Design and Comparison of Two Maize Seeders Coupled with an Agricultural Robot

Abstract

In recent years, the development of robotic vehicles in agriculture has made it possible to reduce human intervention and fatigue in carrying out arduous or repetitive tasks, as well as helping to promote sustainable agriculture to address climate change. However, the great diversity of agricultural tasks and the varied production systems and crops demand a wide range of solutions that can be adapted to robotic vehicles as a power source. These alternatives must be affordable and user-friendly for some users, although more sophisticated solutions must also be developed for others, depending on their specific needs. For this, the present work focuses on the development of two maize seeders with different metering systems coupled to an agricultural robot. The first seeder has a conventional mechanically driven seed metering system with a drive wheel and chain gear, while the second one has an electronically driven metering system based on a DC motor and a digital encoder controlled by a microcontroller. Both seeders were coupled to a remote-controlled robotic vehicle and evaluated on real farmland. Seed distribution in the seed rows was contrasting; the results indicated that the mechanical system performed better in the field than the electronic system. For both seeders, the working capacity was approximately 0.135 ha/h at an average speed of 2.0 km/h. The proposed robot–seeder assembly could help farmers automate and reduce the workload associated with planting, as well as attract young people to the field.

1. Introduction

Ensuring global food security will be a challenge in the coming years because of several factors, including population growth, reduction of available natural resources, urban migration leaving rural communities abandoned, and climate change, which will affect all aspects of food production. The frequency and intensity of droughts, erratic rainfall, and natural disasters represent a shock to the primary sector, which increasingly leads to the need to explore innovative solutions and practices [1,2]. Among the solutions proposed to overcome these obstacles are the implementation of climate-smart and regenerative practices and the development of resistant crop varieties, as well as the application of technology and data analysis for optimized decision-making [3].

In recent years, the automation of agricultural tasks has introduced significant advantages that transform production processes by enhancing resource efficiency and optimizing its usage, reducing labor costs, and improving time management—especially for critical tasks with narrow windows of opportunity. The digitization of these processes with real-time data capturing can, in turn, enhance decision-making processes in the short-, medium- and long-term to ultimately reduce risk, improve productivity, and increase incomes [4,5]. Agricultural robots, also called agrobots, can perform a variety of tasks autonomously or semi-autonomously, including non-standardized activities such as selective weed control, fruit selection, or crop monitoring in challenging production environments [6].

Smallholder farming is the most dominant form of agriculture globally, using approximately 28% of the world’s agricultural land while contributing an estimated 23.5% of the world’s maize production [7,8]. Despite the important role that smallholder farmers play and the fact that the agricultural sector has seen numerous technological advances, these technologies have not always reached small or medium-sized farms [9]. Furthermore, new technologies are usually linked to economies of scale favoring large farms [10], leaving behind smallholder farmers who are unable to access appropriate mechanization [11]. Technologies such as robotics and artificial intelligence (AI), however, present potential opportunities to solve the challenges that hamper the success of smallholder farmers [9].

For maize production, planting represents one of the most important tasks because successful planting directly impacts plant growth and yield at harvest. In smallholder farming systems, manual methods are still widely used for planting, which is not only highly inefficient and time-consuming but also demands significant physical effort, often resulting in health issues due to the high drudgery over the years and creating disinterest in rural youth to engage in farming [12]. In this sense, seed meters play a vital role by ensuring precise seed singulation in the crop and providing an adequate flow that meets desired plant density and spatial arrangements that avoid competition for nutrients [13,14]. Electronically controlled seed meters that automate the seed singulation process have been developed to increase precision and durability while reducing size and maintenance issues compared to conventional mechanical systems [15,16]. However, due to significant cost increases, these kinds of metering devices are rarely found in medium- or small-scale production [9], representing a bottleneck for scaling autonomous mechanization systems in agriculture in the Global South. Designing robot-operated seeders has the potential to reduce labor requirements for planting, improve planting precision, and facilitate timely and uniform seed placement that maximizes crop yields and productivity.

Several solutions to perform crop seeding have been developed. For example, Azmi et al. [17] designed a low-cost seeding robot with a four-wheeled mobile base incorporating a crank-slider mechanism for seed placement; Sneha et al. [18] developed an agricultural robot capable of performing multiple tasks such as plowing, seeding, and even pesticide spraying; Soyoye et al. [19] describe a manually operated seeder for corn with a vertical plate incorporated in the dosing system; Swapnil et al. [20] developed a low-cost seeding machine for various crops for smallholder farmers that incorporates a seed disk mechanism and a wheel for seed metering; Kumar et al. [21] designed and manufactured an intelligent robot for seed sowing focused on reducing human effort by incorporating an individual dosing system using stepper motors. However, most of these seeding solutions have been designed to work in very specific conditions, do not allow coupling to different power sources, and their functionality only applies to the specifically developed technology and not to generic robotic platforms.

Implementing automation and agricultural robots in developing countries presents unique challenges. Although cost is a significant hurdle, automation technologies also need to be inclusive, accessible, and responsive to local needs while considering improvements in environmental sustainability [22,23]. The design process must ensure that prototypes align closely with users’ needs and operational requirements, translating preferences and agronomic demands into technical specifications. Quality Function Deployment (QFD) analysis is a structured approach that could be used for this purpose, as it enables farm equipment designers to focus on critical factors and allows an objective evaluation of the seeders’ performance against user-defined criteria, ultimately guiding improvements that enhance usability, reliability, and productivity [24,25].

For all the above, the authors have embraced these challenges, and the present study focuses on the design, development, and comparison of two maize seeder prototypes with different seed meter actuating mechanisms that are adaptable for different agricultural robots. The design was made using a QFD approach, and the agrotechnical evaluation of the quality parameters of maize crop establishment (such as field capacity, planting depth, and seed distribution in the planting line) was performed, providing insights into the potential for future improvements. Above all, by leveraging cutting-edge technologies such as robotics with low-cost prototype designs, this research seeks to make farming more accessible and attractive to younger generations often discouraged by the labor-intensive nature of traditional farming practices.

2. Materials and Methods

2.1. The Agricultural Robot as Operating Platform

The seeder prototype concepts were developed to work in combination with a modified unmanned ground vehicle (UGV) platform designed by the Agricultural Engineering and Integral Water Use Postgraduate Program at the Chapingo Autonomous University, Mexico, and described in [26,27]. This UGV robot (Figure 1) consists of (A) an aluminum chassis; (B) roller and sprocket chain transmission in a skid-steer arrangement for power transmission; (C) two Ling Ying^® model MY1016Z 250 W brushless DC electric motors to provide movement; (D) two LiFePO4 Li-ion batteries with a nominal voltage of 12.8 V and capacity of 20 Ah, Lithium Power^® model URB 12200; (E) four 25 cm diameter wheels connected to oscillating arms built in a tubular steel profile with spring suspension; (F) a dual 60 A Sabertooth model controller [28]; (G) an Arduino microcontroller model ATmega2560; and (H) a FlySky FS-IA6B 6-channel receiver. The agricultural robot is equipped with a radio control system (FlySky FS-i6) that allows the operator to navigate and maneuver the equipment remotely, controlling direction and speed in real time.

The UGV used in this study was selected due to its compact size to be able to operate in smallholder farmer fields. It has dimensions of approximately 85 × 56 × 53 cm (L × W × H) and a weight of 45 kg, which allows inter-row operations when crops are established and minimizes soil compaction. Additionally, it uses electric motors to independently control the speeds of its wheels, allowing precise turns with a minimum radius. Finally, being an electric vehicle eliminates the need for fossil fuels, contributing to reduced CO₂ emissions and allowing the integration of renewable energy sources, such as solar panels.

2.2. QFD Analysis

The design process was conducted by implementing a Quality Function Deployment (QFD) methodology, which is considered a planning tool that allows for conveying product characteristics and summarizing and converting user feedback into information for designers [29,30]. The QFD analysis was implemented in five phases: (1) a user consultation for extraction and identification of requirements, (2) prioritization of requirements, (3) identification of parameters for the design, (4) identification of relationships between requirements and parameters, and (5) correlation of design parameters.

The procedure for obtaining the requirements was carried out with a group of 37 students from the Agricultural Mechatronic Engineering program at the Chapingo Autonomous University, Mexico (UACh). This group has experience in the execution of agricultural processes and is being trained in the development of electronic systems. This allowed us to accurately define the requirements of future end-users. In addition, by including them in the design process, it is possible to encourage youth to get involved in the development of solutions for agriculture. For the consultative extraction and identification of the design requirements, the students were given one-on-one questionnaires, where they were asked to name the characteristics to be considered for the development of a maize seeder implement coupled to the agricultural UGV robot as previously described.

Subsequently, the prioritization of the requirements was performed with two working groups with expertise in the design of agricultural technologies: the robotics team of the Agricultural Engineering and Integral Water Use Postgraduate Program (IAUIA) of the UACh and the Smart Mechanization Unit team of the International Maize and Wheat Improvement Center (CIMMYT). The first group has extensive experience working in the development of mechatronic solutions. In contrast, the experience of the second group concerns the development of agricultural machinery from a more practical and field perspective and, in particular, on small-scale appropriate equipment. Members of both groups were asked to individually rate the requirements previously obtained using the following scale: 1 → not important at all; 2 → of little importance; 3 → quite important; 4 → very important; and 5 → extremely important.

The exercise produced two maize seeder prototypes concept, as both groups used the “house of quality” diagram (see Figure 2 in Section 2.2), which is a key component used to translate user requirements into engineering targets for product design [31,32]. Later, a morphological analysis was carried out to select alternative components and solutions for the seeders’ sub-functions. Lastly, each seeder prototype was 3D-modeled using SolidWorks^® version 2023 software, after which, both seeders were constructed based on the 3D model designs, and preliminary tests were performed for each seeder to ensure its correct functioning (see Supplementary Material File S1).

2.3. Agrotechnical Evaluation of Seeder Prototypes

The agrotechnical evaluation was carried out at the CIMMYT facilities located in El Batán, Texcoco, State of México. According to the World Reference Base (WRB), the plot soil classification is a Phaeozem Haplico. The soil preparation included two passes with a four-wheel-tractor-pulled disc harrow and one pass with a power tiller to break up the clods of the soil. The agrotechnical evaluation of both seeders consisted of five rows of 25 m long to be sown with three repetitions. The operator maneuvered the UGV robot and seeder attached by means of a remote radio controller along chalk lines previously marked on the soil with a 75 cm distance between them (equivalent to the row distance). The desired distance between seeds was 18 cm (which would result in a plant density of 74,074 plants/ha), and Asgrow^® Albatros hybrid maize seeds were used for sowing.

2.3.1. Effective Field Capacity

Due to the restricted maneuvering conditions outside the plot area and the difficulty involved in constantly performing the robot’s turns, only the effective times were considered, i.e., from the start to the end of each 25 m row section. The times of each of the sowing paths were recorded using a digital stopwatch. The effective work capacity was determined with the following equation:

$$C_{eff}={\displaystyle \frac{A}{T}}$$

(1)

where $C_{eff}$ is the effective field capacity (ha h⁻¹), $A$ is the field coverage (ha), and $T$ is the actual time of operation (h).

2.3.2. Plant Spacing, Sowing Depth and Seed Singulation

Both plant spacing and sowing depth were determined 20 days after the sowing date. For the plant spacing, six sections of 3 m were selected per repetition (18 sections in total for each seeder), and the distances between plants were measured with a flexometer. For the sowing depth, three plants per row were randomly selected, and the soil was gently removed until the seed was found. The distance from the soil surface to the midpoint of the seed was measured.

To characterize seed singulation, the methodology described by ISO 7256/1-1984 [33] was used. Failure (F), double (D), and acceptable (A) indices were calculated with respect to the initially established theoretical spacing value $\left(x_{ref}\right)$ where values less than ${0.5x}_{ref}$ were considered double, values greater than ${1.5x}_{ref}$ were considered failures, and values located within this range were considered acceptable. Cases where there were two consecutive values less than ${0.5x}_{ref}$, were considered triples, and, where three or more occurred, they were considered multiple.

2.3.3. Electric Current Consumption and Battery Range

The electric current values of the robot were recorded during the sowing runs for each of the seeders. For this, a Steren^® MUL-115 clamp meter was connected to the main power supply line of the system. The data transfer was done using the official application of the multimeter connected via Bluetooth to a smartphone. To determine the battery range of the robot, the following equation was used:

T = C/I

where T is the battery range expressed in hours (h), C is the battery capacity expressed in ampere-hours (Ah), and I is the average electric current obtained from the clamp meter readings (A).

3. Results

3.1. QFD Analysis

3.1.1. Extraction and Prioritization of User Requirements

During the QFD analysis, 87% of the students mentioned that precision in sowing through a well-distributed placement and with the minimum number of errors, such as double seeds or lack of seed placement, is of the utmost importance. They also mentioned that, as the device is intended to work together with a specific robot, its design should be compatible with the robot’s software, incorporating microcontrollers for possible communication between elements. Some students also indicated that it should be constructed of inexpensive materials and, above all, that it should not require very specialized maintenance. The results of this session were grouped into general categories and are presented in

Table 1. Requirements expressed by users for a seeder coupled to a farming robot.

General Category	Specific Requirement
Seeding precision	Homogeneous seed placement Adequate sowing depth Seed singulation
Automated	Programmable seeding parameters Remote adjustment of seeding parameters Variable input dosing
Low cost	Construction with economical materials Economical maintenance Spare parts readily available
Robot compatibility	Communication between robot and seeder Continuous feedback to the user
Ease of use	No need for advanced technical knowledge Quick and easy attachment to the agricultural robot Simple steps to start the system
Robustness	Materials resistant to field conditions Operability on uneven terrain
Weight	Constructed of lightweight materials Hoppers with adequate capacity Maximum battery life

.

After obtaining the user requirements, both the IAUIA and CIMMYT groups prioritized the specific requirements, as mentioned in Table 1. Averages were obtained for each category, and the results are presented in

Table 2. Results of the prioritization of requirements for each working group and their level of importance.

Requirements	Relevance for IAUIA	Relevance for CIMMYT
Seeding precision	4.7	4.8
Automated	4.5	1.8
Low cost	3.6	4.4
Robot compatibility	5.0	4.5
Ease of use	2.4	4.3
Robustness	4.2	5.0
Weight	4.5	3.2

.

Due to the different approaches used by the two groups, the prioritization phase yielded different results. The most valued requirement as chosen by the IAUIA group was compatibility with the robot, followed by accuracy, automation, and weight. On the other hand, while the CIMMYT group also prioritized precision, the most valued requirements chosen were ease of use, robustness, and low acquisition and maintenance costs. Given the marked differences in the priorities of each working group, and the fact that a single design would not be able to fully satisfy the needs of both groups, it was agreed that two different seeders would be designed, the first aimed at being a more sophisticated solution and the second at being more robust and easier to maintain.

3.1.2. House of Quality

At this stage, both groups defined the most important elements (or quality characteristics) contributing to the fulfillment of the established user requirements. To ensure seeding precision, the quality characteristics defined by the CIMMYT group were an efficient metering system, a furrow opener that allows adequate depth, and a system to cover the seed and ensure its contact with the soil. Similarly, the IAUIA team considered of top importance the integration of sensors and actuators in the dosing system for a precise seeding. As such, each of the user requirements were transformed into quality characteristics and included in a house of quality diagram shown in Figure 2.

Horizontal rows list user requirements and their weight/importance according to the prioritization of the requirements, while the vertical columns present quality characteristics. The cells of the matrix contain a symbol indicating the degree of correlation between the requirement and the quality characteristic: the solid circle figure ( = 9) represents a strong relationship, the empty circle (○ = 3) represents a moderate relationship, and the triangle (▲ = 1) represents a weak relationship. This level of correlation was defined by each of the groups according to the previously established levels of importance.

To obtain the relative weights of the quality characteristics (shown in the last row), the procedure described by Belhe and Kusiak [32] was followed, multiplying the vector of relative weights of the requirements by each of the vectors of the quality characteristics. For both teams, the quality of the materials was the most important quality characteristic; however, for the IAUIA team, the need to integrate sensors and actuators into the design weighed more than the need for building the prototype with locally available materials for the CIMMYT group.

3.1.3. Morphological Analysis and Selection of Components

The results obtained from the two houses of quality allowed moving on to the next phase, where the different alternatives for the main components of the seeders to be developed were analyzed. The components and the possible solutions considered are presented in

Table 3. Seeder components and possible alternatives for their construction.

Component	Possible Solutions
Chassis material	Steel	Aluminum	Steel + Aluminum	Plastic	Carbon fiber
Agricultural robot attachment	1-point hitch	2-point hitch	3-point hitch	>3-point hitch
Type of seed metering system	Spoon-wheel metering device	Vertical plate with cups metering system	Hole seed-metering wheel	Horizontal plate	Pneumatic
Actuating mechanism of the seed metering system	Sprocket and roller chain mechanical mechanism	Shaft-driven mechanical mechanism	Belt-driven mechanical mechanism	Stepper motor electronic mechanism	Servo-driven electronic mechanism
Furrow opener	Tine type	Disc coulter type	Runner type	Punch type	Concave disc type
Seed hopper material	Plastic	Galvanized steel	Stainless steel
Seed covering and firming device	Single wheel type	Slide type	Finger wheel type
Seeder lifting mechanism	Articulated parallelogram type	Frame bar pivot type	Rigid frame type

.

The selection of components was performed by each group separately. For increasing robustness, the CIMMYT working group opted for mechanical transmission via a pair of sprockets and a roller chain as the actuating mechanism of the seed metering system, and the material selected for the chassis was steel. To increase precision, the seed metering system selected was the hole seed-metering wheel because, according to the experience of the team, the same pressure exerted by the weight of the seed increases the placement of the seed in the holes of the metering wheel. Additionally, the seeder lifting mechanism selected was the articulated parallelogram type because, from the CIMMYT’s team perspective, this mechanism allows the entire seeder body to move vertically to better adapt to the irregularities present in the soil, allowing for greater homogeneity in seeding. To reduce the weight of the seeder, the hopper material selected was plastic, and the seed covering and firming device type was a single wheel. The furrow opener selected was a tine type because it does not require additional vertical force to penetrate the ground. Lastly, for a simple and quick coupling, a 1-point hitch was selected.

The IAUIA working group first selected the components for the actuating mechanism of the seed metering, choosing to do it with a stepper motor due to the level of precision it can provide to perform the rotation. Derived from this selection, a vertical plate with cups metering system was chosen since it requires the least effort to rotate compared to the other mechanisms. To achieve a light and resistant structure, the material selected for the chassis was steel and aluminum, and the hopper was plastic. In the same sense, to reduce weight, a single wheel type was selected, and, for the seeder lifting mechanism, a frame bar pivot type was selected. As for the hitch, the IAUIA group opted for the same tine type to facilitate ground penetration.

3.2. Computer-Aided Design of the Seeders

SolidWorks 2023 was used to draw the components of each seeder in 2-dimensional and 3-dimensional views. Figure 3 shows the rendered 3-D models of the CIMMYT and IAUIA seeder prototypes. Additionally, technical drawings for both seeders and exploded views of their components are provided in Supplementary Material File S2. The mechanical seeder designed by the CIMMYT working group consists of (A) a robot–seeder connecting structure that separates the seeder from the robot to avoid contact between the drive wheel and the frame; (B) an electric piston; (C) a steel chassis that functions as the main structure, taken from the Terradonis commercial seeder model JPH [34]; (D) a seed storage hopper with a capacity of 2.8 kg taken from the GIMEX commercial equipment [35]; (E) a forward-curved furrow opener; (F) a single compacting wheel to ensure uniform pressure on the soil near the seed; (G) a hole seed-metering wheel taken from the GIMEX commercial equipment; (H) a drive wheel to provide movement to the transmission system; and (I) a mechanical transmission system consisting of two sprockets mounted on the drive wheel axis and the seed metering system axis, and a roller chain that allows for transmitting the movement. Thus, as the robot moves, the metering plate rotates and releases a seed at a set distance, which can be modified according to the size of the sprocket.

The seeder with electronic metering system designed by the IAUIA group (Figure 3-bottom) consists of (A) a robot–seeder connecting structure; (B) an electric piston; (C) a steel and aluminum chassis; (D) a plastic seed hopper with a capacity of 0.9 kg; (E) a near-vertical furrow opener; (F) a compacting wheel for both compaction and depth control; (G) a vertical plate metering system; (J) a stepper motor connected to the metering system shaft to control seed metering; (K) a rotatory encoder to measure the rotation of the compaction wheel; and (L) a microcontroller box with an Arduino Atmega2560 and two 5 and 12 V batteries.

3.3. Hardware and Software Description for the Seeder with Electronic Metering System

The electronic system for the IAUIA seeder consists of an Arduino microcontroller model ATmega2560, an E50S8-5000-3-T-5 rotatory encoder, a model TB6600 driver, a NEMA 23 12 V KM-K stepper motor, and a 12 V battery. The connection diagram of the system is shown in Figure 4a. The code was developed in the Arduino IDE and loaded into the ATmega2560. Figure 4b shows the flowchart of the implemented algorithm. When starting the system, the microcontroller sets the configuration and defines the variables and parameters, such as the desired distance between seeds (D_seed), the number of holes of the metering plate (H_p), the radius of the compaction wheel (r), the number of pulses of the encoder (P_e), and the number of steps of the stepper motor (S_m). Then, it calculates the equivalent distance of one encoder pulse (D_pp) and the target distance (D_t) with the following expressions.

$$D_{pp}={\displaystyle \frac{2 \pi r}{P_e }}$$

$$D_t=\frac{D_{seed}}{\left(\frac{S_m}{H_p}\right)}$$

The main program loop (Void Loop) continuously monitors the state of the encoder, calculates the distance traveled by the seeder (Td), and compares the distance traveled with the target distance. If there is a change in the encoder status (i.e., if the seeder has moved), then it accumulates the number of pulses received. The distance traveled is found by multiplying the number of accumulated pulses times the equivalent distance of one encoder pulse. Finally, if the distance traveled is greater than or equal to the target distance, the microcontroller sends N pulses to the motor driver in order to rotate one step. In this way, when the seed drill has traveled the desired seed distance, the stepper motor has rotated by the corresponding fraction so that the metering unit can release one seed.

3.4. Agrotechnical Evaluation of the Seeders

The agrotechnical evaluation of the seeders was performed in July 2023 (Figure 5). The robot with the CIMMYT mechanical seeder attached was able to achieve a slightly higher working speed than when the IAUIA electronic seeder was attached (

Table 4. Agrotechnical evaluation results obtained for both seeders.

Parameter	Mechanical-Driven Metering System Seeder (CIMMYT)	Electronic-Driven Metering System Seeder (IAUIA)
Working speed (km/h)	2.0 ± 0.2 (n = 15)	1.6 ± 0.4 (n = 15)
Effective working capacity (ha/h)	0.15 ± 0.02 (n = 15)	0.12 ± 0.03 (n = 15)
Plant spacing (cm)	17.1 ± 8.8 (n = 305)	12.4 ± 9.2 (n = 417)
Sowing depth (mm)	34.8 ± 5.0 (n = 45)	53.6 ± 9.4 (n = 45)

). Therefore, the working capacity of the former was also higher, enabling the sowing of one hectare in 6.7 h, compared with 8.8 h for the electronic seeder. In terms of the distance between plants, the mechanical seeder achieved an average distance of 17.1 cm, which was closer to the desired plant distance (set at 18 cm). The seed singulation indexes obtained for the mechanical seeder were 66% acceptable, 21% multiple, and 13% failures, whereas, for the electronic seeder, 45% were acceptable, 49% were multiple, and 6% were failures.

The results of the spatial distribution of seeds after planting are shown in Figure 6 and showed that the mechanical seeder produced a greater density of seeds with a spacing closer to 18 cm (the desired plant distance), whereas the electronic-driven seeder yielded smaller densities due to the number of cases in which more than one seed was released (cases where there were multiple seeds).

3.5. Electric Current Consumption and Battery Range

Figure 7 shows the current-consumption values of the agricultural robot during sowing with both seeders. It is evident that there were some variations in the amperage during the runs. The higher values represent the peaks of electrical consumption that were mainly due to higher efforts by the robot when moving forward, probably due to a higher resistance generated by the furrow opener. The average electric consumption for the electronic-driven seeder was 28.3 A, while, for the mechanical seeder, it was 22.4 A. Using these average values, the battery range of the robot would then be 1.4 and 1.8 h for the electronic and mechanical seeders, respectively.

4. Discussion

The present study aimed to design and evaluate two maize seeder machines coupled to a robotic vehicle. The design process was implemented through a QFD analysis methodology but by two groups with different expertise. The IAUIA group decided to build an electronic system compatible with the agricultural robot, whereas the CIMMYT group opted for a robust, durable, and practical design, leaving aside the compatibility. This could be seen throughout, from the prioritization of requirements and conceptualization of the seeders to the selection of the design components in the morphological analysis. The mechanical seeder has a rigid chassis that allows the depth control wheel to be kept immobile, which makes the seeding depth quite homogeneous in well-prepared soils without many clods. However, a limitation is that this same rigidity can be a disadvantage in irregular soils, since obstacles can lift the furrow opener and modify the depth of the seed, varying with the unevenness of the soil. On the other hand, the joints integrated in the electronic seeder (one in the hitch and the other in the connection of the compacting wheel) provide flexibility, allowing for the compacting wheel to be raised and the furrow opener to go deeper. However, as there is no system to limit this flexibility, on uneven ground, the furrow opener can go too deep.

The depth of the furrow openers significantly impacts soil resistance and traction force during sowing. Deeper furrow openers generally lead to greater soil disturbance and increased power demand, especially in compacted soils [36,37]. The articulated arms of the electronic seeder allowed for a greater sowing depth, which can be advantageous compared to the mechanical seeder; however, this generated greater resistance, causing the working speed and effective working capacity to be lower than with the mechanical seeder. This could also explain the higher electric current consumption with the electronic seeder.

The electronic-driven seeder’s high percentage of multiple seed depositions may be due to the fact that vertical plate-type metering systems are usually more sensitive to vibrations caused by ground irregularities. Also, since the seeder lifting mechanism selected was a frame bar pivot type, as the seed deposition is deepened further, the metering system tilts forward, affecting seed singulation.

The seeders developed in this research offer some advantages over other existing systems. For example, although the crank-slider mechanism described by Azmi et al. [17] reported a 92% efficiency in sowing, the vertical plate with cups from the electronic-driven seeder achieved higher efficiency with only 6% failures. More so, the seed spacing of the first mentioned is fixed and does not provide the flexibility that the electronic-driven metering system seeder has, enabling the easy adjustment of the sowing distance in relation to the desired plant density. The robot developed by Sneha et al. [18], which integrates multiple functions, such as plowing, seeding, and monitoring, only operates with its specifically designed robotic platform, limiting its broader adoption, while the presented seeder prototypes and their modular design allow for the use of a variety and more generic robotic platforms. The manual maize seeder with vertical plate developed by Soyoye et al. [19] demonstrated an optimum working speed of 4.97 km/h and an efficiency of 89.7% but only works manually depending on human labor and cannot be coupled to a robotic platform. The presented mechanical and electronic seeder prototype did work significantly slower (with average working speeds of 2.0 km/h and 1.6 km/h, respectively) but could be operated remotely and in an automated fashion with similar working efficiencies (87% and 94%, respectively) making it a very attractive solution in labor-scarce conditions and with minimal physical effort for the operator.

5. Conclusions

In this study, two maize seeders with differently driven metering systems (one mechanical-driven and one electronic-driven) were developed and evaluated in order to satisfy different user requirements and to expand the number of alternative integrated solutions for working with robotic vehicles. The results indicated that the seeder with the conventional mechanical-driven metering system outperformed the electronic-driven metering system in terms of seed distribution, with a higher rate of acceptable seed singulation and a more homogeneous seeding depth. Specifically, the mechanical seeder achieved an effective working capacity of 0.15 ha/h compared to 0.12 ha/h for the electronic system, as well as a higher battery range of the robot with 1.8 h and 1.4 h, respectively.

This improved performance in real field conditions emphasizes the importance of design considerations. The mechanical seeder designed by the CIMMYT working group was characterized by its robustness and ease of use—characteristics that align well with users seeking cost-effective, low-maintenance solutions. However, due to the approach chosen by the working group, this seeder does not allow the full potential of the agricultural robot (such as continuous communication or feedback) to be exploited. In contrast, while the electronic seeder is innovative, the selection of components and variable soil conditions could affect the quality of seed placement, requiring further refinement to optimize its design. In addition, the importance of field-testing during design to ensure adequate performance in real conditions is highlighted.

Although the evaluation of the seeder prototypes was done under real field conditions, inclement weather conditions affected adequate open field conditions during the experimentation period, resulting in a reduced number of replicate data points reducing the generalizability of the results. Therefore, to increase robustness and reliability in future studies, it is recommended to increase the number of replicates, including under varying field conditions. Overall, both sets of proposed seeders represent advances in agricultural automation and may help to attract younger generations of farmers to the field. That said, the design and research of future equipment should take a more holistic approach to encompass both field experience and the advantages of mechatronic systems.