Evaluation of Additive Manufacturing Techniques to Increase Drone Radar Detection Range

Abstract

The deployment of drones in civil airspaces has increased rapidly in recent years. However, their small radar cross-section hinders detectability by existing surveillance systems. This paper evaluates the feasibility of using additive manufacturing techniques with materials with enhanced electromagnetic properties to increase RADAR drone detectability. The examined methods include the production of propellers and landing gear with additive techniques and metal-doped plastics. Field experiments were performed measuring the radar detection range of an off-the-shelf drone with a 24 GHz RADAR module, and findings indicate that the use of specialized materials for drone manufacturing greatly improves drone detectability, increasing the detection range by a factor of 3.9 without changes to the drone design or inclusion of external electromagnetic reflectors.

1. Introduction

The rapid expansion of the drone market has revolutionized numerous sectors, especially in civil applications. Yet, this technological boom faces a formidable obstacle: the seamless integration of crewed and uncrewed aircraft in non-segregated airspaces.

Air traffic services (ATS) for crewed aircraft currently use both primary and secondary RADAR systems to manage airspace. Primary surveillance RADARs (PSR) detect objects by emitting a signal and receiving the reflection from the object [1]. In contrast, secondary RADARs rely on transponders to receive information from aircraft. While equipping all drones with transponders could potentially solve detection issues, this solution presents significant challenges because of the required power and the congestion of the electromagnetic spectrum around the frequency of 1090 MHz at which secondary radar replies are broadcast. In the case of uncrewed aircraft, regulations worldwide require that drones emit their position and identity using direct remote identification (DRID) systems (e.g., Commission Delegated Regulation (EU) 945/2019 or Code of Federal Regulations (USA) 14 CFR part 89) or network remote identification (NRID) interfaces (e.g., Implementing Regulation (EU) 2021/664). However, remote identification devices can fail, thus posing safety concerns. For this reason, detect and avoid (DAA) systems such as the State-of-the-Art ACAS sXu [2] still consider RADAR paramount to prevent mid-air collisions. Moreover, remote identification devices can be removed or switched off if civil drones are used for unauthorized and malicious purposes [3]. Hence, for counter-UAS systems, it is also crucial to integrate robust primary surveillance systems capable of detecting and neutralizing such threats.

The radar cross section (RCS) of an object measures how detectable the object is by RADARs, indicating the effective area that reflects RADAR signals back to the emitter, typically expressed in square meters (m²) or decibels per square meter (dBm²). Among other factors, RCS depends on the electrical size of the target (i.e., the physical dimensions compared to the wavelength) and the material they are made of (metallic targets reflect stronger echoes than dielectric ones). Hence, drones usually have low RCS because of their small size and the materials used in their manufacturing. These factors result in a highly dispersed and random RADAR signature, specifically Swerling-1 [4], complicating detection and tracking efforts. Moreover, PSRs typically operate in the L- or S-band, around 1.3 to 3 GHz, where the longer wavelengths and lower resolution make it even more difficult to detect small targets like drones. As a result, conventional PSRs stand little chance of reliably detecting such low-RCS targets, especially compared to specialized high-frequency drone detection systems. However, even at those high frequencies, the usual drone RCS is quite below the minimum RCS detectable by typical aeronautical radars, around 0 dBm² [5]. For instance, the RCS of a Class C2 drone (less than 4 kg max. take-off weight) at a typical millimeter-wave drone detection radar frequency of 36 GHz is around –10 dBm² [6]. Therefore, improving drone RCS is critical for their safe integration into increasingly crowded airspaces and reducing risks to civil aviation.

In [3], the authors review the different techniques used for drone detection. The authors classify the methods based on the sensor type, dividing them into acoustic, visual, radio-frequency (RF), and RADAR. While effective in specific use cases [7], acoustic, visual, and RF require dedicated sensors for drone detection and suffer from environmental noise, discouraging their implementation in broader surveillance networks. Further research is being conducted to improve the capabilities of RADAR systems to detect and classify small targets such as drones, by focusing on their specific characteristics [8,9] or implementing RADAR systems with heightened sensitivities [4,10]. During flight, the blades of the drones create specific micro-doppler signatures that can be used for detection and classification; however, these signals are still faint compared to other targets in aerospace surveillance [11]. These techniques could also benefit from an increased RCS of the target. Various researchers have tackled the improvement of the RCS of drones by using passive [12] or active [13] reflectors or by designing finely tuned metamaterials [14], achieving significant results.

These implementations require external additions to the drone that would impact its flight performance, reduce the effective payload, or require complicated manufacturing processes.

This paper focuses on a novel approach: using metal-doped materials to increase the RCS of the drone without impacting its design. This is done by modifying drone manufacturing methods with additive manufacturing. Doping a plastic matrix with a small amount of metallic microparticles increases the volumetric conductivity of the resulting compound and thus increases the radar signature without significantly increasing the density.

Therefore, this manufacturing process would contribute to the safe and efficient integration of drones into conventional airspace, a key focus of the European Drone 2.0 [15] strategy. This initiative aims to create a seamless framework that allows drones to operate alongside crewed aircraft while maintaining the highest safety and efficiency standards.

To demonstrate the hypothesis stated in the paragraph above, a study of the dielectric constants and loss factor of commercially available additive manufacturing materials has first been conducted using a microwave calorimeter. Then, models for an off-the-shelf drone kit were designed and manufactured with the materials previously characterized using 3D printing techniques. Those models were evaluated through simulations to assess their structural integrity based on the safety factor. To validate the findings, an experimental test flight was performed using a 24 GHz continuous wave monostatic RADAR in free space conditions to estimate the maximum detectable range and determine which material would optimize the drone’s RCS.

The structure of this paper follows a systematic approach, starting with a Materials and Methods Section detailing all materials used in the study, including specific equipment such as drones and software, along with an explanation of the methodology followed during the manufacturing process and the flight tests. The Results section presents the findings of these tests. This is followed by a discussion of the results, where the implications of the findings are analysed. Finally, the conclusions summarize the key takeaways of the project.

2. Materials and Methods

This section details the materials, tools, and procedures used throughout the study to evaluate the effectiveness of radar-reflective plastics for drone components. By combining 3D printing with material characterisation and structural simulations, this methodology enables replicable assessments of RCS, performed through detectability assessment. It includes plastic compositions, equipment specifications, and experimental flight designs, ensuring results are grounded in consistent environmental and operational conditions.

2.1. Types of Plastics and Filaments

The types of plastics used in this study include polylactic acid (PLA), acrylonitrile butadiene styrene (ABS), and polyethylene terephthalate glycol-modified (PETG), which are common materials for 3D printing. Additionally, this research also incorporates variations of these plastics doped with metallic microparticles to enhance detectability and improve the drone’s RCS. These include PLA doped with copper, PLA doped with iron, and magneto-detectable PETG, which is especially formulated for detection by magnetic detectors in food production environments. While the exact ratio of plastic to metallic particles is not disclosed by the manufacturers, the estimated compositions by weight are 42.55% copper for PLA doped with copper, 28.95% iron for PLA doped with iron, and 16.22% metallic content for magneto-detectable PETG. The ratio of metallic content in the plastic affects the electromagnetic properties of the material, expecting a higher RADAR reflection with higher ratios.

All materials used in this study are available as off-the-shelf filaments with the specific manufacturer and material type being: SMARTFILL green recycled PLA, yellow SMARTFILL ABS, red SMARTFILL PETG, light grey magneto-detectable SMARTFILL PETG, brown HTPLA PROTOPASTA copper composite, and dark grey PROTOPASTA PLA iron composite.

2.2. Microwave Calorimeter

A microwave calorimeter was used to study the loss factors and the dielectric constants of the filaments, specifically, the MW-DETA01. To do so, a small cylinder of 150 mm in length and 96.6 mm in diameter was printed with 100% infill as the dielectric constant depends directly on the infill [12]. Figure 1 shows the measurement setup.

The results from this initial assessment of the electromagnetic properties of the materials were used to guide the selection of specific filaments that could potentially improve the RCS of the drone. This strategy allowed for a thorough analysis of the most promising materials.

2.3. 3D Printer and Slicing Software

The parts were printed using a X1C 3D printer (Bambu Lab, Shenzhen, China) and sliced using Bambu Studio (Version 2.0.3.54, available at https://github.com/bambulab/BambuStudio/releases, accessed on 7 July 2025). The settings used for the slicing were based on the default Fine profile with surface ironing enabled. The most relevant settings are summarized in

Table 1. Bambu Studio slicer settings.

Setting	Value
Layer height	0.12 mm
Line width	0.5 mm first layer, 0.42 others
Ironing	Top surfaces
Wall generator	Arachne
Top/bottom layers	5
Wall loops	3
Infill	25%, gyroid
Nozzle temperature	195° PLA, 220° PETG, 240° ABS

.

The properties were chosen to prioritize part quality, strength, and similarity to the original drone parts. It is important to note that the part density (infill, walls, and top/bottom layers) affects the electromagnetic properties of the final part [12].

2.4. Drone Kit

The drone model kit is the PX4 Development Kit X500 V2 (Holybro, Hong Kong, China). This kit offers ease of part replacement and simple assembly, thanks to the drawings provided by the manufacturer. The total weight of the drone, including the battery, is 1.1 kg, placing it in Class C2 of the UAS classification by EASA.

2.5. 3D Models

Two 3D models were designed in Fusion360 (Version 2.0.20460, available at www.autodesk.com/products/fusion-360, accessed 7 July 2025), propellers and landing gear. The objective is to compare the behaviour of the new materials with that of the original drone kit, so the design closely mimics the original shape. Figure 2 displays the renders of the models.

The propellers and landing gear were subjected to stress simulations to ensure the structural integrity of the drone for safe flight testing. These simulations were validated by measuring the safety factor, which evaluates the strength of a manufactured model under the maximum probable forces it might encounter during operation.

Table 2. Mesh information to obtain the stress simulation.

Mesh Information	Parameter Value
Average element size	3% of the model size
Scale mesh size per part	No
Element order	Parabolic
Create curved mesh elements	No
Max. turn angle on curves (deg.)	60
Max. adjacent mesh size ratio	1.5
Max. aspect ratio	10
Minimum element size (% of average size)	20
Number of refinement steps	0
Results convergence tolerance (%)	20
Portion of elements to refine (%)	10
Results for baseline accuracy	Von Misses stress

contains the mesh settings for the simulations.

The parameters used in the simulation for the different materials are shown in

Table 3. Parameters of the plastic materials.

Parameters	PLA	PETG	ABS
Density (kg/mm3)	1.14 × 10−3	1.14 × 10−3	9.50 × 10−4
Young modulus (MPa)	2100.00	1523.00	2000.00
Poisson’s ratio	0.36	0.36	0.365
Yield strength (MPa)	26.94	23.00	17.00
Ultimate tensile strength	28.10	31.70	25.32

[16]:

As shown in Table 3, PLA doped with copper and PLA doped with iron are not included due to limitations in the available information regarding the properties of these plastic materials, so the values for a generic PLA material are used. Similarly, magneto detectable PETG is not listed, but it can be considered equivalent to standard PETG, as are the doped materials. Nevertheless, it is anticipated that copper-doped PLA and iron-doped PLA would exhibit a slightly lower resistance, attributed to the microparticles that make the models more brittle.

The simulation for the propellers is conducted by approximating the maximum thrust model to be twice the weight of the drone; the weight of the drone is 10.79 N, therefore, the simulated force is 21.58 N. This force is distributed across the four motors and two tips per propeller blade, resulting in approximately 2.69 N per propeller blade tip. The study is simplified by applying a load of 3N at the tips of the blades to analyse how the lift distribution affects the performance of the propeller and evaluate its strength. Figure 3 illustrates the exact points where the force is applied.

The landing gear was tested to validate that it can resist the forces exerted by the aircraft during take-off and landing. Figure 4 shows the setup for this simulation.

2.6. RADAR

To study the maximum detectable range of the drones with different materials, the DK-sR-1200 Radar Module (IMST, Kamp-Lintfort, Germany) was used. The RADAR works in Frequency-Modulated Continuous Wave (FMCW) mode at a 24 GHz carrier frequency, with a 230 MHz bandwidth, beginning at 24,008 MHz. The RADAR has an integrated patch antenna with linear polarization. This RADAR module has a ±65° azimuth and ±25° elevation ranges, conditioning the flight plan design for the field tests [17].

2.7. Flight Tests

Once the materials have been manufactured and confirmed to withstand flight conditions, a flight test has been conducted to evaluate and compare the maximum detection range of the RADAR for each configuration.

As illustrated in Figure 5, the setup involves a RADAR system mounted on a tripod at 1.35 m above the ground. The RADAR is inclined at an angle of 40° from the horizontal and aligned parallel to the drone’s intended flight path. This inclination helps minimize false detections from ground clutter.

The drone takes off and follows a simple trajectory towards a single waypoint. It ascends and travels along the direction of the RADAR beam to reach a final position located 58 m horizontally from the RADAR and 50 m vertically above the ground. This results in a slant range (line-of-sight distance) of approximately 75 m between the RADAR and the drone.

To validate the results of this experiment, the setup and materials from a previous study [12], which employed a RADAR reflector (Luneburg lens) to enhance drone detectability, are reused and serve as a benchmark for comparison. That study demonstrated a maximum detection range of approximately 60.5 m. By replicating similar conditions, this experiment aims not only to confirm the reliability of the current setup but also to compare the performance of newly manufactured materials with those used previously. If similar detection results are achieved, it supports both the validity of the setup and the effectiveness of the new materials.

3. Results

This section presents the results of the study on the plastic materials. First, the results of the dielectric constant and loss factor of the different materials are presented. Then the outcomes of the simulations of the 3D models, analyzed using the safety factor, as well as the results from the 3D printing process, are considered. Lastly, the findings from the experimental flight tests are provided.

3.1. Experimental Study in the Microwave Calorimeter

Figure 6a,b show the results of the dielectric constants and loss factors obtained after analyzing the cylinders from the materials described in Section 2.1.

Materials with higher dielectric constants can store more electromagnetic energy when exposed to microwave radiation [18]. High dielectric constants contribute to a larger RADAR cross section because the material reflects more incident energy back to the RADAR receiver. PLA doped with copper exhibited the highest dielectric constant of 5.27, indicating a superior ability to reflect electromagnetic energy. PLA doped with iron also showed a high dielectric constant of 4.21. Magneto detectable PETG had a dielectric constant of 3.63, moderate among the materials tested. The results show that the specialized materials have a higher dielectric constant when compared to common plastics, and thus, are expected to have a higher RADAR profile.

A high loss factor can be detrimental to detectability because it means the material dissipates more energy as heat rather than reflecting it. This reduces the amount of energy available to be re-radiated or scattered, which can decrease the RCS. PETG demonstrated the highest loss factor of 0.19, indicating significant energy dissipation. PLA had a loss factor of 0.12, which is substantial but lower than that of PETG. PLA doped with iron recorded a moderate loss factor of 0.06. PLA doped with copper showed a lower loss factor of 0.04, suggesting reduced energy dissipation compared to undoped PLA and PETG. Magneto detectable PETG had a relatively low loss factor of 0.03. ABS exhibited the lowest loss factor of 0.01, indicating minimal energy dissipation.

This test of the electromagnetic properties of the materials shows that the doped materials should have better detectability when compared to their non-doped counterparts. The impact these differences have on the drone’s detectability was evaluated through the field experiments detailed in Section 3.4.

3.2. Safety Factor Simulations

After designing the 3D models, it is necessary to make a simulation of the materials to ensure the structural integrity of the drone. The parameter used to validate integrity is the safety factor, which is the ratio between the maximum stress a structure or material can withstand and the stress it is expected to experience in regular operation.

Table 4. Maximum and minimum safety factor values for the propellers.

Safety Factor	PLA	PETG	ABS
Max. safety factor	15	15	15
Min. safety factor	2.01	1.88	1.39
Total displacement (mm)	22.9	32.0	24.3

and Figure 7 contain the results of the simulations for the propellers.

The values that were taken to validate the flight tests are between 1.2 and 1.5 for a protoflight verification approach [19]. Therefore, the three materials can be considered suitable for manufacturing and testing the propellers.

The safety factor values for the landing gear are present in

Table 5. Safety factors of maximum and minimum values of the landing gear.

Safety Factor	PLA	PETG	ABS	Carbon Fiber
Max. safety factor	15	15	15	15
Min. safety factor	2.15	1.83	1.39	15
Total displacement (mm)	9.6	13.3	10.1	0.1

. The results indicate that the landing gears are suitable for manufacturing following the same validation as for the propellers. As can be seen from Figure 8, PLA has the highest safety factor out of the plastic materials, followed by PETG and then ABS. The materials with the current design are not as strong as carbon fibre, but they are still suitable for their intended use.

Based on the results obtained, it can be concluded that both the propellers and landing gear are safe for use. However, it is highly advisable to protect the landing gear with foam on the horizontal tube that contacts the ground to enhance durability and prevent potential damage during landings. These parts are obtained through additive manufacturing techniques, which are often more susceptible to fractures and delamination over time compared to injected materials. This is because additive manufacturing builds parts layer by layer, which can lead to weak interfaces between layers. In contrast, injection moulding produces parts with a homogeneous structure, minimizing the risk of delamination and generally resulting in higher mechanical strength and durability. While injected materials can still experience issues like delamination under certain conditions, the overall reliability and performance of injection-moulded parts typically surpass those produced through additive manufacturing. Overall, the simulations indicate that the 3D models can be safely manufactured and used.

3.3. Manufacturing Results

During propeller manufacturing, PETG proved unsuitable due to its inability to accommodate the complex geometry required for the design. ABS, while offering some flexibility, exhibited high surface roughness that negatively affected aerodynamic performance. More critically, both PETG and ABS failed to meet the minimum safety standards for flight during experimental drone testing, rendering them unsafe for use and leading to their exclusion from further experimentation.

In contrast, PLA presented no manufacturing difficulties. As a result, only plain PLA, PLA doped with copper, and PLA doped with iron were selected for production. Figure 9 presents the weight of each individual propeller.

There are four propellers in total, resulting in a weight increase of 44 g for copper-doped PLA, 28 g for iron-doped PLA, and 4 g for standard PLA. Given the drone’s maximum payload of 2.5 kg, this weight increase is minimal and practically negligible.

The propellers were sanded to reduce imperfections from the 3D printing process and to minimize aerodynamic resistance. Liquid glue was used to securely assemble the model. Figure 10 shows the weights of the different landing gears.

In this case, the only material exceeding the weight of the original half landing gear is copper-doped PLA by 8 g. This additional mass is negligible given the drone’s maximum payload capacity.

3.4. Experimental Flight Test Results

This section presents the results of radar-based detection experiments conducted on various drone configurations to assess their RCS and overall detectability. The primary objective of this analysis is to evaluate how different materials and design modifications affect the radar visibility of drones, thereby enhancing their suitability for airspace monitoring and traffic management applications.

The detections provided by the RADAR system were pre-processed to reduce clutter. Initially, detections were filtered within a ±10° azimuth from the flight path and then manually reviewed to confirm they corresponded to the target. The analysis begins with the results from the original drone, followed by a comparison with those obtained using the doped materials.

Flight tests were not conducted for PETG and ABS due to their low experimental safety, which rendered them unsuitable for safe operation. As a result, the presented results focus only on the original drone, the PLA doped with copper, the PLA doped with iron, and the configuration using the Luneburg lens.

In some graphs, two distinct paths can be observed, which results from the second repetition of the experiment. Conversely, in other cases, only a single path is visible, attributed to the alignment and synchronization of the repetition.

Figure 11 illustrates the drone’s expected flight path in green, with its maximum detection range of 14 m, highlighted in red. At 110 s, the radar detects the drone’s return and reads the data that is plotted on the right side of the graph.

Figure 12 shows the detection distance for the iron-infused PLA. The drone was detected up to 42 m when the original landing gear was replaced by one with iron PLA. When the drone also utilized PLA doped with iron PLA propellers, the maximum detectable distance increased to approximately 45 m, adding 3 m to the maximum detection distance.

Once again, it has been demonstrated that doped materials provide better results than the original drone model, and it can be easily predicted that the 3D models manufactured with these doped materials yield higher detection ranges, validating the results obtained with the microwave calorimeter.

The findings from Figure 13 demonstrate that the most detectable material is PLA doped with copper. In Figure 13a, the maximum detectable distance was 49 m only with the landing gear and 55 m including the propellers, and a 6 m difference when using copper-doped propellers.

Figure 14 contains the results for the detection distance when the drone was equipped with the Luneburg lens.

Figure 15 shows the maximum detection range for each of the materials during testing. Results closely align with previous studies using the same Luneburg lens to increase detectability, with a previous maximum detection range of 60 m [12].

The results obtained indicate a significant improvement in the radar cross-section, which increases the drone’s detectability, thereby fulfilling the project’s primary objective. Additionally, the alternative materials offer similar detectability as the Luneburg lens, but with considerably less weight and greater aerodynamic efficiency.

Improvements in RCS can be extrapolated from improvements in the maximum detection range, as the RCS is proportional to the fourth power of the maximum detection distance if all other variables are constant [20], as shown in Equations (1) and (2):

$$\sigma \propto R_{max}^4$$

(1)

$${\displaystyle \frac{{\mathsf{\sigma }}_1}{{\mathsf{\sigma }}_0}}={\left({\displaystyle \frac{R_{\mathit{max}1}}{R_{\mathit{max}0}}}\right)}^4$$

(2)

Table 6. Improvements in RCS as a factor and dB.

Configuration	Factor	dB
Iron-doped PLA landing gear	80.00	19.08
Iron-doped PLA landing gear and propellers	115.55	20.67
Copper-doped PLA landing gear	149.06	21.76
Copper-doped PLA landing gear and propellers	237.20	23.77
Luneburg lens	255.00	24.08

summarizes the improvements in RCS with the different configurations. The improvements in RCS are significant, obtaining an improvement of 19.08 to 23.77 dB in RCS by changing the materials, without redesigning the parts or including heavy reflectors.

4. Discussion

The results of this study show that additive manufacturing with doped plastic materials can substantially improve the detectability of small drones. The second experimental flight, which was the only flight test performed, demonstrated measurable increases in detection range when copper-doped and iron-doped PLA materials were used. These findings are consistent with the trends observed in the microwave calorimeter analysis, which identified materials with higher dielectric constants and lower loss factors as more reflective and therefore more detectable by radar.

Among all the tested configurations, the drone equipped with copper-doped PLA components, specifically in the landing gear and propellers, achieved the highest detection range of 55 m. Iron-doped PLA configurations also showed improvement, reaching 45 m. In comparison, the original undoped drone was only detected at a maximum range of 14 m. These results confirm that the material composition of drone parts plays a significant role in increasing radar visibility.

The Luneburg lens, used as a benchmark for high detectability, has a maximum detection range of 60 m and an increase of 24.08 dB in RCS, close to the results obtained with the copper-doped PLA (23.77 dB). However, it added significantly more weight to the drone. While the doped PLA components weighed a maximum of 60 g, the Luneburg lens weighed 214 g. This difference of over three times the weight provides a substantial advantage. Using lighter materials allows more payload capacity for additional equipment such as sensors, cameras, or larger batteries, improving mission flexibility and flight efficiency.

Although no direct radar cross-section measurements were taken, the changes in RCS can be extrapolated from the maximum detection range, which serves as a reliable performance indicator in realistic outdoor conditions. The improvements provided by doped propellers suggest that future research could explore further optimizations in part design or surface finish. The consistent contribution of the landing gear to detection improvements indicates that component positioning also affects overall detectability.

These results validate the potential of using doped materials as a lightweight and cost-effective solution for improving radar detection. This approach offers a practical alternative to heavier and more complex reflectors, supporting future efforts to integrate drones into controlled airspace more safely and effectively. Additional studies could investigate alternative doping substances, new geometric designs, or hybrid solutions to further improve performance while maintaining structural efficiency.

5. Conclusions

This research explored the use of additive manufacturing techniques with doped plastic materials to enhance drone detectability. The study combined material characterization, structural simulations, and experimental flight testing to assess the effectiveness of copper- and iron-doped PLA components.

The findings confirm that these doped materials improve detectability while maintaining low weight and structural viability. Compared to conventional detection–enhancing solutions, such as the Luneburg lens, doped components offer a practical and lighter-weight alternative that does not compromise aerodynamic performance. The materials tested in this paper increased the maximum detection range by a factor of 3-3.9, depending on configuration and material. This increase corresponds with an RCS increase of 19 to 23 dB, providing similar results when compared with the lens (24.08 dB). Future work should explore the effectiveness of other materials, such as composites or custom blends of plastics and resin enhanced with additives to improve their electromagnetic response.

Overall, the paper highlights the potential of integrating radar-enhancing properties directly into drone components through design and material selection. This approach supports the development of scalable, cost-effective, and airspace-compatible solutions for improving the visibility of small drones in radar-based detection systems.