Strategic Horizons --
Abstract
The proliferation of unmanned aerial systems (UAS) has reshaped industries ranging from logistics to public safety, demanding rigorous pilot training and strategic equipment selection. This study applies the hybrid SWARA-MOORA-3NAG decision-making model to identify the most suitable drones for initial training. By evaluating operational cost, autonomy, maneuverability, wind resistance, and portability, the analysis ranks the DJI Mini 1, Mini 2, and Mini 3 Pro as optimal platforms. These models balance affordability, ease of use, and performance, aligning with expert recommendations and real-world training conditions. The findings underscore the necessity of structured selection criteria to enhance pilot competency and flight safety, providing a replicable framework for aviation training programs.
***
The use of drones has surged across a wide range of industries, from aerial photography and surveillance to precision agriculture, disaster response, and multi-drone operations. Once limited to experimental applications, unmanned aircraft systems (UAS) have now become integral to commercial, research, and security operations.
Drones exemplify disruptive technology, fundamentally altering traditional methods and expanding operational possibilities. As Clayton Christensen argues, disruptive innovation emerges when entrepreneurs leverage new technologies to gain a competitive edge. True disruption is not just about the product—it is about reshaping the business model.
The focus has shifted from drone hardware itself to the broader systems and services it enables. What once served as a novel tool for small-scale cargo transport has matured into an established logistics model. In 2021, the Brazilian news outlet G1 captured this shift with the headline: McDonald's starts delivering by drone in the Northeast in partnership with iFood.” Operated by Speedbird Aero, drones transported orders from the mall RioMar Shopping in Aracaju across the Sergipe River, cutting delivery times from an average of 40 minutes to just over five. This transformation underscores how drones are redefining efficiency and reshaping market dynamics.
The evolution of drone-based business models extends beyond traditional investment in logistics, systems, and technology. It demands a shift in operational strategy, workforce integration, and most critically—training. Developing a skilled cadre of remote pilots is not a secondary concern; it is a foundational requirement for the industry’s sustained growth.
Emerging technologies—robotics, artificial intelligence, nanotechnology, and quantum computing—are reshaping production and service sectors, forcing a corresponding transformation in education and training frameworks. Drones, as a convergence point for these advancements, now serve as a critical interface between data, automation, and human decision making.
Seen through the broader lens of aviation history, this trajectory was foreseeable. Italian General Giulio Douhet, a pioneering theorist of airpower, anticipated the inevitability of such advances, recognizing the transformative nature of flight long before drones entered the equation: “This machine [aircraft] which man has forged out of his genius and daring after millenniums of trial and failure, is the swiftest and most marvelous invention in the history of transportation. Its eventual development cannot be predicted now, but all the signs point to a long life for it.”
Brazil’s drone sector has expanded at a staggering pace. Data from the Department of Airspace Control’s (DECEA) System for Requesting Access to Brazilian Airspace by Unmanned Aircraft (SARPAS) recorded approximately 410,000 drone flight requests in 2024—a dramatic leap from the mere 95 flights logged in 2016 when the system was first introduced. Initially, SARPAS functioned as a regulatory tool still in its infancy, aimed at familiarizing remote pilots with airspace restrictions and operational requirements. Today, its exponential growth signals the full-scale integration of drones into Brazil’s airspace (see fig.1).
Figure 1. Increased flight requests in Brazil. (Source: DECEA.)
Such rapid proliferation demands stronger regulatory oversight, enhanced supervision, and a structured approach to integrating drones into commercial and professional sectors. Chief among these imperatives is pilot training. Without a rigorous foundation in aviation principles, remote pilots pose risks not only to themselves but to the integrity of national airspace.
Establishing a doctrinal framework is essential to maintaining flight safety and ensuring the responsible evolution of drone operations. Remote pilots must master aeronautical culture, air traffic regulations, micrometeorology, aircraft performance parameters, and standardized phraseology. In short, they must be trained not merely as operators but as aviators, equipped with the knowledge and discipline to navigate a rapidly evolving domain.
Several factors shape the development of technical proficiency, and early training is the ideal stage to instill doctrinal principles in future professionals. Among the many foundational elements, one stands above the rest: equipment selection.
In manned aviation, the question “What kind of equipment is it?” refers directly to the aircraft type, providing immediate insight into its capabilities. The same logic applies to drones. Choosing the right platform is not incidental—it is fundamental.
Drones vary widely in size, weight, speed, endurance, mission profile, and onboard systems. In Brazil, the National Civil Aviation Agency (ANAC) categorizes drone operations into three types:
-
Beyond Visual Line of Sight (BVLOS): No direct visual contact with the aircraft.
-
Extended Visual Line of Sight (EVLOS): Operations assisted by a ground observer.
-
Visual Line of Sight (VLOS): Direct eye contact with the aircraft—the most common mode in Brazil.
Figure 2. Types of Operation (Source: https://fixar.pro.)
ANAC further classifies drones by maximum takeoff weight (MTOW):
Drones under 250 g, categorized as model aircraft, enjoy fewer restrictions and do not require ANAC registration.
For most recreational and professional applications, aircraft under 25 kg dominate the market. However, regulatory compliance extends beyond ANAC. Remote pilots must register with SISANT and adhere to flight authorization protocols via SARPAS. Additional oversight comes from agencies such as:
-
National Telecommunications Agency (ANATEL): Certifies radio frequency compliance.
-
Ministry of Defense (MD): Regulates military and security-related operations.
-
Ministry of Agriculture and Livestock (MAPA): Governs agricultural drone use.
While these regulations exist, enforcement remains a challenge. The sheer volume of drone activity—as seen in the exponential rise in flight requests—places strain on monitoring mechanisms, underscoring the need for stronger compliance frameworks and pilot accountability.
In July 2022, a drone struck a woman during a concert in Brasília, requiring her to undergo surgery. The incident, which took place in the Setor de Clubes Esportivos Sul, unfolded despite the victim’s attempts to avoid the aircraft. This case underscores a persistent gap in public awareness and regulatory enforcement—laws exist, but ignorance of drone safety remains widespread.
This study seeks to engage with the state of the art in drone operations, examining existing literature to identify gaps, challenges, and opportunities for improvement. A rigorous review of current research not only clarifies the theoretical foundations of the field but also illuminates emerging trends and best practices.
By analyzing the most advanced knowledge and real-world applications, integrative research enables us to extract lessons from past innovations, highlight systemic weaknesses, and propose targeted solutions. As drone technology continues to evolve, scholarship must evolve with it—bridging the gaps among theory, policy, and operational reality.
Description of the Problem
The Brasília drone incident, which resulted in serious bodily injury, serves as a stark reminder of the risks posed by untrained operators. It underscores the urgent need for structured decision-making research and reinforces the necessity of mandatory professional training to strengthen the unmanned aviation sector.
The risks are not theoretical. In May 2023, the Energy Company of Paraná (Copel) reported five separate drone-related accidents that damaged power grids across rural areas—including Palotina, Londrina, Nova Cantu, Formosa do Oeste, and Ubiratã—resulting in power outages. As Copel’s drone fleet coordinator Vitor Marzarotto observed, safe flight operations demand training and experience. While regulatory adjustments have simplified flight rules for agricultural drones, this increased accessibility has also heightened the need for competent pilots.
Against this backdrop, equipment selection becomes the first critical decision in a remote pilot’s training. This study focuses on identifying the most suitable drone for initial qualification training (IQT), ensuring that student pilots develop essential skills on an aircraft that balances cost, durability, ease of operation, performance, and real-world applicability.
Decision-making in this context is rarely straightforward. Multiple criteria must be weighed—financial constraints, operational efficiency, learning curve, and mission adaptability. Despite the availability of sophisticated analytical tools, subjectivity remains an inescapable factor. The final decision always rests with the individual pilot or instructor, whose judgment will determine whether an aircraft meets the practical and strategic demands of foundational training. The challenge is not merely selecting a drone—it is ensuring that the choice aligns with both training objectives and broader safety imperatives.
Theoretical Foundation
Operational Research (OR), the foundation of this study, is best understood as “a science composed of numerous techniques and models intrinsically related to the optimization of production systems.”
Decision problems typically involve multiple criteria, often divergent, that influence analysis to varying degrees. According to Carlos Francisco Simões Gomes and colleagues, these criteria are essential for classifying the available alternatives to solve a given problem. Fernanda Gomes de Andrade contends that the Step-Wise Weight Assessment Ratio Analysis (SWARA) method determines the prioritization criteria’s weights based on decision makers’ input. Likewise, Mário Henrique Sombra Beuttenmüller Vilela asserts that Multi-Objective Optimization on the Basis of Ratio Analysis (MOORA) helps decision-makers eliminate poor choices.
Developed by Lucas Ramon dos Santos Hermogenes and colleagues, (2022), this methodology strengthens decision-making by reducing reliance on sensitivity analysis. It ensures that the ordering of problem alternatives remains consistent with the MOORA method’s initial classification. Renan Felinto de Farias Aires and Luciano Ferreira note that multicriteria decision making is among the most widely used approaches for enhancing decision quality across science, government, business, and engineering. Multiple-Criteria Decision Analysis (MCDA) encompasses formal approaches that weigh various criteria, helping stakeholders evaluate significant decisions. MCDA also provides decision makers with a structured framework for assessing alternatives and forming well-founded judgments.
An air force’s aircraft acquisition process, for instance, must account for diverse factors such as service ceiling, operating range, cruise speed, and climb rate. Given these complexities, Multicriteria Decision-Making theory offers a rigorous approach to navigating such high-stakes choices.
Methodology
In the world of drones, quantitative decision-making tools enable the evaluation of a broad range of alternatives based on multiple criteria. These methods bring rationality and objectivity to complex decision making, preventing hasty or uncritical choices.
This study relies on the quantitative and hybrid SWARA-MOORA-3NAG model, introduced by Yousef Bahrami, Hossein Hassani, and Abbas Maghsoudi. SWARA assigns weights to decision criteria, while MOORA identifies the optimal alternative in a straightforward and efficient manner. This approach delivers rapid solutions, making it suitable for both simple and complex problems.
Within the SWARA framework, the decision-maker selects the most relevant criteria, ranks them by priority, and determines how much worse each criterion is compared to its predecessor. The first criterion is assigned a baseline value of zero. From this ranking, informed values (SJ) emerge, with subsequent criteria assigned diminishing values.
The coefficient (KJ) is then calculated using the first equation, prioritizing criteria from most to least important. The SJ of the top-ranked criterion remains zero, while others receive lower percentage values. For example, if cost holds the highest importance and power is considered 25% less significant, the cost coefficient (KJ) is 1, while the power coefficient is 1.25. The KJ coefficient thus quantifies each criterion’s relative importance.
The second equation recalculates the weights using KJ. The numerator of the first criterion remains 1, as its KJ is also 1. For other criteria, the formula divides the weight (WJ) of the preceding criterion by the KJ of the criterion being analyzed.
Following the application of the first three equations, the MOORA method establishes the decision matrix. This matrix, structured by alternatives (j) and objectives (i), undergoes normalization using the fourth equation.
The result is a dimensionless value, ranging from 0 to 1, where values closer to 0 indicate better alternatives. The fifth equation optimizes the model by summing values for maximization and subtracting those requiring minimization.
Next, the sixth equation applies the Tchebycheff Min-Max metric to measure distances between alternatives and the reference point, determining whether maximization or minimization was used.
The seventh equation then subtracts alternative distances to establish absolute ordering from the first normalization.
According to Hermogenes, the first part of the equation (OAN1) represents the cardinal value of each alternative in absolute order. The MOORA method produces Aijm values, while the Tchebycheff Min-Max metric generates Aijmin-max values.
This process repeats until the eleventh iteration. After establishing the absolute ordering from normalization, the twelfth equation sums values for each alternative.
The final absolute global ordering (OAG) of each alternative follows from this sum, with j = 1 marking the first alternative and j = m denoting the final one. OANn represents the absolute ordering across previous steps.
Data Collection and Analysis
A bibliographic survey—drawing from books, academic articles, and both print and digital media—provided the initial foundation for this study. The research focused on 10 drone models manufactured by Dà-Jiāng Innovations Science and Technology (DJI): Tello, Spark, Mini 1, Mini 2, Mini 3, Phantom 3, Phantom 4 Pro, Mavic 2 Zoom, Mavic 3 Enterprise, Inspire 2, and Matrice 210. These models reflect the firsthand experience of the article’s author, who piloted them as part of the Specialization Course on Unmanned Aerial Vehicles (CEVANT) at the Military Fire Department of Rio de Janeiro (CBMERJ) in 2019.
The study evaluated these drones based on five key criteria: operational cost (economy), autonomy (flight time), maneuverability (performance), wind resistance (stability), and portability (logistics). Other potential factors—such as speed, weight, range, and sensors—were excluded from consideration.
Figure 3 depicts a CBMERJ training workshop designed to reinforce situational awareness, operational safety, and teamwork. The exercise required students to coordinate in writing, through voice commands, hand signals, and radio communication—highlighting the importance of precision and responsibility in remote piloting.
Figure 3. Remote Piloting Instruction (CEVANT) (Source: CBMERJ.)
The study’s data, including autonomy and wind resistance, were derived from official DJI manuals. Operational cost figures reflected market averages for 2023.
To complement this methodological approach, the study incorporated insights from Colonel Aviator R1 Jorge Vargas, founder of the Triplo 4 Professional View school, which trains remote pilots across Brazil. The research team also conducted on-site observations of practical training exercises, gathering data on instructor-led activities and the operational factors influencing UAV flights during IQT.
Figure 4. Briefing on drones and equipment. (Source: https://t4drones.com.)
Questions for Colonel Vargas were posed only after applying the SWARA-MOORA-3NAG operational research method, ensuring that comparisons remained objective and free from bias, subjectivity, or undue influence.
Firsthand observation of the school’s remote piloting exercises provided additional insights. The training encompassed all aspects of drone operation, including assembly, disassembly, maintenance, and the preparation of operational environments for simulated exercises.
Field observation further reinforced the study’s findings. It became evident that smaller drones, capable of performing the same tasks as larger models, offer a distinct advantage in IQT. Their ease of handling enhances safety, reducing operational risks for novice pilots.
Proposal for a Solution
The SWARA-MOORA-3NAG computational tool initially processed the problem by defining and inputting the relevant alternatives.
Figure 5. SWARA-MOORA-3NAG Homepage (Source: https://mcda-software-swara-mod1-sm-3nag-k86acd.streamlit.app).
Upon accessing the platform, the first step was to assign a title to the study and record the evaluated alternatives. As shown in figure 4, the selected drone models were entered into the system.
The criteria were then ranked by preference, with cost and profit factors clearly defined. Monotonic criteria were labeled as "C" (cost) and "P" (profit) (fig. 5). Each criterion was weighted on a scale from 0 to 1 to determine its relative importance. For instance, in the case of operational cost, higher values indicated greater expense.
For criteria classified as "L" (monotonic increasing), higher values signified better performance. In the context of drones, key variables such as autonomy, maneuverability, wind resistance, and portability were prioritized. Greater autonomy enables longer practice sessions, improved maneuverability enhances pilot control, higher wind resistance reduces environmental interference, and increased portability facilitates transportation.
Based on practical experience, autonomy was deemed 0.7 times as important as operational cost, while maneuverability was assigned a weight of 0.5 relative to autonomy. Wind resistance, which plays a lesser role at low altitudes, was weighted at 0.3, the same value given to the relationship between portability and wind resistance.
Figure 6. SWARA-MOORA-3NAG Homepage. (Source: https://mcda-software-swara-mod1-sm-3nag-k86acd.streamlit.app.)
The author argues that drone selection depends significantly on the pilot's level of experience. In an indoor setting, where wind and other external factors are irrelevant, wind resistance becomes inconsequential, and autonomy emerges as the most critical factor for sustained practice. In such an environment, the optimal drone is the one that remains airborne the longest. Conversely, in open-air conditions where wind poses challenges for novice pilots, wind resistance takes precedence. If autonomy is low but training time remains sufficient, this limitation does not create operational drawbacks.
Figure 7. Criteria and Weights (Source: generated by the author in the computational tool.)
To compute the global ranking, the market prices of the aircraft (in Brazilian Reais) were factored into the analysis:
Table 1 presents additional data on autonomy and wind resistance, derived from technical specifications. Maneuverability and portability were evaluated using the Likert Scale, ranging from 1 to 5.
According to Rafaela Frankenthal, the Likert Scale allows researchers to derive qualitative insights from structured quantitative data. It can measure attributes such as agreement, frequency, probability, importance, and satisfaction. ith an odd-numbered range (1 to 5), the central value typically indicates neutrality or moderation. In this study, the scale was used to evaluate drone performance, with scores assigned as follows:
1 – Poor
2 – Fair
3 – Good
4 – Very Good
5 – Excellent
Table 1. Decision Matrix (Source: generated by the author in the computational tool.)
Table 1. Decision Matrix (Source: generated by the author in the computational tool.)
| DECISION MATRIX |
| |
Operational Cost |
Autonomy |
Maneuverability |
Wind Resistance |
Portability |
| TELLO |
750.0 |
13.0 |
2.0 |
9.0 |
5.0 |
| SPARK |
2900.0 |
15.0 |
3.0 |
28.0 |
5.0 |
| MINI 1 and 2 |
4300.0 |
30.0 |
5.0 |
37.0 |
5.0 |
| MINI 3 |
6100.0 |
41.0 |
5.0 |
11.0 |
5.0 |
| PHANTOM 3 |
2300.0 |
25.0 |
4.0 |
36.0 |
4.0 |
| PHANTOM 4 PRO |
10500.0 |
30.0 |
4.0 |
36.0 |
4.0 |
| MAVIC 2 ZOOM |
12500.0 |
31.0 |
5.0 |
35.0 |
5.0 |
| MAVIC 3 ENTERPRISE |
26000.0 |
46.0 |
5.0 |
43.0 |
5.0 |
| INSPIRE 2 |
38000.0 |
27.0 |
4.0 |
36.0 |
3.0 |
| MATRICE 210 |
52000.0 |
35.0 |
4.0 |
43.0 |
3.0 |
Table 2 illustrates the normalized matrix computed using the MOORA method, which incorporates the decision matrix by integrating criteria weights, cost factors, profit relations, and the evaluation scale. The right side of Table 2 displays the global absolute distances between alternatives, derived by multiplying the normalized MOORA matrix with the SWARA-generated weights.
Table 2. Normalization in MOORA (Source: generated by the author in the computational tool.)
|
MOORA Ranking—Normalization 3
|
Global Absolute Distance
|
|
MOORA RANKING—Normalization 3
|
|
Global Ranking
|
|
MINI 1 and 2
|
0.471226
|
MINI 1 and 2
|
0.576675
|
|
MINI 3
|
0.440858
|
MINI 3
|
0.520598
|
|
PHANTOM 3
|
0.408646
|
PHANTOM 3
|
0.447235
|
|
MAVIC 2 ZOOM
|
0.408009
|
MAVIC 2 ZOOM
|
0.394702
|
|
MAVIC 3 ENTERPRISE
|
0.403364
|
PHANTOM 4 PRO
|
0.373102
|
|
PHANTOM 4 PRO
|
0.371355
|
SPARK
|
0.218049
|
|
SPARK
|
0.318132
|
MAVIC 3 ENTERPRISE
|
0.128232
|
|
TELLO
|
0.240191
|
TELLO
|
0.090200
|
|
INSPIRE 2
|
0.127284
|
INSPIRE 2
|
-0.538084
|
|
MATRICE 210
|
0.080419
|
MATRICE 210
|
-0882579
|
Figure 8 visualizes the ordering generated by the SWARA-MOORA-3NAG tool. The computational results identified the Mini 1, Mini 2, Mini 3, and Phantom 3 as the three most suitable options for IQT.
Figure 8. Global Absolute Ordering (Source: generated by the author in the computational tool.)
The Phantom 3 lacks onboard sensors, which may make it less suitable for basic training. Figure 9 compares the DJI Mini 1 and Mini 2, both weighing approximately 249 grams and sharing similar characteristics. These lightweight, highly maneuverable drones are easily transportable, making them excellent choices for IQT.
Figure 9. Mini and Mini 2 (Source: https://drdrone.ca.)
Recently, DJI has promoted the Mini 2 SE, a new variant of the Mini 2, through social media campaigns emphasizing its affordability and suitability for beginner pilots. While it retains the Mini 2’s cost-efficiency, design, flight characteristics, and weight (with only minor reductions), the key distinction lies in its image processor: the Mini 2 supports 4K video, whereas the Mini 2 SE records at 2.7K. Since this difference does not impact flight training, the Mini 2 SE aligns well with this study’s findings.
Figure 10. Mini 2 SE and Mini 2. (Source: https://www.dji.com/br.)
Although the Mini 3 Pro and Mavic 2 Zoom differ significantly in size, maneuverability, and weight, the classification model grouped them as similarly viable choices for novice pilots.
Figure 11. Mini 3 Pro and Mavic 2 Zoom. (Source: https://shre.ink/Mini3vsMavic2Zoom.)
In an interview, Colonel Vargas was asked to identify the ideal unmanned aircraft for IQT. He recommended the Mini 2, citing its affordability, lightweight frame, portability, and ability to instill confidence in student pilots due to its forgiving design. However, he suggested transitioning to the Mini 3 Pro after initial training, as its built-in screen enhances usability.
The global absolute ranking assigned negative values to the Inspire 2 (-0.5381) and Matrice 210 (-0.8826). This does not imply that these models are unsuitable for training, but rather that they may not be ideal for IQT. Their high operating costs and limited portability make them less practical for introductory flight instruction.
Figure 12. Inspire 2 (on the left) and Matrice 210 (on the right). (Source: https://dronecerto.com.br/blog.)
Discussion of Results
The SWARA-MOORA-3NAG method proved to be a highly effective decision-making tool for drone selection. The criteria applied aligned closely with established requirements in the unmanned aerial systems sector. Despite the inherent trade-offs among the weighted factors, the method produced a well-balanced and satisfactory global ranking.
The Mini 1 and Mini 2 emerged as the top-ranked options, followed by the Mini 3 Pro. This outcome aligns with insights provided by the founder of Triplo 4 Professional View, reinforcing the validity of the selection process.
A key takeaway from the analysis is the close proximity of the results among the highest-ranked alternatives. This suggests that these compact drones effectively meet the predefined requirements, making them strong candidates for training and operational use.
Final Considerations
The SWARA-MOORA-3NAG method proved to be a highly effective analytical tool for addressing complex decision-making scenarios with numerous constraints. The findings reinforce the viability of applying a multicriteria methodology—whether in collaborative or independent decision-making processes.
The results obtained provide valuable support for managers and policymakers facing intricate choices involving multiple alternatives and criteria, particularly in procurement decisions such as the selection of unmanned aircraft.
Ultimately, the study confirmed the efficiency and effectiveness of this method. Its straightforward mathematical modeling allows for seamless application to problems incorporating both qualitative and quantitative criteria. As a hybrid approach, SWARA-MOORA-3NAG delivers a structured, data-driven framework, offering decision makers robust and actionable insights.
CAP Eduardo Araújo da Silva, Brazilian Air Force University (UNIFA)
Captain da Silva is a PhD student in aerospace sciences and holds a master's degree in education. He specializes in military law and technological education and has bachelor's degrees in law and public security, along with undergraduate studies in pedagogy and air traffic management. He serves as head of the UAS Coordination and Control Section of the Operations Subdepartment at the Airspace Control Department (DECEA) and is an operational safety inspector and certified airspace control investigator. He has completed the strategic policies in the face of complex threats course at the Brazilian Higher Warfare School (ESG), the unmanned aerial vehicles specialization course at the Military Fire Department of Rio de Janeiro, and the United Nations staff officers and military observers courses at the Brazilian Navy’s Naval Peace Operations Centre (COpPazNav). He has also served as a UAS doctrine instructor at the Ministry of Defense’s Brazilian Peace Operations Joint Training Center (CCOPAB).
Marcos dos Santos, PhD Federal Fluminense University (UFF)
Dr. dos Santos is a researcher, professor, and consultant specializing in high-level decision-making for public and private organizations. He teaches in the Systems and Computing Graduate Program at the Military Engineering Institute (IME) and the Production Engineering Graduate Program at Fluminense Federal University (UFF). He holds a post-doctorate in space science and technology from the Aeronautics Institute of Technology (ITA), a PhD in production engineering from UFF, and a master’s in production engineering from the Alberto Luiz Coimbra Institute for Graduate Studies and Research in Engineering, a prominent research and education center at the Federal University of Rio de Janeiro.
Notes