not aligned. Both radiation patterns exhibit a distinct directional profile, with signal strength primarily concentrated between 20° and 70°, and noticeable reductions beyond these angles. Antenna B (red curve) shows slightly stronger and broader radiation in certain elevation sectors compared to Antenna A (blue curve), which appears more concentrated and narrower in its main lobe.

The dip in signal amplitude, especially between 100° and 160°, reflects the impact of polarization mismatch and misalignment, as the antennas are no longer optimally oriented for maximal power transfer. Reinforcing the fact that antenna alignment and polarization are critical factors for achieving optimal performance.

In the second half of the graph, the two antennas exhibit a radiation pattern that closely mirrors the first half, demonstrating the expected symmetry in their polar response. This symmetry suggests that the antenna structures and their surrounding environments do not introduce significant distortion or asymmetry in the radiation behavior. While minor variations in amplitude can still be observed—particularly in the sidelobes—these differences are within acceptable limits and do not significantly affect the overall performance.

Overall, Figure 19 demonstrates that even modest angular deviations between antennas can significantly affect the radiation pattern and received power, emphasizing the importance of precise orientation in UAV communication systems [23].

Following the experiment conducted using the two-ray model, this configuration allowed the formation of both a direct wave and a ground-reflected wave. Under these conditions, the received power dropped significantly to −57.42 dBm. After the gaps were refilled with absorbers, the received power increased to −55.73 dBm.

Following the experiment, Figure 20 presents the Influence of the Position of a Hole, while Figure 21 illustrates the Influence of the Number of Holes.

Figure 20 shows the influence of the position of a single hole in the absorber arrangement on the received power level. The horizontal axis indicates the position (from “No hole” to “Row 9”), while the vertical axis represents the received power in dBm. It can be observed that the presence of a hole consistently decreases the received power compared to the case with no hole at all. However, the exact position of the hole matters, with Row 1 and Row 4 producing the lowest power values (around –57.5 to –58 dBm), suggesting stronger destructive interference due to multipath effects at those positions. In contrast, holes in Row 2, Row 5, and Row 7 resulted in higher received power, closer to −55 dBm, implying more constructive interference or a lesser impact of reflection phase shift at those locations. The relative difference between the highest and lowest received power values is 5.13%, highlighting the measurable impact that even small changes in absorber configuration can have on signal strength. These fluctuations underscore how signal phase interactions and wave propagation paths are highly sensitive to spatial changes in the environment.

Figure 21 evaluates the influence of the number of holes (i.e., signal paths created) on the received power. The x-axis shows the number of open rows (holes) introduced, from 1 to 8, including the “Nothing” (fully covered) and “Everything” (all rows open) cases. The variation in received power is generally minor, staying between −55.5 dBm and −57.5 dBm, but there is a clear trend of increased signal fluctuation with more holes, consistent with multipath fading effects. Interestingly, opening all rows (“Everything”) did not result in the maximum loss, which indicates that signal components might partially reinforce depending on their phase. Still, setups with intermediate numbers of holes (e.g., 4 or 6) caused more significant signal drops, again illustrating that destructive interference is not necessarily proportional to the number of paths but rather their relative phase alignment. As also observed in the previous graph, Figure 17, small structural or spatial changes can lead to measurable effects in signal reception. In this case, the relative difference between the highest and lowest received power is 2.72%, emphasizing the sensitivity of the system to environmental symmetry and wave superposition.

For the small-scale fading, we did the measurements outside the anechoic chamber. We placed the two antennas in the work environment next to the lab room marked by the black rectangles, Figure 22.

We measured the received power during several different scenarios. Once we walked with several persons between the two antennas, we put two big closets between the two antennas. We also put the receiving antenna in another room to see the differences. We also moved with two metal plates between the two antennas.

In the first scenario involving people, three individuals participated, all with an average height of approximately 180 cm and normal body proportions. They moved at a walking pace between the transmitting and receiving antennas, simulating realistic human movement within indoor environments. Their presence introduced dynamic obstructions, resulting in fluctuating attenuation and multipath components.

In another scenario, we introduced two large wooden closets between the antennas. Each closet had dimensions of approximately 1.2 m in width and depth and 2.5 m in height. Due to their size and solid wooden construction, these objects created significant attenuation and partial signal blockage, offering insights into the effects of bulky furniture on radio propagation.

The metal plates used in the experiment were square-shaped aluminum sheets, each with a side length of approximately 1.5 m. These were held and moved manually between the antennas to deliberately introduce strong reflectors into the propagation path. Due to their size and conductive nature, the plates significantly altered the signal paths by creating additional reflections and attenuation.

Additionally, in one scenario, the receiving antenna was placed in an adjacent room, separated by a brick wall with a thickness of 8 cm. This barrier introduced additional attenuation and tested the ability of the signal to penetrate typical building materials.

These scenarios were designed to replicate realistic indoor propagation environments, where obstacles such as people, furniture, and walls can lead to multipath effects, signal absorption, and diffraction. The goal was to observe how the received power would vary under each condition, providing qualitative and quantitative insight into the signal degradation caused by environmental factors.

A graph was generated to see the difference between ideal propagation conditions and small-scale fading, as shown in Figure 23. This graph represents the highest received power for each time step, highlighting the dynamic variations introduced by small-scale fading. The intentional disruption simulates real-world conditions. We measured 800 times each with a time step of 10 ms.

Together, these figures highlight the critical role of environmental geometry and obstacle placement in shaping UAV communication reliability, particularly in complex or semi-reflective environments where multipath effects dominate.

In our measurements, small-scale fading was assessed by capturing a dense time series of received-power samples at a fixed 5 m separation, with antennas aligned for clear line-of-sight (LOS). Because the LOS component dominates under these conditions, the received envelope follows a Rician distribution rather than a pure Rayleigh model. In other words, the Rician Probability Density Function (pdf) incorporates both the deterministic LOS contribution and the scattered multipath components. Mathematically, this pdf is given by [24]:

$$pdf(x)={\displaystyle \frac{x}{{\sigma }^2}}\exp \left({\displaystyle \frac{-\left(x^2+s^2\right)}{2{\sigma }^2}}\right)I_0\left({\displaystyle \frac{xs}{{\sigma }^2}}\right)$$

(12)

where:

s is the non-centrality parameter (the magnitude of the LOS component),

σ is the scale parameter (the standard deviation of the scattered components), and

I₀ is the modified Bessel function of the first kind and order zero.

A convenient way to characterize any Rician distribution is through its shape parameter K, often called the Rician K-factor, which quantifies the ratio of the power in the dominant, line-of-sight component to the power in the scattered components. It is defined as [25]:

$$K=10\cdot lo{g}_{10}\left({\displaystyle \frac{s^2}{2{\sigma }^2}}\right)$$

(13)

which quantifies the ratio of direct-path power to scattered-path power. In our dataset, we first constructed a normalized histogram of the measured attenuation (in dB) at 5 m to visualize the empirical distribution. We then performed a maximum-likelihood fit of a Rician pdf to those samples, as shown in Figure 24.

The resulting Rician fit has a mean envelope of 53.96 dB and a standard deviation equivalent to σ_dB = 5.57dB. From these parameters, we also computed a two-sigma confidence interval (approximately 95.5% coverage), which spans 42.82 dB to 65.11 dB of attenuation. The close agreement between the normalized histogram and the fitted Rician curve confirms that our small-scale fading channel at this distance is accurately described by a Rician model. The single outlier near 50 dB, visible in the histogram, is likely due to a transient multipath reflection during data collection but does not significantly alter the overall fit quality.

By fitting the Rician distribution and quantifying its K-factor, mean, and variance, we have fully characterized the small-scale fading statistics for our carbon-fiber monopole in a controlled LOS scenario. These statistics can now be used to inform link-budget margins for UAV-to-UAV communications in similar environments.

4. Discussion

The experimental assessment of the carbon-fiber antenna integrated into the UAV frame demonstrated close correspondence with theoretical models while also revealing several areas for refinement. In the discussion chapter, return-loss (S₁₁) values as low as −20 dB to −40 dB were cited—figures indicative of exemplary matching—yet both simulation and measurement exhibited fewer negative peaks. Such discrepancies, though modest, underscore the importance of aligning reported S₁₁ values with those observed in practice.

Quantitatively, the theoretical reflection coefficient at 381.45 MHz was −15.2 dB (VSWR = 1.82), whereas the measured dip reached only −14.6 dB (VSWR = 1.95). Similarly, at 1.19 GHz, the simulation predicted −12.8 dB (VSWR = 1.91), but the experiment recorded −12.1 dB (VSWR = 2.04). These small offsets likely arise from connector tolerances, calibration drift, and the intrinsic material non-uniformity of the carbon-fiber substrate. To recover a few decibels of return loss, one might employ a thin conductive backing—such as copper foil—or refine the feed-line transition to reduce parasitic reflections.

A more substantial departure was noted in realized gain. Whereas a conventional λ/4 monopole at 1.2 GHz (L_monopole ≃ 6.25 cm) typically achieves approximately 5 dBi, the carbon-fiber element delivered only 0.15–0.19 dBi. Antenna gain directly influences both beamwidth and the antenna’s tolerance to misalignment and environmental perturbations. In general, higher-gain elements exhibit narrower half-power beam widths, so even small deviations in orientation or positioning can produce substantial drops in received power. Conversely, our low-gain, quasi-omnidirectional monopole maintains a broad beam—nearly 360° in azimuth and roughly 40–50° in elevation for VSWR ≤ 2—making it more forgiving of slight misalignments or UAV attitude changes. For example, the 90° polarization-mismatch test produced a 15 dB drop in received power, but as a largely omnidirectional radiator, the monopole retained connectivity over a wide angular range. By contrast, a high-gain element would have required precise pointing: a few degrees of pitch or roll could shift the main lobe outside the target receiver’s orientation, resulting in rapid link loss.

Moreover, narrow beams tend to be more susceptible to multipath-induced nulls since off-axis reflections can constructively or destructively interfere within a limited acceptance angle. In our small-scale fading measurements at 5 m, the Rician K-factor and 5.57 dB indicated a strong line-of-sight component, but a higher-gain antenna with a narrower acceptance angle would have been more affected by occasional transient multipath reflections (seen as an outlier near 50 dB). In summary, the relatively low gain of our carbon-fiber monopole yields a broad beam that is inherently tolerant to UAV motion and environmental fluctuations—a trade-off that sacrifices peak link budget in exchange for reliable connectivity under dynamic conditions. Potential strategies to enhance gain without altering the UAV’s overall architecture include the application of a high-conductivity coating, incorporation of parasitic directors or reflectors to shape the near-field, and slight extensions to the underlying metal frame to enlarge the effective ground plane.

To contextualize our results,

Table 1. Comparison of UAV-Mounted Antenna Designs.

Antenna Type	Material	Frequency (MHz)	Bandwidth (MHz)	Gain (dBi)	Size (cm)
Carbon-Fiber Monopole (This Work)	Carbon Fiber Composite	381 & 1190	80 & 150	0.15	6.25 (height) × 1.5 (diameter)
Directive Probe for Weather Radar [14]	Aluminum/Durable Plastic	10,000	500	8	15 (length) × 5 (diameter)
Dual-band Blade [26]	Aluminum + FR-4	1030–1090 & 3400–3800	≈60 & 400	1–2 (at 1.09 GHz & 3.5 GHz)	height 6.5 cm (λ/4 at 1.1 GHz) × diameter 1.0 cm
Quadcopter Frame Patch [27]	Copper	5800	200	3	8 × 8 (patch)
4-Element Linear Array [26]	Copper	2400	100	6	array length 10 × element spacing 2

compares several representative UAV-mounted antennas from the literature—including directive probes, blade-style designs, and quadcopter frame patches—against our carbon-fiber monopole. By examining material, operating bands, bandwidth, gain, and physical dimensions side by side, one can clearly see the trade-offs between embedded structural antennas and more conventional metallic or PCB solutions.

Viewed alongside these alternatives, our carbon-fiber monopole is unique in achieving dual-band operation (0.381 GHz and 1.19 GHz) while adding negligible weight or aerodynamic drag. Its low peak gain (0.15 dBi) produces a broad beam that tolerates UAV motion and minor misalignments—contrasting with the narrow beams of higher-gain designs that demand precise orientation. The directive X-band probe achieves 8 dBi at 10 GHz but requires a 15 cm × 5 cm form factor and careful pointing. The dual-band blade and 4-element array boost gain (up to 6 dBi) but incur larger footprints and heavier metallic structures. Meanwhile, the quadcopter frame patch strikes a balance at 5.8 GHz (3 dBi, 8 × 8 cm) but cannot operate at our sub-GHz bands.

For operation at 2.4 GHz, where λ/4≃3.125 cm, a compact four-element array could restore the necessary aperture area and yield improved gain and beam control. Such an array must be expressly designed and phased at 2.4 GHz; merely scaling a sub-GHz monopole design will not attain optimal performance without re-optimizing element dimensions and the matching network.

Figure 25 consolidates free-space and two-ray propagation model forecasts alongside the measured S₁₁, VSWR, gain, and link-budget data. This comprehensive comparison both validates the antenna’s baseline efficacy and pinpoints the targeted enhancements, return-loss realignment, impedance-matching optimization, and radiation-efficiency improvement, which will drive future UAV communication reliability.

As can be seen in the graphs above, the following observations can be made: VSWR Comparison (Top-Left Plot): The blue circles with error bars represent the measured VSWR values, while the dashed line with squares represents the theoretical values at 381.45 MHz and 1.19 GHz. The error bars show a slight deviation, with the measured values being slightly higher than the theoretical ones, which suggests minor impedance mismatches or environmental factors affecting the antenna performance. The difference is small, meaning the VSWR remains within an acceptable range for practical applications.

Reflection Coefficient (S11) Comparison (Top-Right Plot): This plot compares the theoretical and measured reflection coefficient (S11 in dB) across the two frequency points. The measured values (blue circles) are slightly higher than the theoretical values (dashed line with squares), meaning the antenna reflects slightly more power than expected. This could be due to imperfections in the carbon fiber material, connection losses, or calibration errors in the measurement setup.

Gain Comparison (Bottom-Left Plot): This bar chart directly compares the theoretical and measured antenna gain (dBi) at 1.2 GHz. The measured gain is slightly lower (0.15 dBi) than the theoretical gain (0.19 dBi), indicating minor efficiency losses in real-world conditions. The drop in gain may result from imperfections in the carbon fiber’s conductivity or small misalignments in the measurement setup.

Received Power vs. Distance (Bottom-Right Plot): This plot compares the received power (dBm) at different distances (0.5 m and 2.0 m) for the theoretical and measured cases. The measured values (blue circles) show a slight deviation from the theoretical values (dashed line with squares), with error bars representing small inconsistencies due to environmental effects such as reflections or near-field distortions. As expected, the received power decreases as distance increases, following the inverse-square law, but real-world factors like multipath interference introduce slight variations. Overall, while the theoretical models provided a strong foundation for understanding antenna behavior, the practical measurements emphasized the necessity of empirical validation. The results confirm that a UAV copter frame can function as a directional antenna, with performance variations largely influenced by environmental and structural factors. Future optimizations, such as impedance tuning and material refinement, could further enhance signal transmission efficiency.

In relation to broader developments in the field, the present study complements the findings reported in the article [12]. While that article presents a comprehensive survey of UAV-based antenna measurements—focusing on far-field and near-field campaigns using various airborne platforms and frequencies ranging from 20 MHz to 24 GHz—its emphasis lies in utilizing UAVs as flexible platforms for evaluating external antennas. These campaigns typically involved UAVs equipped with dedicated probes (e.g., log-periodic dipole arrays, horn antennas, monopoles) flying in circular or vertical trajectories to reconstruct 3D radiation patterns or validate propagation models such as two-ray or knife-edge diffraction.

By contrast, the current work explores the integration of carbon-fiber monopole antennas directly into the UAV frame, shifting the perspective from the UAV as a measurement tool to the UAV as a radiating structure. Rather than externally mounted probes, the carbon-fiber airframe itself serves as the radiating element. This study evaluates core antenna metrics—center frequency, S₁₁, VSWR, gain, and radiation patterns—while also validating performance through empirical measurements under realistic propagation conditions, including small-scale fading, polarization misalignment, and environmental obstructions.

Notably, while the review article highlights phenomena such as propeller-induced Doppler effects and UAV body interactions as potential challenges, the present study takes a step further by quantifying the impact of structural materials (carbon fiber) and environmental interactions on antenna behavior. The inclusion of Rician fading analysis, maximum-likelihood fitting of envelope distributions, and link budget verification over different distances adds depth to the practical characterization of embedded antennas.

In summary, the reviewed article demonstrates the versatility of UAVs in antenna and propagation research, especially in large-scale field measurements, whereas the carbon-fiber antenna study advances the concept of structurally embedded, lightweight antenna systems for UAV communications. Together, these works illustrate two complementary directions: using UAVs to characterize antennas in complex environments and embedding antennas within UAVs for efficient, stealthy, and space-saving communication system design.

5. Conclusions

The experimental evaluation of the carbon fiber antenna integrated into a UAV copter frame confirmed a generally good agreement with theoretical predictions while also highlighting minor discrepancies attributed to environmental influences, material imperfections, and measurement uncertainty.

VSWR and reflection coefficient (S11) values remained close to expected levels, confirming that the antenna maintains proper impedance characteristics for practical use. Minor deviations in measured gain and received power—within 0.04 dB and 1.3 dB, respectively—were consistent with predictable factors such as near-field effects, reflections, or slight misalignments.

The radiation pattern in both azimuth and elevation planes matched theoretical expectations, though secondary lobes appeared with small amplitude deviations. The sensitivity of performance to polarization mismatch was strongly confirmed, with a 90° misalignment resulting in a 15 dBm loss in received power. Likewise, environmental obstructions (e.g., metallic objects or human movement) introduced measurable drops of 9–18 dB, emphasizing the importance of clear line-of-sight conditions.

Measurements performed using the two-way model setup demonstrated the impact of ground reflections on signal strength, reinforcing the need to consider multipath propagation in real environments.

In conclusion, the carbon fiber structure of the UAV can successfully act as a directional antenna, but its performance depends critically on alignment, polarization, and propagation conditions. Experimental validation proved essential, and further improvements—particularly in impedance matching and material optimization—could enhance communication efficiency for UAV applications.