Evaluation of Affordable Agricultural Drones for Small and Medium Farms

Abstract

Smart technologies are increasingly used in agriculture, with drones becoming one of the key tools in agricultural production. This study aims to evaluate affordable drones for agricultural use in the Posavina region, located in northern Bosnia and Herzegovina. To determine which drones deliver the best results for small and medium-sized farms, ten criteria were used to evaluate eight drones. Through expert evaluation, relevant criteria were first established and then used to assess the drones. The selected drones are designed for crop monitoring and are priced under EUR 2000. Using the fuzzy A-SWARA (Adapted Step-wise Weight Assessment Ratio Analysis) method, it was determined that the most important criteria for drone selection are control precision, flight autonomy, and ease of use, all of which are technical attributes. The fuzzy MARCOS method revealed that the best-performing drones are also the most affordable. The drones D5, D4, and D8 demonstrated the best results. These findings were confirmed through comparative analysis and sensitivity analysis. Their features are not significantly different from those of more expensive models and can, therefore, be effectively used for smart agriculture. This study demonstrates that drones can be a valuable tool for small farms, helping to enhance agricultural practices and productivity.

1. Introduction

Recent changes in agriculture increasingly focus on using modern technology to improve production, especially through the development of smart systems. Drones in agriculture offer many benefits for small and medium-sized farms by helping to reduce costs and increase production efficiency. They can be used for tasks such as crop mapping, plant health monitoring, irrigation planning, and the careful use of pesticides and fertilizers [1]. Drones provide faster and more complete crop assessments than traditional methods while also reducing the need for human labor. The data they gather allow farmers to make better decisions, leading to improved agricultural outcomes.

A key advantage of drones is their ability to collect data in real time [2]. With this up-to-date information, farmers can react quickly and effectively to changes in crop conditions. Drones can also cover large areas in a short time, and by using multispectral cameras, they detect changes in crops before they are visible to the eye. This allows for precise actions, reducing the amount of pesticides and fertilizers needed, which helps lessen the environmental impact. By using drones, it is possible to identify whether certain crops lack water and to apply precise irrigation only to those crops. Protecting the environment should remain a central focus in agricultural production [3].

Drones in agriculture have attracted much attention from researchers who see their potential to boost production. Research by Javaid et al. [4], Panday et al. [5], Rejeb et al. [6], Nhamo et al. [7], and others shows that drones can help cut costs and increase crop yields. Studies by Hafeez et al. [8] and Inoue [9] also suggest that drones make it faster and easier to assess crops, improving the use of resources like water, fertilizers, and pesticides. However, there are challenges when using drones on small farms. If precision farming methods are applied, the cost–benefit ratio may not always work in favor of smaller farms [10], raising questions about whether drones are worth the investment for them.

Although drones are widely used in large-scale farming [11], they are harder to apply on small and medium-sized farms. The main reasons are the high costs of buying and maintaining drones [12], which greatly influence farmers’ decisions, as well as the technical knowledge required to operate them. These factors can be obstacles for smaller farms with limited budgets. However, as technology develops and market competition increases, drones are becoming more affordable and economically accessible [13], making them a more realistic option for smaller farms.

This study aims to identify affordable and practical drones for small and medium-sized farms in the Posavina region, located in northern Bosnia and Herzegovina. This region, the largest plain in the country, has the highest potential for agricultural growth [14]. The goal of this research is to evaluate affordable drones that meet the technical needs and budget limits of farmers in Posavina. Small and medium-sized farms are central to agriculture in this area, and the adoption of drones could greatly improve their productivity and sustainability [15]. However, to make drone use successful, it is necessary to identify models that best match the specific needs of these farms. This research provides an overview of affordable drones on the market, offering guidelines for their evaluation using decision-making methods. It will help answer the question of how to choose the best drone by balancing technical features and costs, especially when budgets are tight.

Based on this, the main goal of this paper is to evaluate drones suitable for small and medium-sized farms using a fuzzy approach with specialized methods. The fuzzy approach is used in this research because complete information is not available, and precise evaluations cannot be used; instead, assessments are expressed through linguistic values [16]. The application of multi-criteria decision-making (MCDM) methods is performed to evaluate drones based on the relevant criteria. The evaluation of both the criteria and drones is carried out subjectively, relying on expert opinions [17] while incorporating uncertainty into the decision-making process [18]. This approach aims to provide farmers with guidelines on choosing the drone that best fits their specific needs. The fuzzy approach relies on the A-SWARA (Adapted Step-wise Weight Assessment Ratio Analysis) and MARCOS (Measurement Alternatives and Ranking according to Compromise Solution) methods. Here, the fuzzy A-SWARA method helps determine the importance of various criteria based on subjective evaluations [19], while the fuzzy MARCOS method ranks drone options by finding a balance across different factors [15].

This research offers several practical and theoretical contributions, particularly in the application of drones in agriculture. The main contribution is to demonstrate that drones can be effectively used on small and medium-sized farms to enhance agricultural production and increase farmers’ work efficiency. Additionally, the contributions of this research are reflected in the following:

• Guidance for farmers: The study provides clear, practical advice for small and medium-sized farmers on which drone models are best suited for agricultural tasks. By weighing technical and economic factors, this guidance highlights the strengths and weaknesses of specific models, making decision making easier.
• Innovative evaluation method: The use of the fuzzy A-SWARA and MARCOS methods offers a new way to evaluate drones in agriculture, focusing on subjective decision making that can be adapted for similar studies in the future.
• Sustainable agriculture: This drone evaluation contributes directly to sustainable farming by helping optimize resource use in agriculture, which can reduce costs, improve efficiency, and promote ecologically and economically sustainable production practices—vital for the long-term health of agriculture.
• Supporting small farm modernization: This research encourages a broader conversation about applying new technologies to small farms and positioning drones as an effective modernization tool. This can increase the competitiveness of small and medium farms.
• Economic development in rural areas: By lowering costs and boosting productivity, this research can support economic growth in rural regions. Drones help farmers use resources more efficiently, increase income, and improve living standards in rural communities.

Following the introduction, this paper is organized into four main sections that build upon one another. First, the materials and research methods will be explained, detailing the steps and tools used in the study. The results section will apply these methods to a real-world example, guiding the selection of the best-suited drone for small and medium-sized farms in the Posavina region. The discussion will examine the findings in detail, comparing them with similar research. Finally, the last section will summarize the key results, discuss limitations, and suggest directions for future research.

2. Materials and Methods

This research was conducted in several phases:

-

Selection of experts, drones, and criteria;

-

Evaluation of criteria and alternatives;

-

Application of fuzzy SWARA and MARCOS methods;

-

Analysis of results using comparative and sensitivity analysis.

In the first phase, drones were assessed by a panel of experts based on key characteristics and selected criteria. Ten experts were chosen for their knowledge of agricultural production in Posavina and extensive practical experience with drones. Most of these experts are professors from agricultural faculties. Drones for observation were selected based on the price range and availability in Bosnia and Herzegovina, with a maximum price limit set at around EUR 2000. Therefore, this study focuses on drones used primarily for crop monitoring.

To ensure objectivity, drones were labeled rather than named directly, so none were rated “best” in an absolute sense. It is important to note that this research relies on the subjective evaluations of experts familiar with the Posavina region, and different regions or farming conditions could lead to different rankings of the drones. The basic characteristics of the drones examined are summarized in

Table 1. Technical characteristics of selected drones.

Designation	Camera	Range	Flight Autonomy	Weight	Obstacle Sensors	Max Speed	Price Range	Payload
D1	20 MP	7 km	30 min	1375 g	Yes	72 km/h	EUR 1500–1800	500 g
D2	21 MP	4 km	25 min	320 g	No	55 km/h	EUR 700–900	200 g
D3	48 MP	10 km	34 min	570 g	Yes	68 km/h	EUR 900–1000	200 g
D4	8 MP	8 km	43 min	700 g	No	60 km/h	EUR 400–600	100 g
D5	12 MP	8 km	35 min	790 g	No	65 km/h	EUR 500–600	200 g
D6	48 MP	12 km	34 min	249 g	Yes	57 km/h	EUR 1000–1100	100 g
D7	50 MP	10 km	28 min	249 g	Yes	54 km/h	EUR 800–900	300 g
D8	4 K	1 km	26 min	495 g	No	45 km/h	EUR 300–500	100 g

. However, these characteristics do not encompass all of the criteria used in this study. Some criteria can only be evaluated after these drones are tested in practice. For example, image quality is not determined solely by the number of megapixels but is also influenced by factors such as sensors, hardware and software used by the cameras, and various other parameters. As a result, certain criteria can only be assessed when these drones are applied in real-world conditions.

The features of these drones indicate that they are designed for small and medium-sized farms, as they can cover small areas on a single battery charge. They are intended for crop monitoring on these farms, while advanced functions such as spraying or seeding are not available. With a single battery charge, they can typically cover areas of a few hectares, although this coverage can be increased by replacing the batteries. Additionally, these drones are lightweight and cannot carry extra loads. In most cases, it is possible to attach certain sensors to them, provided they do not exceed the drone’s weight limit. Due to their specific features and affordability, these drones are a good choice for monitoring and recording crop conditions on farms. This allows farmers to gather useful information and take appropriate measures to improve agricultural production.

The observed drones are evaluated based on ten criteria, focusing on technical and economic characteristics. These criteria are as follows:

• Criterion (C1) camera quality [20,21]: Assesses camera resolution, stabilization, and image quality, rated from very low to excellent quality.
• Criterion (C2) flight autonomy [21,22,23]: Evaluates flight duration on a single charge, rated from very short to long-term autonomy.
• Criterion (C3) steering precision [22,24]: Measures accuracy of controls and GPS, rated from unreliable to highly precise steering.
• Criterion (C4) stability in air [22,25]: Evaluates drone stability in poor weather, rated from very poor to excellent stability.
• Criterion (C5) obstacle avoidance sensors [24,26]: Assesses the effectiveness of sensors, rated from no sensors to superior crash protection.
• Criterion (C6) control range [21,23,24]: Measures maximum control distance, rated from very short to extremely large range.
• Criterion (C7) ease of use [25,27]: Evaluates control simplicity, rated from very complex to very easy to use.
• Criterion (C8) portability and weight [22,24]: Assesses weight and ease of transport, rated from heavy and hard to carry to light and portable.
• Criterion (C9) battery charging time [21,23]: Measures battery charging speed, rated from slow charging to very fast charging.
• Criterion (C10) value for money [20,27]: Rates the price–performance ratio from very poor to excellent value.

Each criterion and its importance were rated on a nine-level scale, ranging from least to most important. Due to differences across criteria, the scale was adjusted as needed for each. To make these ratings consistent, linguistic evaluations were converted to numerical values from 1 to 9. Likewise, the alternatives were scored using this same scale. Because these ratings are given in linguistic terms, they were converted into fuzzy numbers to apply fuzzy evaluation methods. For example, the lowest rating was converted into a fuzzy number (1, 1, 2) and the highest into (8, 9, 9). Other values were converted in a similar manner.

To practically apply these ratings, fuzzy methods were used to determine the importance of each criterion and to rank the alternatives. Since many methods are available, this research focused on two that have proven effective in similar studies: the fuzzy SWARA [28] and fuzzy MARCOS [29] methods. The fuzzy SWARA method established the importance of each criterion based on expert evaluations, while the fuzzy MARCOS method ranked the selected drones according to these criteria. Through this approach, we aimed to recommend which drones are best suited for agricultural production in the Posavina region.

The fuzzy SWARA method, initially developed by Keršulienė et al. [30], provides a way to determine the importance of various criteria through a step-by-step approach. However, this study uses a simplified variant known as A-SWARA, which facilitates group evaluation of criteria importance. The A-SWARA method includes the following steps:

Step 1. Experts assess the importance of criteria using linguistic values.

Step 2. Linguistic values are converted into fuzzy numbers.

Step 3. An aggregate rating for each criterion is formed.

Step 4. Criteria are ranked based on their overall scores.

Step 5. Normalized values are generated by dividing the aggregate scores by the highest fuzzy number of the top-ranked criterion.

Step 6. Final weights for each criterion are determined by dividing individual values by the sum of all normalized values.

Once the importance of each criterion is set, the ranking of alternatives is performed using the fuzzy MARCOS method developed by Stević et al. [31]. This approach assesses the relationship between each alternative and reference value representing ideal and anti-ideal points, following these steps [32]:

Step 1. Experts evaluate the alternatives based on the chosen criteria using linguistic values.

Step 2. Linguistic values are transformed into fuzzy numbers, creating the initial fuzzy decision matrix.

Step 3. The initial decision matrix is expanded.

Step 4. The matrix is normalized using the following:

$$\tilde{n}=\left(n_{ij}^l,n_{ij}^m,n_{ij}^u\right)=\left({\displaystyle \frac{x_{id}^l}{x_{ij}^u}},{\displaystyle \frac{x_{id}^l}{x_{ij}^m}},{\displaystyle \frac{x_{id}^l}{x_{ij}^l}}\right)ifj\in C$$

(1)

$$\tilde{n}=\left(n_{ij}^l,n_{ij}^m,n_{ij}^u\right)=\left({\displaystyle \frac{x_{ij}^l}{x_{id}^u}},{\displaystyle \frac{x_{ij}^m}{x_{id}^u}},{\displaystyle \frac{x_{ij}^u}{x_{id}^u}}\right)ifj\in B$$

(2)

where l, m, and u are the lower, middle, and upper bounds of the fuzzy numbers, respectively.

Step 5. Weighted values of the normalized matrix are calculated as follows:

$${\tilde{v}}_{ij}=\left(v_{ij}^l,v_{ij}^m,v_{ij}^u\right)={\tilde{n}}_j\times {\tilde{w}}_j$$

(3)

Step 6. Calculation of Si matrix is performed as follows:

$$S_i={\sum }_{i=1}^n{v}_{ij}$$

(4)

Step 7. Calculation of the utility degree K_i is carried out using the following:

$${\tilde{K}}_i^{-}=\left({\displaystyle \frac{{\tilde{S}}_i}{{\tilde{S}}_{ai}}}\right)=\left({\displaystyle \frac{s_i^l}{s_{ai}^u}},{\displaystyle \frac{s_i^m}{s_{ai}^u}},{\displaystyle \frac{s_i^u}{s_{ai}^u}}\right)$$

(5)

$${\tilde{K}}_i^{+}=\left({\displaystyle \frac{{\tilde{S}}_i}{{\tilde{S}}_{id}}}\right)=\left({\displaystyle \frac{s_i^l}{s_{id}^u}},{\displaystyle \frac{s_i^m}{s_{id}^u}},{\displaystyle \frac{s_i^u}{s_{id}^u}}\right)$$

(6)

Step 8. Calculation of the fuzzy matrix ${\tilde{T}}_i$ is carried out using the following:

$${\tilde{T}}_i={\tilde{t}}_i=\left(t_i^l,t_i^m,t_i^u\right)={\tilde{K}}_i^{-}+{\tilde{K}}_i^{+}=({\tilde{k}}_i^{-l}+{\tilde{k}}_i^{+l},{\tilde{k}}_i^{-m}+{\tilde{k}}_i^{+m},{\tilde{k}}_i^{-u}+{\tilde{k}}_i^{+u}),$$

(7)

and by setting the fuzzy number $\tilde{D}$ as follows:

$$\tilde{D}=\left(d^l,d^m,d^u\right)=\underset{i}{\max }{\tilde{t}}_{ij}$$

(8)

Step 9. Defuzzification of fuzzy numbers is performed with the following:

$${df}_{def}={\displaystyle \frac{l+4m+u}{6}}$$

(9)

Step 10. Determination of the utility function f(K_i) is carried out as follows:

$$f\left({\tilde{K}}_i^{+}\right)={\displaystyle \frac{{\tilde{K}}_i^{-}}{{df}_{def}}}$$

(10)

$$f\left({\tilde{K}}_i^{-}\right)={\displaystyle \frac{{\tilde{K}}_i^{+}}{{df}_{def}}}$$

(11)

Step 11. Calculation of the final utility function is performed as follows:

$$f\left(K_i\right)={\displaystyle \frac{K_i^{+}+K_i^{-}}{1+\frac{1-f\left(K_i^{+}\right)}{f\left(K_i^{+}\right)}+\frac{1-f\left(K_i^{-}\right)}{f\left(K_i^{-}\right)}}}$$

(12)

3. Results

The selection process for a suitable drone for agricultural production in Posavina begins with evaluating the importance of specific criteria. This assessment relies on expert ratings, where each criterion is scored on a scale of 1 to 9, with 1 representing the lowest importance and 9 indicating the highest, or extreme, importance (

Table 2. Criteria importance evaluation.

Experts	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10
Expert 1	8	8	9	8	7	7	9	6	7	8
Expert 2	7	8	9	7	7	6	9	6	7	9
Expert 3	7	9	9	7	7	6	9	6	8	9
Expert 4	6	9	9	7	7	8	9	6	9	9
Expert 5	6	8	9	6	7	5	9	5	8	9
Expert 6	5	9	9	7	6	6	7	6	8	6
Expert 7	6	9	9	6	5	6	7	6	8	9
Expert 8	7	8	8	6	6	6	8	7	8	8
Expert 9	6	9	8	7	6	6	8	6	8	6
Expert 10	6	8	8	6	6	5	8	5	7	8

).

After assigning ratings, linguistic values are converted to fuzzy numbers, and a total score across all criteria is calculated. The next step involves calculating the weights for each criterion. First, criteria are ranked by the total score. Then, each total score is divided by the highest fuzzy value, and the weight for each criterion is determined by dividing individual normalized values by the sum of these normalized values. According to this approach, the results show that experts consider criterion C3, steering precision, as the most important factor in drone ranking (

Table 3. Criteria importance results.

Id	$\mathbf{Sum}\left({\mathit{s}}_{\mathit{j}}\right)$	$\mathbf{Normalization}\left({\mathit{n}}_{\mathit{j}}\right)$	$\mathbf{Weight}\left({\mathit{w}}_{\mathit{j}}\right)$
C3	(77, 87, 90)	(0.86, 1.00, 1.00)	(0.10, 0.11, 0.11)
C2	(75, 85, 90)	(0.83, 0.94, 1.00)	(0.09, 0.11, 0.11)
C7	(73, 83, 88)	(0.81, 0.92, 0.98)	(0.09, 0.10, 0.11)
C10	(71, 81, 86)	(0.79, 0.90, 0.96)	(0.09, 0.10, 0.11)
C9	(68, 78, 87)	(0.76, 0.87, 0.97)	(0.08, 0.10, 0.11)
C4	(57, 67, 77)	(0.63, 0.74, 0.86)	(0.07, 0.08, 0.10)
C5	(54, 64, 74)	(0.60, 0.71, 0.82)	(0.07, 0.08, 0.09)
C1	(54, 64, 74)	(0.60, 0.71, 0.82)	(0.07, 0.08, 0.09)
C6	(51, 61, 71)	(0.57, 0.68, 0.79)	(0.06, 0.08, 0.09)
C8	(49, 59, 69)	(0.54, 0.66, 0.77)	(0.06, 0.07, 0.09)

), followed by C2, flight autonomy, and C7, ease of use. The least important, according to experts, is C8, portability and weight.

The next phase evaluates the selected drones based on the defined criteria (

Table 4. Drone evaluation by selected criteria.

Expert 1	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10
Drone 1	5	6	5	4	6	6	6	7	8	6
Drone 2	6	5	5	4	6	5	4	5	7	5
Drone 3	5	5	5	5	7	8	5	8	5	6
Drone 4	5	6	7	3	5	8	6	8	6	5
Drone 5	6	6	7	8	5	5	4	5	6	8
Drone 6	5	6	7	5	6	8	6	8	5	6
Drone 7	6	6	6	6	6	5	8	7	7	5
Drone 8	5	8	6	4	8	7	5	6	4	5
⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮
Expert 10	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10
Drone 1	4	5	5	3	5	5	6	6	7	6
Drone 2	5	5	4	3	5	4	5	4	6	6
Drone 3	4	4	4	4	4	7	4	7	4	5
Drone 4	5	5	6	2	5	7	5	7	5	4
Drone 5	5	6	6	7	4	6	4	4	5	7
Drone 6	4	4	6	4	4	7	5	7	4	5
Drone 7	6	5	6	5	5	4	7	6	6	6
Drone 8	5	7	5	3	7	6	4	5	3	5

). Experts first rate each drone, and these linguistic ratings are then converted to fuzzy numbers. This allows for the formation of a summary decision matrix, which is the basis for ranking drones. In forming this matrix, each expert is given equal importance, averaging all matrices created from expert evaluations. With the aggregate decision matrix complete, the fuzzy MARCOS method is applied, including steps for expanding the aggregate matrix, normalizing values, and weighting these normalized values according to criteria weights. This paper does not detail the calculations of the fuzzy MARCOS method, as it has already been validated in over 1000 studies.

Once values for each alternative are calculated, the utility degree and utility function are determined, followed by the defuzzification process to obtain crisp numbers. Finally, a utility function is calculated, which indicates that, according to expert evaluations, Drone 5 performs best, followed by Drones 4 and 8, with Drone 2 ranked the lowest (

Table 5. Fuzzy MARCOS method results.

	$\mathit{d}{\tilde{\mathit{K}}}_{\mathit{i}}^{-}$	$\mathit{d}{\tilde{\mathit{K}}}_{\mathit{i}}^{+}$	$\mathit{d}\mathit{f}\left({\tilde{\mathit{K}}}_{\mathit{i}}^{-}\right)$	$\mathit{d}\mathit{f}\left({\tilde{\mathit{K}}}_{\mathit{i}}^{+}\right)$	${\mathit{K}}_{\mathit{i}}$	Rank
Drone 1	0.819	1.245	0.562	0.370	0.592	6
Drone 2	0.777	1.181	0.533	0.351	0.525	8
Drone 3	0.817	1.242	0.560	0.369	0.589	7
Drone 4	0.871	1.324	0.597	0.393	0.682	2
Drone 5	0.879	1.337	0.603	0.397	0.697	1
Drone 6	0.841	1.278	0.577	0.379	0.629	5
Drone 7	0.854	1.299	0.586	0.386	0.653	4
Drone 8	0.868	1.319	0.595	0.392	0.676	3

).

This ranking reflects a balance among the criteria considered. To confirm this ranking, a comparative analysis is conducted using the same decision matrix and criteria weights but employing different fuzzy methods [33,34]. Five additional methods are used in this analysis: fuzzy RAWEC (Ranking of Alternatives with Weights of Criteria), fuzzy WASPAS (Weighted Aggregated Sum Product Assessment), fuzzy SAW (Simple Additive Weighting), fuzzy MABAC (Multi-Attributive Border Approximation Area Comparison), and fuzzy ARAS (Additive Ratio Assessment). Each method has unique steps: for instance, fuzzy RAWEC applies two types of normalization, fuzzy WASPAS balances two methods, fuzzy SAW ranks based on individual values, fuzzy MABAC ranks relative to the average, and fuzzy ARAS uses a new utility function relative to the maximum value. As these methods also differ in normalization approaches, results may vary.

This analysis reveals some differences in ranking, especially for Drone 4 (Figure 1). According to one method, this drone ranks first, while another method places it fourth. However, most methods agree that Drone 5 is the best performer. Additionally, the ranking obtained by the fuzzy MARCOS method aligns closely with the other methods, thereby supporting the results produced by this approach.

Following the comparison with other fuzzy methods, a sensitivity analysis is conducted. This analysis helps determine how shifts in criteria weights affect the final ranking of the drone alternatives. Sensitivity analysis can be performed in various ways, and for this study, specific criteria weights are reduced by 30%, 60%, and 90%, with the weights of the remaining criteria adjusted to balance these reductions. This approach allows us to see how a reduction in the weight of one criterion influences the overall ranking. Since there are 10 criteria with three levels of reduction each, this results in 30 distinct scenarios for analysis.

The sensitivity analysis results indicate that even slight changes in criteria weightings can impact the ranking order (Figure 2). For instance, Drone 5, which originally ranked highest, maintained the top position in 13 scenarios but ranked lower in the others. In Scenario 12, Drone 5 dropped to fifth place when criterion C4, stability in air, was reduced by 90%. This outcome suggests that Drone 5 excels in stability; however, when the importance of this feature is minimized, its ranking declines. A similar examination of Drone 8 reveals that it performs well in flight autonomy (C2) and obstacle avoidance (C5) but could improve in steering precision (C3) and battery charging time (C9) to become the top-ranked option. Based on these findings, Drones 4, 5, and 8 emerge as the top choices, with experts highlighting their strengths for small to medium farms.

4. Discussion

With the rise of technology, agriculture is also undergoing substantial transformation [35]. Agriculture in many regions, including Posavina [36], is moving toward adopting smarter technologies. These advancements were initially limited to large farms but are now reaching medium and smaller farms due to the availability of more affordable technologies [37]. However, the adoption rate of smart technologies is between 15 and 20 percent, as indicated by research conducted by Gabriel and Gandorfer [38]. Drones, in particular, are gaining traction in the agricultural sector [6]. They serve multiple functions and are becoming integral to farming operations. Research by McCarthy et al. [39] has shown that drones should be used on small and medium-sized farms to increase productivity, reduce costs, and improve food security. This research focused on selecting suitable drones for medium and small farms in Posavina for several reasons:

• Posavina, the largest lowland area in Bosnia and Herzegovina, holds significant potential for expanding agricultural production.
• Bosnia and Herzegovina, as a developing country, relies heavily on traditional farming practices, with limited adoption of smart technologies.
• Drone prices have decreased, making this technology more accessible to farmers.
• Drone use in Bosnia and Herzegovina’s agriculture sector remains limited, highlighting the need for studies like this to guide adoption in the region.

The motivations behind this research underscore its potential contribution to advancing agricultural practices in Bosnia and Herzegovina. Given the diverse range of drones on the market, each designed with specific applications in mind, this study focused on eight drones priced under EUR 2000, each intended for crop monitoring. However, crop monitoring is not achievable without the use of specific software, and the future of this field lies in the integration of artificial intelligence, as noted by Guebsi et al. [40]. Therefore, future research should pay particular attention to software solutions for crop monitoring analysis. It is also essential to emphasize that artificial intelligence is becoming increasingly common in practice [41], including in agriculture.

To determine which of these drones would best meet the needs of agricultural producers in Posavina, fuzzy methods were used, as expert evaluations were expressed through linguistic values. Research by Więckowski et al. [42] demonstrated that linguistic values and expert evaluations are most commonly applied when qualitative criteria are used. Additionally, MCDM methods are often used to obtain such results [43]. This research leveraged the SWARA method for collective decision making, based on its initial principles [30], with certain steps removed and others added for a more streamlined approach. The modified A-SWARA method used collective expert evaluations to determine the weight of the criteria. According to the evaluations and the results obtained with A-SWARA, drones needed to have strong control, high flight autonomy, and ease of use. Consequently, the focus on drone selection centered on technical characteristics that make them easy to operate, even for agricultural producers who may have limited technical knowledge. To implement these features effectively in agriculture, Merz et al. [21] suggested the development of appropriate software interfaces that allow for both online drone management and the advancement of automatic systems. Such systems would enable drones to continuously monitor crop conditions on farms independently. The less demanding the drones are in terms of required knowledge, the simpler they are to use. Results from this method showed that portability and weight were the least important criteria; all drones in this selection were lightweight, with a maximum weight of 1375 g, making them highly portable. Research by Delavarpour et al. [25] has proven that lighter drones are more commonly used in agriculture because they ascend more quickly and operate at relatively lower altitudes. Additionally, such drones are easier for a single person to handle.

For drone selection, the fuzzy MARCOS method was also applied, and it was used to rank the drones. As a relatively new multi-criteria analysis method with proven practical effectiveness, MARCOS was an appropriate choice for this study. The results revealed that avoidance sensors were not a significant factor in drone selection, as the two top-ranked drones did not have this feature. However, in complex agricultural production, where diverse farming activities are present, it is necessary for drones to have these sensors, as noted by Ahmed et al. [24]. This may be due to Posavina’s vast lowland terrain, where drones are likely used for cereal cultivation with few obstacles to navigate. Additionally, the most favorable drones were among the least expensive of those reviewed, an important factor in Bosnia and Herzegovina as a developing country where cost plays a major role in technology adoption. Farmers are likely to prefer affordable drones over costlier models. Another interesting finding was that camera megapixels did not significantly impact selection; even a 2 MP camera provides adequate quality for monitoring crop conditions [44].

Further analysis confirmed that more drones could be effectively used in Posavina’s agricultural production. In particular, the sensitivity analysis showed that three drones performed best, suggesting that major investments in high-end models are not necessary. These findings indicate that drones are no longer a luxury restricted to large farms but are increasingly accessible for all farm sizes. However, alongside these benefits, certain challenges remain. Effective use of drone photography requires specialized software and a capable computer. Additionally, farmers must have the necessary knowledge to maximize the benefits drones offer in agriculture. Moreover, drones enable the implementation of sustainable agricultural production and enhance the competitiveness of small and medium-sized farms [45].

5. Conclusions

This research applied fuzzy methods to evaluate cost-effective drones suitable for agricultural production, specifically tailored to the needs of the Posavina region, the largest lowland area in Bosnia and Herzegovina. By incorporating expert evaluations, this study determined the relative importance of various criteria and ranked drones accordingly. Using the A-SWARA method, it was established that the most critical criteria in drone selection are C3—steering precision, C2—flight autonomy, and C7—ease of use. The fuzzy MARCOS method further identified that drones D5, D4, and D8 exhibit the most favorable characteristics and are the top choices for agricultural use in Posavina. It should be noted that this result was obtained solely based on expert opinions and the use of qualitative criteria. Therefore, future research should incorporate a combination of both qualitative and quantitative criteria to make better-informed decisions. This study aimed to demonstrate how drones can be used in agricultural production in Posavina and how more affordable drones can also be effectively utilized. For this reason, the primary goal of this research was not to select the best drone but to show that drones are essential for improving agricultural production.

Like any study, this research has certain limitations. Specifically, the criteria and selection of drones included may affect the results; alternative criteria or additional drones might lead to different rankings. Future research should aim to identify the most relevant criteria for drone selection and apply those in evaluations. Furthermore, it is necessary to use quantitative criteria in future studies, with a focus on selecting specific types of drones. The choice of drones was influenced by the models available on the Bosnia and Herzegovina market and limited by budget considerations, meaning that other drones could potentially be included in future studies. The primary focus here was to demonstrate the feasibility of drones for medium and small farms. Sensitivity analysis showed that each drone has distinct characteristics that may appeal differently based on individual farm requirements, suggesting that future research should develop adaptable decision-making models based on specific research objectives and farm needs.