Design and Tests of a Large-Opening Flexible Seedling Pick-Up Gripper with Multiple Grasping Pins

Abstract

The pick-up gripper, as a core component of automatic transplanting systems, presents challenges in reliably grasping seedlings. In this study, a large-opening flexible seedling pick-up gripper was designed based on standard trays and seedling characteristics. Structural design and force analysis of the grasping mechanism were conducted to develop a functional prototype. As this represented the first prototype of this new gripper, multi-factor orthogonal tests and performance tests under local conditions were performed to evaluate its grasping effectiveness. It was found that the end diameter of the pick-up pin and the extraction speed for lifting plug seedlings vertically had the most significant effects, followed by the penetration depth and grasping force. The optimum grasping effectiveness was achieved when the end diameter of the pick-up pin was 1.2 mm, the penetration depth in the top straight line of the pick-up pin was 40 mm, the grasping force for squeezing root lumps was 0.4 MPa, and the extraction speed for lifting plug seedlings in a vertical direction was 900 mm/s. For typical vegetable seedlings, the average success rate in transplanting was up to 95%. Under the combined actions of penetrating, squeezing, and extracting operations, plug seedlings could be efficiently picked out for efficient transplanting.

1. Introduction

Vegetables are an essential food in the daily diet [1]. Therefore, the effective supply of vegetables is a primary issue for everyone’s livelihood. Currently, China is the largest producer and consumer of vegetables in the world, with an output of 23.03 million hectares and a yield of 828.68 million metric tons in 2024. In response to the rising demand for agricultural products and diminishing resources, the adoption of automated agricultural systems offers a promising solution to enhance productivity and reduce labor costs [2]. Compared with direct seeding for vegetable production, seedling transplantation has numerous overall benefits, such as increased seedling survival and vegetable yield, a short growing period, and so on [3]. With the development of seedling technology, research on ensuring precision in automatic transplanters and their high-speed components has been promoted to reduce the labor and time required for continuous transplanting operations [1,4].

Various work processes and mechanical structures for automatic transplanting have been designed and optimized to meet the requirements of extracting, transferring, discharging, and planting seedlings [5]. Choi et al. [6] designed a sophisticated five-bar mechanism and a pincette-type driver with two pick-up pins to extract plug seedlings from their growth cells and transfer them to the place where they would be planted into the ground. The performance tests showed that this pick-up device could extract 30 seedlings per minute with a success ratio of 97% using 23-day-old seedlings. Although the working efficiency was low, it proved the feasibility of automatic transplanting. The use of an automatic transplanter could reduce the labor requirement of seedling transplantation by carrying out repetitive tasks in an accurate and reliable operating mode. Kang et al. [7] developed a two-row transplanter to promote the mechanization of vegetable transplanting. This transplanter was able to automatically supply seedling trays and transport picked seedlings to the planting hopper. It was further found that the fork-type manipulator was suitable for picking up seedlings. For efficient transplanting, Zhou et al. [8] proposed a simple transplanting mechanism structure with differential internally engaged non-circular planetary gear trains. In theory, this non-circular rotating trajectory could meet the requirements of multi-objective transplanting. Yue et al. [9] designed a reciprocating seedling-picking device driven by full air pressure through mechanical analysis between the seedling-picking claw and the seedling pot. These studies provide feasible solutions for the innovative design of automatic transplanters for vegetable seedlings.

The successful integration of seedling transplanting into a robot requires an operational gripper. Alongside increased research developments in transplanting technologies, the core working gripper needs to be compatible with the seedling characteristics if the full potential of automation is to be realized [10,11]. The existing designs range from simple grippers to custom end-effectors with the capabilities of dexterous hands. Ryu et al. [12] developed an end-effector using a pneumatic system to successfully grip, hold, and release the plug seedling during transplanting. The preliminary test results showed that the shovel-type fingers had optimal transplanting success rates. Khadatkar et al. [13] designed a robotic transplanter for plug seedlings that worked by grasping their plant stems. However, most vegetable seedlings are so fragile that they are not suitable for this transplanting operation. Gao et al. [14] designed an inclined inserting-type picking manipulator for the automatic transplanting of plug seedlings. The manipulator could directly extract the seedling after the steel needles were inserted into the growth medium at an angle. The optimal combination of operation parameters was obtained to achieve the maximum value of the limit break force. Tong et al. [15] analyzed the working process of a spade end-effector with four shovel-type pins in inserting and clamping the substrate around the tray cell. However, the size and position accuracy of the shovel-type pins must be able to grasp the root lump in the cell while causing little damage. Tian et al. [16] proposed a multi-needle seedling gripper that was able to grab as much substrate as possible from the tray cell. Taking the medium integrity rate as an evaluation index, the simulation analysis of EDEM was used to explore the influence of different needle diameters, insertion depths, and working speeds on success. The design and research conclusions of the above references have provided important references for the reliable design of pick-up grippers. In general, the main challenge in grippers, apart from further streamlining the structure and testing the applicability to various seedlings, is clarifying the working mechanism of the mutual coordination between the grasping theory of the end-effector and the internal structural deformation of the root lump.

This study aimed to develop a flexible pick-up gripper based on the principles of mechatronics and pneumatics. The growth characteristics of the plug seedlings and the grasping effects of their root lumps were fully considered in the design to avoid excessive damage. Multi-factor combination tests were used to optimize the key structural parameters and further analyze the actual working performance. This research could provide innovative ideas for the development of precise seedling transplanting devices.

2. Materials and Methods

2.1. Overall Design and Analysis

2.1.1. Mechanical Structure and Working Principle

Plug seedlings are cultivated in independent tray cells, which need to be grasped using a specialized manipulator [17]. Therefore, a structurally simple gripper design is required to achieve low loss and highly efficient removal of seedlings from their growth trays for further growth and development. As shown in Figure 1, a pincette-type seedling pick-up gripper was designed with a large-opening cavity to accommodate fragile seedling plants. In terms of specific mechanical structures, the gripper incorporates multiple fine pins to flexibly grasp the root lumps of seedlings (Figure 1: red straight arrow). The design consisted of an industrial pneumatic fixture, two innovative cylinder fingers, and several connecting parts. Two cylinder fingers were symmetrically fixed to the limited bilateral pinching jaws of the pneumatic fixture by the connecting parts. Their opening degree was adjusted by using the countertop bolts. Each cylinder finger was composed of a double-acting mini cylinder, two pick-up pins, a positioning plate, and an adjustment plate. The pick-up pins were welded as a fork, of which the opening size was changed according to the tray cell dimensions. The shank of the pick-up pins was connected to the piston rod of the mini cylinder. Their penetrating tips were designed to pass through the adjustment plate that was bolted to the positioning plate. These pick-up pins were designed to extend in and out. With the help of the reciprocating opening and closing operations of the pneumatic fixture (Figure 1: red curved arrow), the mini cylinder activated the fork-type pins to grasp seedlings and then release them at the discharging position.

The working principle of the new seedling pick-up gripper is described below and in Figure 2:

-

At the beginning of a working cycle, two cylinder fingers pull their fork-type pins back. The pneumatic fixture is bounced off, causing the cylinder fingers to open. In this case, the pick-up gripper is moved by the automatic transplanting manipulator towards the seedling’s root lump with little damage to its upright plant (Figure 2a).

-

When the pick-up gripper is adjusted to be normal to the tray cell surface, the cylinder fingers push the fork-type pins to obliquely penetrate the root lump along both sides of the tray cell (Figure 2b). Then, the multiple pin assembly moves toward the seedlings until it can grasp the maximum number of their root lumps.

-

The pneumatic fixture is driven to close, causing the fork-type pins to squeeze the root lump gradually as they penetrate the tray cell (Figure 2c). With the help of this operation, the pick-up gripper applies some grasping forces to hold the seedling firmly. Finally, the pick-up gripper is moved up to extract the seedling from the tray cell (Figure 2d).

-

After extracting the seedling, the automatic transplanting manipulator moves the pick-up gripper to transfer the seedling to the place where it is to be transplanted. As opposed to grasping seedlings, the pick-up gripper efficiently discharges the seedling through the combined operations of opening the cylinder fingers and then retracting the fork-type pins (Figure 2e). In this way, the seedling falls with the inertia action while standing upright.

2.1.2. Dimension and Force Analysis of Key Components

The grasping mechanism was developed through iterative theoretical calculations. As shown in Figure 3a, the overall dimensional parameters of the fork-type pins were determined based on the inner size of the tray cell. The slightly flexible injection molded polystyrene plastic trays are widely used for vegetable seedling production, with overall dimensions of 540 mm length × 280 mm width. They contain 50 cells, 72 cells, and 128 cells in 5 × 10, 6 × 12, and 8 × 16 arrangements, respectively. The shape of each cell is like an inverted truncated pyramid. For the maximum penetration depth into the bottom of the tray cell, the fork-type pins were designed to meet the following requirements:

$$\left\{\begin{array}{c}{d}_0=b-∆d_0\\ d_1=\sqrt{{\displaystyle \frac{{\left(a-b\right)}^2}{4}}+h^2}-∆d_1\end{array}\right.$$

(1)

where d₀ is the width of the fork-type pins, mm; d₁ is the effective penetration depth of the fork-type pins, mm; Δd₀ and Δd₁ are the safety interval distances between the tip of the fork-type pins and the tray cell, respectively, mm; and a, b, and h are the top side, bottom side, and height dimensions of the tray cell, respectively, mm.

Taking the 128-cell tray as the design reference, the cell dimensions are 32 mm top × 13 mm bottom × 42 mm height. Previous studies concluded that the optimal safety interval distance was 2 to 3 mm for grasping the maximum amount of root mass [6]. Therefore, the width between the two pin tips was set to 10–11 mm. The diameter of the pick-up pin should also be considered to ensure a reasonable interval. According to Equation (1), the maximum penetration depth into the bottom of the tray cell was calculated to be less than 43.06 mm without considering the safety distance. The other dimensions which were suitable for various types of plug trays might be adjusted by the use of different connecting parts. In this way, the effective stroke of the mini cylinder could be determined to push the fork-type pins to obliquely penetrate the root lump.

The schematic diagram of the grasping mechanism is shown in Figure 3b. Further, a large-opening cavity among two pairs of pick-up pins was designed to accommodate the seedling plant (Figure 3: blue dotted line). According to the stem height of the seedling plant, the length of BC for the positioning plate was defined as follows:

$$l_{BC}\times \cos \beta \ge H$$

(2)

where l_BC is the length from the assembly point of the mini cylinder to the penetration point of the pick-up pins, mm; β is the grasping angle of the pick-up pins, °; and H is the stem height of the seedling plant, mm.

In the direction of the seedling leaf width, the length of AB for the positioning plate was calculated as follows:

$$\left\{\begin{array}{c}{l}_{{O_{1}O}_2}+2\times \left[l_{AB}\times \cos \beta +l_{O_{2}A}\times \sin \left(\theta +\alpha -\beta \right)\right]\ge {\displaystyle \frac{L}{2}}\\ l_{AB}\times \cos \alpha +l_{O_{2}A}\times \sin \theta -l_{BC}\times \sin \alpha \le {\displaystyle \frac{a-l_{{O_{1}O}_2}}{2}}\end{array}\right.$$

(3)

where l_AB is the length from the end point of the industrial pneumatic fixture to the assembly point of the mini cylinder, mm; l_BC is the length from the assembly point of the mini cylinder to the penetration point of the pick-up pins, mm; l_O1O2 is the spacing of the rotation center of the grasping plates of the industrial pneumatic fixture, mm; l_O2A is the length from the right center to the end point of the industrial pneumatic fixture, mm; a is the top width of the tray cell, mm; α and β are the penetration angle and the grasping angle of the pick-up pins, °; θ is the angle between the line O₂A and the vertical direction as the industrial pneumatic fixture opens, °; and L is the leaf width of the seedling plant, mm.

As the pick-up pins were inserted into the root lump along the cell wall, the grasping angle was close to the conical degree of the tray cell ranging from 10° to 12°. The optimal stem heights that were suitable for mechanized transplanting ranged from 90 mm to 115 mm for typical seedlings, such as cucumber (Cucumis sativus L.), tomato (Lycopersicon esculentum Mill.), and pepper (Capsicum annuum L.) [10,11]. Therefore, the length of l_BC was estimated to reach about 117 mm based on the maximum stem height. After rounding, l_BC could be designed at the value of 120 mm. The mini industrial pneumatic fixtures that were readily available on the market had such size parameters. The spacing range of l_O1O2 was about 10–20 mm, the length ranges of l_O2A and l_O2O were about 20–35 mm, and the angle value of θ was nearly 9°. Therefore, the length of l_AB was estimated to range from 16 mm to 18mm according to Equation (3). Taking its average value, the length of l_AB was designed as 17 mm. Finally, the overall dimensions of the connecting parts could be designed based on the industrial pneumatic fixture used.

As shown in Figure 3b, the adhesion forces between the root lumps and the cell walls were established since the limitation of the seedlings’ growth space might lead their roots to coil around the perimeter of the tray cells [18]. Therefore, the pick-up pins must apply certain grasping forces to overcome the growth adhesion forces. In the vertical direction, the pulling force generated by the grasping operation should be at least equal to the seedling gravity and the equivalent adhesion forces.

$$F_P=G+{\sum }_{VD}\left(F_{A1}+F_{A2}+F_{A3}\right)$$

(4)

where F_P is the pulling force generated by the grasping operation, N; G is the gravity of the seedling, N; ${\sum }_{\mathit{VD}}{\mathit{F}}_{\mathit{A}}$ are the equivalent adhesion forces in the vertical direction, N; and F_A1, F_A2, and F_A3 are the adhesion forces between root lumps and different cell walls, N.

It was assumed that the seedling’s root lump was a homogenate, and the grasping forces on the pin surfaces were uniform. To simplify the force analysis, the pulling force was generated by the grasping operation of the pick-up pins in the non-deformable state. Thus, the pulling force was approximately equal to the resultant force of the normal force and the frictional force at the penetration sides of the pick-up pins.

$$\left\{\begin{array}{c}{F}_P=2\times l_{CD}\times \left(F_N\times \sin \beta +F_f\times \cos \beta \right)\\ F_f=\mu \times F_N\end{array}\right.$$

(5)

where F_P is the pulling force generated by the grasping operation, N; F_N is the normal force generated by the grasping operation at the penetration sides of the pick-up pins, N; F_f is the frictional force generated by the grasping operation at the penetration sides of the pick-up pins, N; L_CD is the effective penetration depth of the pick-up pins, mm; β is the grasping angle of the pick-up pins, °; and μ is the friction coefficient between the root lumps and the pick-up pins.

It could be seen from the analysis of the grasping mechanism that the single grasping action had a lever effect (Figure 3b). According to the force analysis of the lever fulcrum at the point of O₂, the following mechanical equation was obtained:

$$\left\{\begin{array}{c}N\times l_1=F_Q\times l_2\\ N={l_{CD}\times F}_N\end{array}\right.$$

(6)

where N is the equivalent normal force at the penetration sides of the pick-up pins, N; F_Q is the pushing force generated by the industrial pneumatic fixture, N; F_N is the normal force generated by the grasping operation at the penetration sides of the pick-up pins, N; l₁ and l₂ are the equivalent lengths from the fulcrum to the contact point, respectively, mm; and L_CD is the effective penetration depth of the pick-up pins, mm.

For the grasping operation of the pick-up pins in the non-deformable state, the value of N was approximately equal to the resultant force of the normal force at the penetration sides of the pick-up pins. The pushing force was generated by the air pressure on the horizontal piston of the industrial pneumatic fixture. Therefore, the value of F_Q was calculated as follows:

$$F_Q=P_0\times {\displaystyle \frac{\pi }{4}}\times d^2$$

(7)

where F_Q is the pushing force generated by the industrial pneumatic fixture, N; P₀ is the working air pressure of the industrial pneumatic fixture, Pa; and d is the cylinder diameter of the industrial pneumatic fixture, mm.

By combining Equations (5)–(7), the cylinder diameter of the industrial pneumatic fixture could be calculated as follows:

$$d=\sqrt{{\displaystyle \frac{2\times l_1\times F_P}{\pi \times l_2\times P_0\times \left(\sin \beta +\mu \times \cos \beta \right)}}}$$

(8)

where d is the cylinder diameter of the industrial pneumatic fixture, mm; F_P is the pulling force generated by the grasping operation, N; P₀ is the working air pressure of the industrial pneumatic fixture, Pa; β is the grasping angle of the pick-up pins, °; μ is the friction coefficient between the root lumps and the pick-up pins; and l₁ and l₂ are the equivalent lengths from the fulcrum to the contact point, respectively, mm.

The adhesion forces were represented by complicated growth adsorption processes, which were related to the combined effects of the root soil consolidation ability and water film attraction [19]. As an approximate calculation, mechanical force tests of pulling seedlings were conducted to estimate the approximate equivalent adhesion forces. In the practical application, the resultant force between the seedling gravity and the equivalent adhesion forces was 0.97 N to 2.93 N for typical seedlings. In the design, the pick-up pins were inserted into the root lump with a wide opening. The grasping angle of the pick-up pins was close to the conical degree of the tray cell. Based on the comprehensive analysis of the mechanism spatial scale, the equivalent length ratio of the leverage for l₁ and l₂ was taken as the intermediate value of 9:1. The working air pressure of the industrial pneumatic fixture was defined as the common value of 0.3 MPa. Therefore, the cylinder diameter of the industrial pneumatic fixture was calculated to be at least 9.25 mm according to the maximum seedling extraction conditions. This provided a basis for selecting the appropriate industrial pneumatic fixture. During the usage process, the grasping forces could be further adjusted via the working pressures for different seedling extraction requirements. In addition, the grasping force could also be moderately adjusted through controlling the limit distance of the countertop bolts.

2.2. Prototype Optimization and Performance Tests

In actual production scenarios, the unevenness of seedling growth, such as seedling heights, biomass, and plant materials, might lead to differences in seedling quality. As the grasping components are in direct contact with the seedlings, it was necessary for the flexible pick-up gripper to adapt to the internal deformation characteristics of the seedlings’ root lumps.

2.2.1. Prototype Construction of Testing Device

According to the overall design, a physical prototype of the large-opening flexible seedling pick-up gripper was constructed to examine its working efficiency (Figure 4a). The mini-A GR04.100-type industrial pneumatic fixture from Shenzhen Nast Machinery Equipment Co., Ltd. was used to perform the opening and closing actions of the pick-up gripper. The average grasping time was up to 0.8 s with an estimated service life of one million cycles. The cylinder finger was made of the MA-16×50-S-CA double-acting cylinder from Taiwan Airtac International Group. The operating speed ranged from 30 m/s to 800 m/s and the working air pressure ranged from 0.1 MPa to 1.0 MPa. The connecting parts between the cylinder fingers and the fork-type pins were fabricated via sheet metal processing techniques.

A robotic transplanting manipulator was designed in a previous study using the Cartesian coordinate system (labeled X-axis linear actuator and Y-axis linear actuator) to specify the working position of the gripper with a transplanting rate of 1000~1200 plants/h [19]. Here, the new large-opening flexible seedling pick-up gripper was assembled onto the transplanting manipulator for testing. The double-row chain conveyors were synchronized to move the source tray and the destination pots into the working space of the pick-up gripper. The SY-7120-5g-02-type solenoid valve from SMC Corporation (Tokyo, Japan) was used to control cylinder actions with flow rate of 4.5 dm³/(s bar) and operating speed ranging from 30 m/s to 800 m/s. Also, a GR300-10-type pressure-regulating valve from Airtac International Group was used to control the operating pressure for different needs. Its operating air pressure ranges from 0.05 to 0.9 MPa. According to the grasping requirements, the instant-response pneumatic circuits were designed to sense the cylinders at each station. Specifically, when the electromagnetic coil was powered, the inlet and outlet of the solenoid control valves were quickly switched in the control of the cylinder to perform the corresponding actions. Finally, a control system with an integrated power source was developed to enable point-to-point motion of the gripper between the source tray and the destination pots.

2.2.2. Optimization Tests of Grasping Parameters

Once the plug seedlings were successfully extracted from their tray cells, they would have a high probability of being properly transplanted into the soil or the growing pots. As shown in Figure 4b, several mechanical factors were found to influence the grasping effectiveness of the gripper. The front-end diameter of the pick-up pins should be moderately small to reduce penetration damage [6]. Based on the main root sizes of typical vegetable seedlings [20], the pick-up pins were made of 304 stainless steel wires with diameters ranging from 1.0 mm to 1.6 mm. In order to ensure that the pick-up pins grasped the maximum amount of root mass, the penetration depth in the top straight line needed be extended as much as possible. Therefore, different penetration depths were examined in the tests. As the seedlings may force the roots to coil around the perimeter of the tray cells, the pick-up gripper must hold the seedlings firmly for extraction. The grasping force for squeezing root lumps and the extraction speed for lifting plug seedlings vertically out of the cells were investigated in the process of transplanting [18]. In view of the above analysis, orthogonal optimum tests were conducted with the multi-factor levels listed in

Table 1. Factors and levels of the optimal grasping parameters.

Factor	Symbol	Level
		1	2	3	4
Pin diameter/mm	A	1.0	1.2	1.4	1.6
Penetration depth/mm	B	25	30	35	40
Grasping force/MPa	C	0.2	0.3	0.4	0.5
Extraction speed/(mm/s)	D	600	900	1200	1500

.

The multi-factor tests of the optimal grasping parameters were designed based on the L₁₆ (4⁵)-type orthogonal experimental table. The last column in the table was regarded as the error column. There were 16 groups of combined tests. In each test, 16 seedlings were transplanted continuously, and the overall process was repeated 5 times. Taking the 35-day-old tomato seedlings in the 128-cell trays as testing objects, they were suitable for transplanting in a farm in Zhenjiang, Jiangsu province, China. The seedlings used were watered a day before transplanting. The moisture contents of the seedling’ root lumps were kept at the medium level of 55%~60% during the optimal tests [21]. As shown in Equation (9), the integrity rate of the shaped substrate body was used as the response value to evaluate the grasping effectiveness of the gripper under different sets of factors and levels. Finally, the corresponding data were recorded, and the statistical analysis was conducted on SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). The least significant difference (LSD) method was used for multiple comparisons of these data, and the significance of value (significant when p < 0.05) was indicated.

$$\mathit{Y}={\displaystyle \frac{m_{\mathit{t}}{-m}_{\mathit{f}}}{m_{\mathit{t}}}}$$

(9)

where Y is the integrity rate of the shaped substrate body, %; m_t is the total mass of the shaped substrate body, g; and m_f is the broken mass of the shaped substrate body during the grasping process, g.

2.2.3. Performance Tests Under Actual Conditions

As shown in Figure 5, three typical vegetable seedlings (35-day-old tomato seedlings, 40-day-old pepper seedlings, and 32-day-old Chinese cabbage seedlings) were used to verify the optimal grasping parameters. The seedlings used were mature and deemed ready for transplanting by the local growers. In order to accurately evaluate the practicality and adaptability of the large-opening flexible seedling pick-up gripper, these seedlings were transplanted from 128-cell trays to 50-cell trays. In order to account for variability in the seedlings, every species was tested 5 times with a four-week interval. The overall transplanting process was observed and recorded on-site using a CCD camera. The corresponding verifiable results were recorded, and data analyses were conducted.

3. Results and Discussion

3.1. Optimization Tests

The optimal grasping parameters are presented in

Table 2. Optimal results of the grasping parameters.

No.	Factor (Levels)					Y/%
	A	B	C	D	E (Error)
1	1 (1.0 mm)	1 (25 mm)	1 (0.2 MPa)	1 (600 mm/s)	1	83.78
2	1	2 (30 mm)	2 (0.3 MPa)	2 (900 mm/s)	2	85.63
3	1	3 (35 mm)	3 (0.4 MPa)	3 (1200 mm/s)	3	87.09
4	1	4 (40 mm)	4 (0.5 MPa)	4 (1500 mm/s)	4	83.65
5	2 (1.2 mm)	1	2	3	4	88.44
6	2	2	1	4	3	86.91
7	2	3	4	1	2	93.57
8	2	4	3	2	1	95.64
9	3 (1.4 mm)	1	3	4	2	85.87
10	3	2	4	3	1	89.35
11	3	3	1	2	4	87.43
12	3	4	2	1	3	89.41
13	4 (1.6 mm)	1	4	2	3	87.05
14	4	2	3	1	4	88.95
15	4	3	2	4	1	84.26
16	4	4	1	3	2	88.13
K1	85.04	86.29	86.56	88.93	88.26
K2	91.14	87.71	86.94	88.94	88.30
K3	88.02	88.09	89.39	88.25	87.62
K4	87.10	89.21	88.41	85.17	87.12
R	6.10	2.92	2.83	3.77	1.18
S	A2	B4	C3	D2

. Based on the range analysis (R values), the most influential factors on the grasping integrity rates were the pin diameter and the extraction speed, followed by the penetration depth and grasping force. The optimal combination of parameters was identified as A₂B₄C₃D₂. The optimum grasping effectiveness of the new gripper could be achieved when the end diameter of the pick-up pin was 1.2 mm, the penetration depth in the top straight line of the pick-up pin was 40 mm, the grasping force for squeezing root lumps was 0.4 MPa, and the extraction speed for lifting plug seedlings vertically was 900 mm/s.

As shown in

Table 3. Analysis of variance (ANOVA) for the orthogonal tests.

Source	Sum	DOF	Mean Square	F Value	p Value	Significant Test
A: Pin diameter	77.30	3	25.77	20.19	0.02	*
B: Penetration depth	17.46	3	5.82	4.56	0.12	ns
C: Grasping force	20.66	3	6.89	5.39	0.10	ns
D: Extraction speed	38.69	3	12.90	10.10	0.04	*
Deviation	3.83	3	1.28
Sum	157.93

Note: ns, no significant effect; * significant at 0.01 < p < 0.05.

, the statistical analysis of variance (ANOVA) for the orthogonal tests further showed that the end diameter of the pick-up pin and the extraction speed for lifting plug seedlings in the vertical direction had significant effects (0.01 < p < 0.05) on the grasping effectiveness of the new gripper. The other factors of the penetration depth and grasping force had no significant effects (p > 0.05). The ANOVA results were consistent with the findings from the range analysis, reinforcing the significance of the pin diameter and extraction speed. The multi-factor tests could be used to evaluate the grasping effectiveness of the gripper under different sets of factors and levels.

After conducting the multi-factor tests, it was found that the end diameter of the pick-up pin and the extraction speed were crucial to the success in separating plug seedlings from their growth cells. At the pick-up pin end diameters ranging from 1.0 mm to 1.6 mm, the integrity rates of the shaped substrate body in transplanting were 85.04%, 91.14%, 88.02%, and 87.10%, respectively. With the increase in the pin diameter, the grasping effectiveness first increased and then decreased. For the material properties of 304 stainless-steel wire, the pick-up pins with a medium diameter of 1.2 mm could maintain a reliable grasping force to pick up seedlings [6,19]. The material properties of the pick-up pins were different. Given the variation in material properties, the future work could explore alternative pin materials (e.g., composites or coated metals) to suit different seedling types. The ease of manufacturing pins should also be taken into consideration. In the multi-pin grasping operation that involved penetrating the root lump, the pick-up pins always caused some damage. Based on the working principle of painless medical needles [22,23], the tip of the pick-up pins could be sharpened to obtain variable diameter pins for soil insertion. In this way, a low-loss state was obtained for penetrating and grasping the seedling’s root lump. After the root lump of a seedling was grasped by the pick-up gripper, the robotic transplanting manipulator attempted to lift the seedling vertically. This lifting motion might be associated with some increased forces causing the pick-up pins to break the shaped substrate body, and limiting their ability to maintain appropriate grasping of the seedling [18]. As the extraction speed for lifting plug seedlings reached the maximum value of 1500 mm/s, the grasping effectiveness of the new gripper deteriorated significantly and reached the lowest level. Since the plug seedling was a root–substrate complex, a reaction process involving separation from the cell wall was required to achieve success in picking up the seedlings [6,19]. Further studies are required to determine the reasonable extraction rate with a minimum transplanting time to achieve the most efficient separating motion, also considering the seedlings’ complicated growth adhesion properties [24].

During the grasping operation, the fork-type pins were pushed to penetrate the root lump obliquely along both sides of the tray cell. The average integrity rates of seedlings in automatic transplantation were 86.29%, 87.71%, 88.09%, and 89.21% for penetration depths of 25 mm, 30 mm, 35 mm, and 40 mm, respectively. The grasping effectiveness of the gripper constantly increased with the increase in the penetration depth in the top straight line. In the wide-open grasping state, the fork-type pins were moved deeply into the root lumps to facilitate the removal of seedlings. This was in agreement with the finding that the best performance was obtained when the pick-up pins penetrated the root lump as deeply as possible [6]. When the penetration depth of the pick-up pins reached more than three-quarters of the cell depth, the integrity rate of seedling retrieval under different transplanting operations could remain at a relatively high level. The working pressure of the industrial pneumatic fixture for squeezing root lumps was kept at 0.4 MPa to achieve the optimum grasping effectiveness of the new gripper at the highest integrity rate of 89.39%. The increased pressure also seemed to facilitate seedling grasping at medium moisture levels. On the whole, it was important for the pick-up gripper to grasp the maximum number of root lumps and hold them firmly during transplanting.

3.2. Performance Tests

The transplanting success rate reflects the gripper’s overall efficiency in extracting, transferring, discharging, and planting seedlings. Statistical analysis was conducted on five rounds of transplanting data for every species. The corresponding results of the performance tests are shown in

Table 4. The corresponding results of the performance tests.

Seedling	No. of Seedlings Fed	No. of Extraction Failures	No. of Substrate Breakages	No. of Seedlings Damaged	No. of Planting Failures	Success Rate in Transplanting/%	Average Success Rate/%
Tomato	128	2	2	0	2	95.31	94.22 ± 1.62
	128	3	3	2	2	92.19
	128	1	2	1	1	96.09
	128	3	4	1	1	92.97
	128	1	2	2	2	94.53
Pepper	128	2	1	2	1	95.31	95.16 ± 1.39
	128	2	2	3	2	92.97
	128	1	2	2	1	95.31
	128	2	1	1	0	96.88
	128	1	1	3	1	95.31
Cabbage	128	2	2	0	1	96.09	96.41 ± 1.18
	128	1	3	1	1	95.31
	128	1	3	0	1	96.09
	128	0	2	0	0	98.44
	128	1	2	1	1	96.09

. The average success rates in transplanting were 94.22 ± 1.62%, 95.16 ± 1.39%, and 96.41 ± 1.18% for tomato, pepper, and cabbage seedlings, respectively. Once these plug seedlings were successfully extracted out of their growth cells, they had the highest probability of being properly transferred and discharged for planting into the growing flats (Figure 6a). The subsequent cultivation indicated that all the seedlings survived.

When the tray cell was loosely filled with the growth substrate, the seedlings’ root system was the key element for bearing grasping forces [18]. Most extraction failures often occurred with young seedlings because their root lumps were not well developed. There were 10 extraction failures for tomato seedlings, 8 for pepper seedlings, and 5 for cabbage seedlings. In addition, the seedling’s root system was so poorly tangled that the vibration shocks of transferring and discharging forced the root substrate to break during the continuous action process (Figure 6a–c). For efficient transplanting, a dense root structure would usually provide the highest probability for extracting, transferring, and discharging seedlings. Therefore, seedling cultural production should consider mechanical handling [25]. It might be possible to improve the root systems by using more fibrous growth media or a soil stabilizer. Biochar is a good medium for soil improvement, which can promote seedling growth. On the basis of seedling cultivation, the effect of the biochar added in peats on the growth of seedlings was studied by testing the growth status of the seedlings, parameters of root systems, and compressive strength of the substrates [20]. Therefore, by enhancing root growth until the time of transplanting, the strength and adaptability of the shaped root lump can be maximized to protect against damage and achieve reliable grasping [26]. In the design, the opening cavity between two pairs of pick-up pins was widened to accommodate the seedling plant. For the three typical vegetable seedlings studied, most of the seedlings were undamaged during the transplanting process. As shown in Figure 6b,c, the older seedlings were plump, with an abundance of leaves causing their leaves to become tangled with the pick-up pins. Moreover, the tilted seedlings and extruding leaves from neighboring cells were prone to being pinched and damaged. This resulted in damage to 6 tomato seedling plants and 11 pepper seedling plants. Meanwhile, the stems of the cabbage seedlings were short, and only two plants were damaged (Figure 6d). For common vegetables, the leaves of seedlings were torn off by the pick-up pins, which may affect their rejuvenation and revival. However, minor damage to cotyledons had minimal impact on subsequent seedling growth [6]. The success rates in transplanting seedlings could be further increased by addressing these challenges. Further, the working path of the gripper could be optimized to shift between large plants to overcome damage to the seedlings during transplanting. There were planting failures for each type of seedling in this study. In many cases, the seedlings’ crown widths were so broad that their leaves were also tangled with the gripper. It was not easy to discharge such seedlings when they were planted into the destination pots. These results further confirmed that short and upright seedlings were well suited to mechanized transplanting [27].

Although the prototype seemed to be studied in the laboratory, it was intended at the beginning of this study to concentrate our efforts on making the function of the new gripper more accurate. Taking three typical vegetable seedlings as the transplanting objects, the performance tests confirmed that the large-opening flexible seedling pick-up gripper was a workable design that caused minimal damage to the seedlings’ roots and the plants. The basic actions of the gripper can be defined as efficiently grasping, firmly transferring, and freely discharging seedlings. On the whole, the new gripper could complement manipulator motions with minimum restriction. Compared with other designs, the large-opening seedling pick-up gripper is structurally simple, functionally accurate, and flexibly feasible for the seedling plants [14,15,16]. The flexible grasping mechanism was also designed to respond to several variables such as standard trays and seedling characteristics [6,24]. In order to achieve high-speed mechanized transplanting, the seedling quality must be uniform and sturdy. Some other essential factors are that the plug seedlings should be easily extracted from the tray cells, and their roots should be evenly distributed in the loose substrate particles. Therefore, vegetable seedlings should be produced via specialized agronomy practices for use with the gripper, considering factors such as the shortness of seedling plants and root formation [20,27]. Moreover, the working path of the gripper could be optimized to overcome damage to the seedling branches and leaves in transplanting. Flexible automation, which introduces a procedure to accommodate several changing objects, might become an important alternative [28]. Further studies of the control strategy of the robotic transplanting manipulator should also be conducted to eradicate current bottlenecks in enhancing the adoption level of mechanization. This new pick-up gripper did not compensate for blank cells and unhealthy seedlings in the trays. However, if computer vision systems could be used for a precise identification of objects, the automatic transplanting performance would be improved by abandoning the blank cells and unhealthy seedlings in the trays [29]. The multi-gripper system could significantly improve the operational efficiency. Furthermore, a set of cooperative transplanting systems with multiple grippers could be constructed to achieve a row of automatic transplanting seedlings working simultaneously. Large-scale production applications of the new gripper should be performed under various environmental conditions. Therefore, the durability and reliability of grasping seedlings should be evaluated considering seedling substrate materials, seedling growth stages, and working vibration conditions of the transplanter.

4. Conclusions

On the basis of the standard trays and seedling characteristics, a large-opening flexible seedling pick-up gripper was designed and evaluated in a laboratory. The pick-up gripper with a large-opening cavity for accommodating fragile seedling plants was a pincette-type mechanism utilizing multiple fine pins for flexible grasping of root lumps. The physical prototype of the pick-up gripper was constructed, and multi-factor orthogonal tests were conducted to evaluate the effectiveness of grasping. It was found that the end diameter of the pick-up pin and the vertical extraction speed significantly influenced the gripper’s performance (0.01 < p < 0.05). The grasping effectiveness was first increased and then decreased with the increase in the pin diameter. Excessive extraction force at higher speeds increased the risk of damaging the shaped substrate body. For efficient transplanting, the pick-up gripper must grasp the maximum number of root lumps and hold them firmly. The optimum grasping effectiveness of the new gripper was achieved when the end diameter of the pick-up pin was 1.2 mm, the penetration depth in the top straight line of the pick-up pin was 40 mm, the grasping force for squeezing root lumps was 0.4 MPa, and the extraction speed for lifting plug seedlings in a vertical direction was 900 mm/s. The performance tests further revealed that the average success rate in transplanting was up to 95% for typical vegetable seedlings. The large-opening flexible seedling pick-up gripper could successfully perform extraction, transfer, discharge, and planting of seedlings. In future, vegetable seedling qualities should be combined with automatic transplanting operations.