Design and Experimental Study of an Automatic Seedling Picking and Feeding Device for a Strawberry Bare-Root Seedling Transplanter

Abstract

The automated transplanting of bare-root strawberry seedlings faces challenges, including a high reliance on manual feeding and a lack of adaptable equipment. Moreover, existing picking mechanisms are typically limited to plug seedlings rather than multi-stem crops. To overcome these limitations, this study proposes an automatic seedling picking and feeding system synergizing an intermittent conveying device and a manipulator. First, the physical and mechanical properties of bare-root strawberry seedlings were measured. Through radial stem compression tests, a non-destructive clamping safety threshold bounded by a bio-yield point of 28 N was determined. Secondly, an intermittent conveying device based on variable-span V-shaped supports was designed. Furthermore, based on the clamping safety threshold and spatial kinematic modeling, a seedling picking end-effector was developed. This end-effector adopts a “gather first, clamp later” operation strategy to guarantee grasping accuracy and ensure reliability by preventing seedling detachment during transportation. Finally, a system test bench was built to conduct a three-factor, three-level orthogonal experiment to investigate the effects of seedling picking frequency, gripping position, and clamping gap on operation quality. At a seedling pick-up frequency of 20 plants/min, gripping position of 25 mm, and gripping gap of 5 mm, the system reached 98% success rate in picking and feeding, validating its stability and providing core equipment and theoretical support for fully automatic strawberry transplanters.

1. Introduction

Given its high economic value, strawberry production must be mechanized to improve overall agricultural profitability [1,2]. In large-scale strawberry cultivation, bare-root strawberry seedlings are widely adopted due to their advantages such as low breeding costs, easy storage and transportation, and strong environmental adaptability [3,4]. However, as the shortage of agricultural labor becomes increasingly severe, automated transplanting technology has become a key bottleneck restricting the development of the strawberry industry [5,6,7]. Currently, the transplanting operations for bare-root strawberry seedlings are still predominantly semi-automatic or manual, and no mature, fully automatic bare-root seedling transplanter has yet emerged [8]. To achieve the leap from semi-automatic to fully automatic transplanting of bare-root strawberry seedlings, three core technical bottlenecks must be addressed: First, the morphological characteristics of bare-root seedlings are complex, which makes visual recognition and positioning difficult; second, the seedling feeding and supply process struggles to adapt to the morphological differences in bare-root seedlings, heavily relying on manual posture-setting and feeding, which results in poor operational continuity; third, in the core execution process of seedling picking and feeding, existing rigid mechanical structures generate significant motion impact, which easily causes mechanical damage when clamping fragile stems, leading to high seedling damage and drop rates. These issues severely restrict the development of strawberry bare-root seedling transplanters toward high precision and non-destructive operations [9,10,11,12,13].

In recent years, domestic and international scholars have conducted extensive research on the mechanization of transplanting for strawberries and other crops. The Rain Flo 1600 transplanter (manufactured by Rain‑Flo Irrigation, East Earl, PA, USA) and the transplanters used by Driscoll’s (based on models from Checchi e Magli, Italy) are suitable for single-row on-ridge operations with large plant spacing for bare-root strawberry seedlings and potted seedlings; however, both lack automatic seedling picking devices and are classified as semi-automatic equipment [14]. Liu et al. developed a high-ridge, double-row strawberry plug seedling transplanter [15], Liu et al. designed a hole-punching strawberry transplanter [16], and Kang et al. specifically designed an automatic seedling picking device for strawberry transplanters [17]. These devices perform excellently in automatic seedling supply and conveying mechanisms, but their automatic picking systems are all designed for strawberry plug seedlings supported by substrate blocks, and cannot be directly applied to bare-root seedlings that lack substrate wrapping and have exposed roots.

In the field of transplanting other substrate-free bare-root crops, scholars have made numerous beneficial explorations. For instance, regarding sweet potato bare-root seedlings, some studies designed an automatic seedling supply and planting control system based on pre-processed seedling belts. This technology pre-fixes individual sweet potato seedlings onto flexible seedling belts equipped with Velcro, and, combined with motor-driven seedling feeding rollers and a dual flexible disk mechanism, successfully achieves the continuous automatic conveying and clamped planting of sweet potato seedlings [18]. Meanwhile, for the mechanized transplanting of Welsh onion bare-root seedlings, researchers designed a clamping and planting mechanism based on the combination of flexible disks and elastic pressure rollers [19]. However, neither the belt-based sweet potato seedling supply mechanism nor the corrugated disk clamping mechanism for Welsh onions is suitable for bare-root strawberry seedlings. If bare-root strawberry seedlings are forcibly fed into such pre-processed seedling belts or corrugated squeezing disks, it is not only difficult to untangle the root systems, but the strong flat belt clamping or radial disk squeezing can also easily break the clustered and brittle petioles, causing irreversible mechanical damage to the shortened stems.

Regarding the seedling picking and feeding end-effector, stem-clamping seedling picking has become a current research hotspot because it can effectively avoid complex root systems [20]. Zhang Ni designed a combined pot-pushing and stem-clamping picking device [21], Wang Xiu designed a stem-clamping picking device for vegetable transplanting [22], Liu Zhicheng developed an automatic transplanter for greenhouse tomatoes [23], and Lin Zhenhua designed a novel stem-clamping and extracting method based on a non-circular gear system [24]. These stem-clamping picking devices typically cooperate with photoelectric sensors or mechanical limits for precise positioning, but they mostly adopt rigid, direct release during the feeding and releasing stage, lacking a flexible transition. Furthermore, these existing devices are primarily designed for morphologically regular, single-stemmed vegetable plug seedlings such as peppers and tomatoes. Unlike conventional vegetable seedlings, bare-root strawberry seedlings feature multiple clustered stems and leaves, extensive root systems, and a lack of substrate support. If existing pneumatic seedling picking end-effectors and positioning/feeding methods are directly adopted, not only is it difficult to achieve precise posture-setting, but seedling injury will also occur due to uncontrollable clamping force and motion impact [25], and seedlings are highly prone to falling off during transportation. Therefore, it is necessary to develop a dedicated non-destructive clamping and precise, stable feeding mechanism tailored to the physical characteristics of bare-root strawberry seedlings.

Therefore, this study aims to develop a precise and stable automated seedling picking and feeding system to address the heavy reliance on manual labor and the poor conveying continuity in bare-root strawberry transplanting. Specifically, this research focuses on the development and bench test evaluation of a novel intermittent conveying device based on variable-span V-shaped supports, alongside a non-destructive robotic manipulator for seedling picking and feeding. Based on the physical characteristics of the clustered strawberry stems, this study centers on two core hypotheses: First, we hypothesize that combining variable-span V-shaped supports with sensor-based intermittent control can accommodate the morphological complexity of bare-root seedlings, thereby achieving precise posture adjustment and a continuous seedling supply. Second, we hypothesize that, compared to traditional rigid clamping mechanisms, restricting the manipulator’s gripping force to within the experimentally determined biological yield boundary of the stems will significantly reduce seedling damage and dropping rates.

2. Materials and Methods

2.1. Determination of Physical Properties of Strawberry Bare-Root Seedlings

The material properties of bare-root strawberry seedlings are the foundation for the parameter analysis of the key components of the automatic seedling picking and feeding device to achieve automatic picking under mechanical constraints. In this study, “Hongyan” bare-root strawberry seedlings—grown for 60 to 70 days after stolon propagation and possessing 3 to 5 expanded leaves—were selected as the research object to conduct external dimension measurement experiments. The ‘Hongyan’ variety is one of the most widely cultivated strawberry cultivars in China, particularly in Jiangsu Province and its surrounding regions, where this study was conducted. It is characterized by moderate plant vigor, well-developed root systems, and a typical multi-stem clustered morphology, which can fully represent the morphological challenges encountered in automated transplanting. Therefore, this variety serves as an ideal benchmark material for validating and evaluating the proposed design. A custom-built inclined plane friction testing device was used to measure the coefficient of static friction between the seedling stems and typical contact materials.

The position of the center of gravity is the core mechanical basis determining the magnitude of eccentric loading on the mechanical gripper during high-speed transportation. Given that bare-root strawberry seedlings exhibit an overall slender and irregular morphology, the offset of their longitudinal center of gravity plays a decisive role in grasping stability. Therefore, based on the principle of rigid body lever equilibrium, a horizontal suspension balance method was employed in this study to determine the axial position of the center of gravity.

2.1.1. Basic Characteristic Parameters

A total of 100 60-day-old bare-root strawberry seedlings with a moisture content of 71% were randomly selected for morphological measurement. The measurement positions for each parameter are illustrated in Figure 1.

In this study, the parameters are defined as follows: plant height (l_a) is the vertical distance from the root crown to the highest leaf tip; root length (l_b) is the length of the longest root measured from the root crown; leaf height (l_c) is the height of the leaf canopy above the root crown; stem diameter (l_d) is the diameter of the stem at 30 mm above the root crown; leaf spread (l_e) is the maximum horizontal width of the leaf canopy; and the position of the center of gravity (l_g) is the axial distance from the root crown to the center of gravity, determined using the horizontal suspension balance method.

2.1.2. Radial Compression Measurement of Bare-Root Strawberry Seedlings

The bow back of bare-root strawberry seedlings is an important agronomic characteristic [26]. To avoid damaging the bow back during clamping, and based on the center of gravity measurements, the grasping point was determined to be located 20 to 30 mm above the root. To determine the appropriate clamping force for the seedling picking end-effector, radial compression tests were conducted on the stems of the bare-root strawberry seedlings. Seedlings with three stems were selected, and the tests were performed on the stems at a position 20 mm above the root to measure the compression–displacement curve of each stem, as shown in Figure 2. Radial compression tests were performed on a TA.XTplus Texture Analyser (Stable Micro Systems, Godalming, Surrey, UK), which features a force measurement accuracy of better than 0.5% of the reading, a force resolution of 0.1 g, a displacement resolution of 0.001 mm, and a speed accuracy of better than 0.1%. The force sensor was calibrated with standard weights before each test to guarantee measurement reliability. A P/36R probe was selected as the loading head. The loading speed was set to 1 mm·s⁻¹, and the target strain was set to 80%. The collection of experimental data commenced when the probe detected a force of 0.1 N. By recording the force-displacement curves during the compression process, the variation laws of the forces acting on the stems of bare-root strawberry seedlings under pressure were analyzed to characterize their compressive mechanical properties.

2.2. Overall Scheme and Working Principle of the System

2.2.1. Overall System Structure Design

The overall structure of the fully automatic hole-punching and directional transplanter for bare-root strawberry seedlings is shown in Figure 3a. It is mainly composed of a machine frame, a hole-punching device, a directional planting device, an automatic seedling picking and feeding device, a soil-covering device, a seedling tray, an electrical cabinet, a switch box, and a seat. As the core operational unit, the structure of the automatic seedling picking and feeding device for bare-root strawberry seedlings is shown in Figure 3b. It primarily consists of a supporting frame, an intermittent conveying device, an automatic seedling picking and feeding manipulator, and a control cabinet. Specifically, the intermittent conveying device is fixed to the supporting frame via a support plate and is responsible for the intermittent stepping supply of the seedlings; the automatic seedling picking and feeding manipulator is mounted above the frame via a connecting plate and is responsible for executing compliant clamping and spatial transportation. These modules operate synergistically to jointly form the core operational system that realizes the conveying, picking, and feeding of the bare-root seedlings.

2.2.2. Sequence Planning of Automatic Operations

During operation, bare-root strawberry seedlings are manually placed into the adjustable support device of the conveying unit. After pressing the start button, the machine can execute the continuous automatic operations of seedling conveying, picking, and feeding. The horizontal transport stroke between the picking point and the dropping point is 350 mm, as determined by the overall layout of the transplanter. Detailed parameter specifications are provided in Section 3.2. The specific steps are illustrated in Figure 4.

Manually adjust the spacing of the supports according to seedling variety and average root length, matching the span to seedling length for stability.

Step 1: The stepper motor drives the conveyor belt intermittently for pre-feeding, sequentially delivering 3 seedlings to the picking station to initiate the picking-dropping cycle.

Step 2: When the seedling reaches the preset picking point, an induction plate triggers the proximity switch, the PLC stops the conveyor belt, the vertical module lowers the end-effector to the preset gripping position on the support plane, and the motor drives the linkage mechanism to close the claws. The claws first gather the stems, then secure the clamping.

Step 3: The vertical module rises to detach the seedling from the support. The horizontal module transports the seedling laterally right above the planting device. The vertical module lowers to the planting device plane. After the planting device clamps the bare root, the seedling picking end-effector open in reverse, and the vertical module ascends.

Step 4: The horizontal module transports the end-effector back above the picking point to wait for the next seedling.

Cycle Steps 1–4 until operations are complete.

2.3. Key Component Design and Analysis

2.3.1. Design of Adjustable-Span Intermittent Conveying Device

The intermittent conveying device for bare-root strawberry seedlings designed in this study, as shown in Figure 5a, primarily consists of a conveyor belt with its base and side frames, a stepper motor, double-ended adjustable support frames, a proximity switch, and sensing plates. Serving as the core power source of the conveying system, the output shaft of the stepper motor is coupled with the drive shaft to precisely control the intermittent motion stroke and start-stop logic of the conveyor belt. On the surface of the conveyor belt, multiple double-ended adjustable support frames are fixed at equal intervals along the direction of motion to provide reliable support and posture constraint for the bare-root strawberry seedlings. A proximity switch is fixedly mounted on the side frame of the conveyor belt at the end opposite to the driving end, and the outer side of each support frame is correspondingly equipped with a sensing plate that moves synchronously with it [27].

During the manual feeding stage, the bare-root strawberry seedlings are placed transversely on the double-ended adjustable support frames. The roots are uniformly oriented toward the outer support end of the conveyor belt, while the crowns rest on the inner support end, and the seedlings are conveyed forward in a stepping motion. This two-point support layout leaves the central stem area of the bare-root seedling suspended, providing an ample operational window for the end effector. This ensures that the gripper can accurately descend and grasp the seedling at the predetermined gripping point.

Due to the significant variations in length among individual bare-root strawberry seedlings, using a fixed-span support during the conveying process can easily cause shorter seedlings to slip off or the center of gravity of longer seedlings to shift, thereby reducing the success rate of subsequent gripping operations. To address this issue and achieve reliable support and stable posture constraint for bare-root seedlings of varying lengths, a variable-span support mechanism was designed in this study. This mechanism primarily consists of two components: a fixed bracket and an adjustable support slider. The fixed bracket is securely attached directly to the conveyor belt, serving as the reference support end. The adjustable support slider is mounted via a guide rail mechanism, allowing it to slide laterally across the conveyor belt and be locked into the desired position.

The structural parameters of the conveying device are annotated in Figure 5b. To ensure the stability of the seedling’s center of gravity while keeping the central stem area completely suspended for interference-free gripper operation, the structural parameters of the conveying device must satisfy the following geometric constraints:

$$\left\{\begin{array}{l}D=L_i-\left(\Delta L_r+\Delta L_c\right)\\ S\ge L_{max}-L_{min}\\ P\ge W_{max}+\delta \\ \tan \left({\displaystyle \frac{\alpha }{2}}\right)<{\displaystyle \frac{1}{\mu }}\end{array}\right.$$

(1)

In the formula, $D$—Effective span between the support slider and the fixed bracket, mm; $L_i$—Length of the bare-root strawberry seedlings in the current operating batch, mm; $\Delta L_r$—Reserved overhang allowance at the outer side of the root, mm; $\Delta L_c$—Overlap allowance of the crown on the outer side of the support slider, mm. $P$—Center installation pitch of adjacent support frames on the conveyor belt, mm; $W_{max}$—Maximum limit leaf spread diameter of the bare-root strawberry seedlings, mm; $\delta$—Redundant safety clearance to prevent physical interference, mm.

2.3.2. Design of Picking End-Effector Based on Seedling Characteristics

The pick-and-place manipulator is the core execution component of the automatic picking and feeding system for bare-root strawberry seedlings, responsible for completing the entire operational process of precisely grasping the bare-root seedlings from the conveying device, transporting them smoothly, and accurately positioning and releasing them into the planting device [20].

As shown in Figure 6, the pick-and-place manipulator mainly consists of three parts: a horizontal motion module, a vertical lifting module, and a seedling-picking end-effector installed at the end of the sliding table of the horizontal module. To meet the precise positioning requirements under high-speed continuous operation, the spatial movement of the manipulator adopts an orthogonal configuration similar to a Cartesian coordinate system, where the horizontal and vertical axes are perpendicular to each other and their movements are mutually independent; the kinematic solution has no coupling, and the trajectory has no singular points, which can ensure high spatial positioning accuracy. According to the requirements of the operational rhythm and span, a ball screw linear guide sliding table mechanism is selected for horizontal movement. This mechanism features high transmission efficiency, high positioning accuracy, and strong structural rigidity. Its power source utilizes a servo motor, transmitting power through components such as couplings, thrust bearings, ball nuts, linear guides, support bearings, and lead screws, in order to satisfy the manipulator’s requirement for high-frequency, high-speed, and smooth long-distance reciprocating transportation between the picking point and the dropping point. The lifting motion stroke is relatively short, primarily responsible for controlling the precise height when the end-effector descends for seedling picking and dropping. Considering the accuracy of point-to-point movement and the need to accommodate morphological differences among different batches of seedlings, this module employs a stepper motor matched with a precision lead screw drive, ensuring the stability and stopping accuracy of the descending and ascending actions in the Z-direction. As the component in direct contact with the bare-root strawberry seedlings, the seedling-picking end-effector utilizes a stepper motor to drive its internal screw slider and symmetric linkage mechanism. Combined with the “gather-first, clamp-later” operational strategy, it achieves non-destructive wrapping and flexible clamping of the root-stem junction of the bare-root strawberry seedlings by precisely controlling the opening/closing angle and motion stroke of the flexible claws, effectively avoiding physical interference and mechanical damage. To ensure the coherence of the fully automatic continuous operation of the entire machine, the pick-and-place manipulator must strictly execute the following seven action sequences within a complete working cycle: “move to picking point $\to$ descend for picking $\to$ clamp seedling stem $\to$ ascend to detach $\to$ transport horizontally $\to$ descend for dropping $\to$ reset”. Based on the comprehensively selected motion forms, the overall working requirements of the machine, and the three-dimensional spatial layout requirements mentioned above, the core kinematic parameters and structural dimensions of each module of the pick-and-place manipulator are selected and configured.

To achieve flexible and non-destructive clamping, this study breaks through the limitations of traditional rigid metal grippers by adopting a flat, straight nylon plate structure for the end-effector claws. The relatively low elastic modulus of nylon endows the claws with excellent bending compliance. The schematic diagram of the seedling-picking end-effector mechanism is shown in Figure 7. Because the top of the mechanism is suspended by a fixed hinge support, when the stepper motor drives the central slider upward, the claws swing inward symmetrically in an arc around the pivot point. During the closing process, the claws do not remain parallel; instead, as the swing angle changes, their extension lines intersect in space, ultimately forming an inverted V-shaped converging envelope space in the clamped state.

To ensure the successful clamping of the bare-root seedling stems, the following conditions must be met:

(1). The opening degree at the end of the claws must be larger than the maximum measured stem diameter at the upper part of the new stem of the bare-root strawberry seedlings.

(2). The clamping force must be limited below the physical failure strength of the stem to prevent irreversible mechanical damage.

(3). The maximum static friction generated between the claws and the stem surface must be sufficient to resist the horizontal inertial forces generated during rapid acceleration or sudden stops of the manipulator. Additionally, the claws must provide sufficient anti-overturning frictional torque on the contact surface to counteract the overturning torque that causes the bare-root seedling to tilt to one side due to the eccentricity of its center of gravity during such rapid acceleration or deceleration.

(4). Because the claws form a V-shape when closed, an upward wedge-shaped thrust is also generated. Therefore, the design of the included angle must satisfy the frictional self-locking condition to prevent the seedling from being squeezed outward.

Based on the physical morphology of the stems of bare-root strawberry seedlings, the following gripper motion plan for seedling extraction is designed:

The gripper descends vertically via the Z-axis module in its fully open position. At this stage, the opening distance at the bottom of the dual flat jaws exceeds the natural spread of the three scattered stems. The gripper surrounds the scattered stems in a V-shaped configuration, ultimately stopping at the horizontal plane of the root-stem junction. The stepper motor then drives the slider upward, causing the gripper jaws to swing inward. Because the gripper closes in a converging V-shape, the flat inner walls of the jaws first make contact with the outermost loose stems. By leveraging the tangential guiding force of these flat surfaces, the three splayed stems are forced to slide along the inner walls toward the central axis, gathering into a single bundle. As the mechanism advances, the angle of the dynamic V-shaped space reaches its designated value. The bilateral jaws then fit securely against the outer walls of the bundled stems at an inclined angle, successfully completing the grasping process of the bare-root strawberry seedling.

In order to explore the motion trajectory of the gripping point, kinematic modeling is performed on the mechanism. The schematic diagram of the seedling picking gripper mechanism is shown in Figure 7. Since the mechanism is symmetrical, the left half is taken for independent modeling. The center point of the top fixed base frame is taken as the origin $O\left(\mathrm{0,0}\right)$. The X-axis is positive horizontally to the right, and the Y-axis is positive vertically downward. According to the definition of the coordinate system, the coordinates of the two hinge points on the left are as follows: The left top fixed hinge $P_0$: $\left(-w_f/\mathrm{2,0}\right)$ The left slider moving hinge $P_2$: $\left(-w_s/2,d\right)$. The left mechanism is essentially a variation in a slider-crank mechanism. Let the angle between the connecting rod $L_1$ and the positive direction of the X-axis be ${\theta }_1$, and the coordinates of the hinge point $P_1$ be $\left(x_1,y_1\right)$:

$$\left\{\begin{array}{l}{x}_1=-w_f/2-l_1{\cos \theta }_1\\ y_1=l_1{\sin \theta }_1\end{array}\right.$$

(2)

The hinge point $P_2$ is located on the slider, and its coordinates satisfy the geometric constraint that the distance between it and $P_1$ is $L_2$:

$${\left(x_2-x_1\right)}^2+{\left(y_2-y_1\right)}^2=l_2^2$$

(3)

Substituting $x_1$ and $y_1$ into the equation and expanding:

$${\left(-{\displaystyle \frac{w_s}{2}}-\left(-{\displaystyle \frac{w_f}{2}}-l_1{\cos \theta }_1\right)\right)}^2+{\left(d-l_1{\sin \theta }_1\right)}^2=l_2^2$$

(4)

Let the constant $\Delta W=W_f-W_s$:

$${\left(\Delta w+l_1{\cos \theta }_1\right)}^2+{\left(d-l_1{\sin \theta }_1\right)}^2=l_2^2$$

(5)

After rearranging, a trigonometric equation for ${\theta }_1$ can be obtained:

$$2\Delta w{l}_1{\cos \theta }_1-2d{l}_1{\sin \theta }_1=l_2^2-l_1^2-\Delta w^2-d^2$$

(6)

After calculating ${\theta }_1$, the exact coordinates $\left(x_1,y_1\right)$ of $P_1$ can be determined. Subsequently, it is necessary to calculate the spatial pose of the connecting rod $L_2$ (i.e., the absolute angle ${\theta }_2$ of the vector $P_1{P}_2$), because the gripper fingertip is fixed to the extended structure of $L_2$:

$${\theta }_2=\mathrm{atan}2\left(y_2-y_1,-W_s-x_1\right)$$

(7)

According to the diagram, the gripper finger is a single rigid body containing $L_2$, $L_3$, and $L_4$. $L_3$ and $L_4$ are converted into fixed structural angles relative to the vector $P_1{P}_2$. Let the structural deflection angle of $L_3$ relative to $L_2$ be ${\delta }_3$, and the structural deflection angle of $L_4$ relative to $L_2$ be ${\delta }_4$. The coordinates $\left(x_1,y_1\right)$ of the fingertip point $P_1$ at the end of the left gripper are:

$$\left\{\begin{array}{c}{x}_{tip}=-W_s+L_3\cos \left({\theta }_2+{\delta }_3\right)+L_4\cos \left({\theta }_2+{\delta }_4\right)\\ y_{tip}=y_2+L_3\sin \left({\theta }_2+{\delta }_3\right)+L_4\sin \left({\theta }_2+{\delta }_4\right)\end{array}\right.$$

(8)

Since the mechanism is completely symmetrical, the total opening degree $S$ is twice the absolute value of the left fingertip’s X-coordinate: $S=2x_{tip}$.

To avoid damaging the strawberry bare-root seedling during the gripping process, the seedling is treated as a cylinder for force analysis, as shown in Figure 8. During the transient process when the gripper finally closes and compresses the root-stem junction, an angle exists between the two gripper jaws, and a force analysis is conducted on this configuration.

Since the working surface of the gripper is flat, the stem is treated equivalently as a cylinder with a radius of $R_c$. Therefore, the gripping model is microscopically simplified into a bilateral line contact problem. Let the normal squeezing forces exerted by the unilateral gripper jaws on the stem epidermis be $F_{n1}$ and $F_{n2}$, and the contact length be $L$. According to the Hertz contact theory, the elastic modulus is $E^{*}$, and the equivalent radius of curvature of the system is $R^{\mathrm{*}}=R_c$. Then the maximum microscopic contact stress ${\sigma }_{max}$ on the central axis of the contact surface is:

$${\sigma }_{max}=\sqrt{{\displaystyle \frac{F_n{E}^{*}}{\pi R_{c}L}}}$$

(9)

The stress distribution of the gripper contact is highly concentrated. To prevent the local ${\sigma }_{max}$ from exceeding the physical failure strength of the strawberry epidermal cells, it is necessary to limit the magnitude of the normal force $F_n$.

Assume that in the final clamping position, the included angle between the two flat gripper jaws is $\beta$. Let the normal positive pressure perpendicular to the jaw surface ultimately transmitted by the driving mechanism to the gripper be $F_n$. Because the gripper jaws are inclined, the normal force $F_n$ will generate an upward wedge thrust component $F_u$ in the vertical direction:

$$F_u=2F_n\sin \left({\displaystyle \frac{\beta }{2}}\right)=F_h\cdot \tan \left({\displaystyle \frac{\beta }{2}}\right)$$

(10)

At the moment of clamping, this vertical upward resultant force $F_u$ can easily cause the seedling to slip upward along the gripper jaws. To ensure that the seedling is not “squeezed out”, the maximum static friction force generated on the surface of the gripper jaws must be able to counteract this wedge component force. Let the coefficient of static friction between the gripper jaws and the strawberry epidermis be $\mu$. According to Coulomb’s law of friction, the total maximum static friction force on the bilateral planes is $F_f=2\mu F_n$. The necessary and sufficient physical self-locking condition to prevent axial slipping at the moment of clamping is:

$$F_u\le F_f⟹2F_n\sin \left({\displaystyle \frac{\beta }{2}}\right)\le 2\mu F_n\cos \left({\displaystyle \frac{\beta }{2}}\right)⟹\tan \left({\displaystyle \frac{\beta }{2}}\right)\le \mu$$

(11)

The tangent of the half-angle of the final V-shaped included angle $\beta$ between the two flat jaws must be less than or equal to the experimentally measured material static friction coefficient $\mu$.

Once self-locking clamping is completed, the manipulator will enter the 3D spatial transportation stage. At this time, the seedling is subjected to the combined effect of gravity $mg$ and horizontal inertial force $a_x$. To prevent the seedling from falling off during high-speed spatial transfer, the resultant force of the friction force and the vertical normal component force must overcome the external dynamic load:

$$2\mu F_n\cos \left({\displaystyle \frac{\beta }{2}}\right)+2F_n\sin \left({\displaystyle \frac{\beta }{2}}\right)\ge m\sqrt{{\mathrm{g}}^2+a_x^2}$$

(12)

During the horizontal transportation process, an eccentricity $L_g$ exists between the center of gravity and the gripping center. To prevent the seedling from tilting in its posture during high-speed transportation, the gripper must provide sufficient anti-overturning friction torque at the effective contact surface with the stem. The maximum anti-overturning friction torque $M_f$ that can be provided under the combined action of the bilateral gripper jaws is:

$$M_f={\displaystyle \frac{1}{2}}\mu F_{n}L.$$

(13)

To ensure the absolute vertical stability of the seedling during the spatial transfer process, it is necessary to satisfy the condition that the anti-overturning friction torque is greater than or equal to the overturning torque, i.e., $M_f\ge M$, from which the lower limit constraint condition of the normal gripping force to prevent deflection can be derived:

$$F_n\ge {\displaystyle \frac{2m{a}_x{L}_g}{\mu L}}$$

(14)

Combining the equations with the $F_m$ biological yield red line for radial compression of the plant, the full-process gripping conditions under this dual-jaw V-shaped spatial architecture are ultimately established as follows:

$$\left\{\begin{array}{l}\tan {\displaystyle \frac{\beta }{2}}\le \mu \\ 2\mu F_n\cos {\displaystyle \frac{\beta }{2}}+2F_n\sin {\displaystyle \frac{\beta }{2}}\ge \sqrt{{\left(mg\right)}^2+{\left(m{a}_x\right)}^2}\\ 2F_n\cos {\displaystyle \frac{\beta }{2}}\le F_m\\ {\displaystyle \frac{1}{2}}\mu F_{n}L\ge m{a}_x{L}_g\end{array}\right.$$

(15)

2.4. Control System Design

2.4.1. Hardware System

The hardware circuit diagram of the control system for the entire test bench is shown in Figure 9, which is mainly divided into a signal input module, a control processing module, a drive execution module, and a power supply module. In the signal input module, there are 4 proximity switches used for positioning, along with one each of a start/stop switch, a reset switch, and an emergency stop switch. The control processing module is primarily the PLC processor, which is selected based on the number of input and output ports: XD10-60T10-C.

The drive execution module includes the stepper motors and servo motors responsible for the seedling feeding, picking, and dropping operations, as well as their corresponding motor drivers. The power supply module is connected to a 220 V household power supply and converts AC to direct current DC through a switching power supply to provide power to the other components of the test bench.

The specific models and versions of the selected electronic components, including sensors, motor drivers, and motors, are detailed in

Table 1. The version of the components.

NO.	Equipment	Model
1	Seedling feeding motor	42 Stepper motor (Leadshine, Shenzhen, China)
2	Seedling feeding motor driver	DM542 (ShidaiChaoqun, Beijing, China)
3	Seedling feeding point proximity switch	TL-W3MC1 NPN Normally Open (Omron, Kyoto, Japan)
4	Gripper motor	35 Stepper motor (Aohude, Guangzhou, China)
5	Gripper driver	DM420 (ShidaiChaoqun, Beijing, China)
6	Gripper proximity switch	Microswitch (Yueyi, Wenzhou, China)
7	Horizontal module driver	SD300-20AL-GBN (ShidaiChaoqun, Beijing, China)
8	Horizontal module motor	sd60aea04030-sc3-ap (ShidaiChaoqun, Beijing, China)
9	Horizontal module proximity switch	TL-W3MC1 NPN Normally Open (Omron, Kyoto, Japan)
10	Vertical module motor	28 Stepper motor (ZDT, Guangzhou, China)
11	Vertical module driver	DM420 (ShidaiChaoqun, Beijing, China)
12	Vertical module proximity switch	TL-W3MC1 NPN Normally Open (Omron, Kyoto, Japan)
13	PLC	Xinje xd560t10-c (Xinje, Wuxi, China)

. The controller plays a central role in the overall system, taking responsibility for data computation and the coordinated scheduling of input and output signals. To ensure the hardware selection meets the system’s control requirements, the controller must be equipped with at least 5 input channels and 14 output channels, as shown in

Table 2. Input and output of the PLC.

Input	Description	Output	Description
X0	Start-Stop button	Y0	Horizontal module pulse
X1	Scram button	Y1	Seedling picking vertical motor pulse
X2	Horizontal module motor zero point	Y2	Seedling picking gripper pulse
X3	Conveyor motor zero point	Y3	Conveyor device pulse
X4	Vertical module motor zero point	Y10	Horizontal module direction
X5	Seedling picking gripper motor zero point	Y11	Seedling picking vertical motor direction
X6		Y12	Seedling picking gripper motor direction
X7		Y13	Conveyor device motor direction

. This serves as the key criterion for selecting the PLC.

2.4.2. Software System

Figure 10 illustrates the software flowchart of the seedling picking device, detailing the cyclic process of the coordinated conveying and grasping operations for strawberry seedlings. The entire system is primarily divided into three stages: the reset phase, the bare-root seedling conveying phase, and the picking and releasing phase. During the reset phase, the feeding and picking mechanisms are secured in their initial positions, with the picking mechanism positioned directly above the picking point. Consequently, during the bare-root seedling conveying phase, a pre-feeding operation is executed first by delivering three seedlings. Once seedlings are present at the picking point, the continuous picking cycle commences. In the picking and releasing phase, the system executes the picking action by sequentially triggering the corresponding sensors to complete a descending clamping motion. After grasping a seedling, the mechanism moves directly above the releasing point to drop it. Upon releasing the seedling, the device returns to the position directly above the picking point to grasp the next one, thus forming a continuous picking and releasing cycle. Within the PLC (XDPPro) software architecture, the system employs rising-edge trigger logic to execute the DRVA instruction, ensuring high-speed absolute positioning control for each module.

2.5. Test Stand Experiment

To verify the picking performance of the automatic bare-root strawberry seedling picking device, and to evaluate the conveying accuracy of the bare-root seedlings during intermittent transport and spatial handling, as well as the kinematic stability and damage-free gripping effect of the pick and place manipulator, this study constructed an automatic seedling picking and feeding device test bench and conducted tests at a laboratory in Jiangsu University (Zhenjiang, China) on 20 July 2026, as shown in Figure 11. The experimental hardware system utilized a XINJE XD5-60T10 PLC as the main control unit, which, combined with stepper and servo motor drivers, proximity switches, and micro switches, formed the control and execution system. The setup was also equipped with auxiliary testing tools, including FSR402 thin-film pressure sensors, vernier calipers, vernier height gauges, stopwatches, marker pens, and a laptop installed with control and data acquisition software. The benchmark materials selected for the experiment were bare-root “Hongyan” strawberry seedlings, a variety widely used in production. These seedlings were propagated via stolons and grown for 60 to 70 days, featuring uniform growth and 3 to 5 expanded leaves. All seedlings were strictly screened prior to the experiment to ensure their physical characteristics met the agronomic requirements for fully automatic transplanting.

2.5.1. Experimental Design

To investigate the key parameters affecting the efficiency and accuracy of automatic seedling picking, this study selected picking frequency, gripping position, and gripping gap—which significantly influence the grasping success rate—as the core experimental factors.

The picking frequency (A) determines the production rhythm of the entire machine and directly impacts the complex dynamic loads borne by the manipulator during the three-dimensional spatial handling phase. An excessively high frequency causes a surge in dynamic loads, increasing the risk of grasping failure or seedling drop. Conversely, a frequency that is too low reduces the overall operational efficiency, failing to meet the performance requirements of high-efficiency transplanters. The gripping position (B) is critical in determining the spatial posture of the bare-root seedling after it is grasped. An improper gripping position can disrupt the center-of-gravity balance of the suspended seedling, causing posture deviation during transport and thereby severely affecting the subsequent dropping accuracy. The gripping gap (C), defined as the distance between the bottom edges of the flat plates on both sides of the gripping mechanism, is determined by the descent height of the manipulator’s vertical module. A proper gap is essential for ensuring damage-free and stable gripping. If the descent height is too small (resulting in a gap that is too narrow), it frequently leads to missed stems or mechanical damage to the seedlings; if the descent height is too large (resulting in a gap that is too wide), it results in unstable gripping, increasing the risk of seedlings slipping out during transport.

To evaluate the primary and secondary effects of the experimental factors, range analysis was employed to process the success rate data. The experiment utilized an $L_9(3^3$) orthogonal array. For the nine parameter combinations, 50 independent gripping cycle tests were executed for each group. The ratio of the number of successful pickings to the total number of attempts was taken as the comprehensive seedling picking success rate for that group, and range analysis was subsequently used to determine the order of influence of the factors.

The specific calculation rules for the range analysis are as follows: First, $K_{ji}$ is defined as the sum of the test indices for factor j at level i, and $k_{ji}$ is its corresponding mean value ($k_{ji}=K_{ji}/3$). The optimal level for each factor can be determined by comparing the values of $k_{ji}$. Next, the range value $R_j$ for factor j is calculated as the difference between the maximum and minimum mean values ($R_j=\max \left(k_{ji}\right)-\min \left(k_{ji}\right)$). A larger $R_j$ value indicates a more significant influence of the factor on the seedling picking success rate.

In summary, this study aims to improve the comprehensive performance of automatic seedling picking by optimizing these three critical variables. Therefore, these three parameters were selected as the experimental factors for a three-factor, three-level orthogonal experiment, with the specific factor levels presented in

Table 3. Table of factor levels.

Level	Picking Frequency (Seedlings/min)	Gripping Gap (mm)	Gripping Position (mm)
1	10	5	20
2	15	6	25
3	20	7	30

.

2.5.2. Test Methods and Evaluation Indicators

Seedling picking success rate $Y_1$: The bare-root strawberry seedlings are successfully transported from the conveyor belt to the picking point, extracted by the picking manipulator, and smoothly delivered to the dropping point without falling or experiencing posture changes.

$$Y_1={\displaystyle \frac{N_1}{N}}$$

(16)

In the formula, N₁—number of successfully picked seedlings; N—total number of seedlings.

The experiment was carried out strictly in sequence according to the nine sets of parameter combinations arranged in the orthogonal array. The specific seedling feeding and picking processes are illustrated in Figure 12a and Figure 12b, respectively. To ensure the accuracy of spatial positioning, all kinematic modules of the system strictly executed a return-to-zero (homing) reset procedure prior to each test cycle. For each parameter combination, 50 independent gripping cycle tests were repeated. The operational status of each cycle was recorded in detail to calculate the seedling picking success rate for that specific group. Upon completion of the experiment, range analysis is performed. The objective was to ascertain the primary and secondary effects of each experimental factor on the picking performance and, ultimately, to optimize and deduce the best combination of operational parameters.

3. Results and Discussion

3.1. Study Results on the Physical Characteristics of Bare-Root Strawberry Seedlings

3.1.1. Results of Basic Characteristic Parameters

Through measurement experiments on the morphological parameters of bare-root strawberry seedlings and the contact characteristics between the seedling stems and the nylon plate, the relevant morphological parameters were successfully obtained, with specific data detailed in

Table 4. Analysis of Basic Characteristic Parameters.

Parameter	Mean	Standard Deviation	Coefficient of Variation (%)
Plant length la (mm)	213	28.5	13.3
Root length lb (mm)	90	16.5	18.2
Leaf height lc (mm)	89	24.2	27.2
Stem diameter ld (mm)	17	1.25	18.2
Seedling width le (mm)	104	26.8	25.6
Distance from center of gravity to origin lg (mm)	17	4.30	25.27
Mass $m$ (g)	4.054	2.242	55.30
Coefficient of friction with nylon board	0.414	0.073	3.24

. According to the data in the table, the coefficients of variation for the morphological parameters of the bare-root strawberry seedlings are all greater than 10%, indicating a high degree of dispersion and inconsistency in their morphology. This demonstrates that the significant morphological differences among the seedlings impose higher adaptability requirements on the automatic picking device. The aforementioned parameters can provide a dimensional reference for the design of the intermittent conveying device and the picking mechanism’s end-effector, as well as for determining the gripping position, the gripper opening size, and the spatial configuration of the picking device.

3.1.2. Results of Radial Compression Tests on Bare-Root Strawberry Seedlings

As shown in Figure 13, the variation trends of the radial compression force-displacement curves for the stems of all samples are highly consistent, and the force-deformation process can be distinctly divided into two stages.

The first stage is the elastic deformation zone (section OA), where the compressive load increases approximately linearly with the displacement. The stress applied during this stage does not cause substantial damage to the plant tissues, and the stem can completely recover its original shape relying on its own elasticity after unloading. Point A is identified as the “bio-yield point” of the stem. Upon passing point A and entering the second stage, with the continuous application of the load, the stem epidermis and internal vascular bundle cells rupture, resulting in irreversible structural mechanical damage. Such tissue destruction severely hinders the subsequent growth and development of the seedlings and significantly reduces the transplanting survival rate. The tests determined that the minimum bio-yield force corresponding to point A at this location is 28 N. In the current research and development of automated transplanting equipment, most directly adopt the rigid gripping methods designed for standard plug seedlings. However, when gripping bare-root seedlings with high morphological variability, this traditional rigid structure is highly prone to instantaneous gripping force overload, thereby crushing the fragile internal vascular bundles. This study determined the critical bio-yield value of 28 N for bare-root strawberry seedlings, establishing a mechanical boundary for the automatic seedling picking mechanism.

Combined with the derivation of kinetic formulas, under the conditions that the seedling picking end-effector operate at $\beta ={20}^{\circ }$, $\mu =0.45$, and without considering vertical acceleration, to ensure a damage-free and slip-proof compliant grip by the claws on the stems of bare-root strawberry seedlings, the theoretical safety range for the normal gripping force is as follows:

$$0.15 \mathrm{N}\le F_n\le 14.22 \mathrm{N}$$

(17)

In this study, the FSR402 thin-film pressure sensor was calibrated using a static loading method with weights ranging from 50 to 1000 g in increments of 50 g. During actual seedling picking tests, when the closing gap of the claw was set to 5 mm, 6 mm, and 7 mm, the generated gripping forces were 1.2 N, 0.9 N, and 0.7 N, respectively. The system leverages the passive compliance of the nylon material; through its own macroscopic bending deformation, the non-metallic claw effectively compensates for the potential overdrive displacement caused by the 35-stepper motor driving the linkage. This purely electromechanical passive flexible compensation mechanism ensures that the actual gripping force is maintained within the theoretical safety range. Research by Hu et al. on a flexible gripping device for pepper plug seedlings indicated that seedling picking end-effector s can adapt to applied loads through physical deformation, significantly reducing contact stress and thereby effectively minimizing mechanical damage to fragile stems [28]. This demonstrates that the mechanical output characteristics of the proposed design meet the operational requirements for highly adaptable, damage-free, and slip-proof gripping of bare-root strawberry seedlings.

3.2. Design Results of Key Components

3.2.1. Design Results of the Conveying Device

Based on the previously measured dimensional characteristics of bare-root strawberry seedlings, the key structural parameters of the intermittent conveying device were established: the effective conveying width of the device is l_b = 250 mm and the total length is l_a = 560 mm. To accommodate the length differences in various seedlings, the span D of the double-ended adjustable support is dynamically adjustable within the range of 110 mm to 170 mm. Ten sets of adjustable support devices are arranged at equal intervals with a pitch P = 110 mm along the entire conveyor belt, with a V-shaped support angle set to $110°$ and a support height of 10 mm. The physical conveying device is shown in Figure 14.

Currently, automatic seedling supply systems for soilless crops mostly utilize fixed-span conveyor belts or corrugated squeezing disks. When dealing with bare-root strawberry seedlings that exhibit extreme length variations, these existing devices are highly prone to causing short seedlings to slip or long seedlings to shift their center of gravity due to the fixed support reference, severely reducing the accuracy of subsequent gripping. The dynamically adjustable span of 110~170 mm designed in this study accommodates the high morphological coefficient of variation in bare-root strawberry seedlings. It ensures absolute stability of the center of gravity for all types of seedlings and suspends the middle stem area, reserving a working window for the end effector. In their study on a sweet potato transplanter, He T et al. pointed out that inconsistencies in morphological characteristics, such as the rhizome diameter of seedlings, are the core inducing factors for gripping detachment [29]. The variable-span physical constraint strategy adopted in this paper fundamentally overcomes this common issue, demonstrating the reliability of this dimensional parameter in the seedling supply process for non-standardized crops. Furthermore, ten sets of adjustable support devices are arranged at equal intervals with a pitch P = 110 mm along the entire conveyor belt, with a V-shaped support angle set to 110° and a support height of 10 mm. Existing belt-type intermittent conveying mechanisms often lack three-dimensional posture constraints for the scattered stems and leaves of crops during stepping starts and stops. This design flaw causes materials to easily undergo lateral rolling or posture deflection under the inertial effect of motor acceleration and deceleration, rendering the preset seedling picking point positioning ineffective. The 110° V-shaped support angle and 10 mm support height set in this study achieve a conformal physical limit and gathering of the crown and root of bare-root seedlings without damaging the tender petioles. At the system execution level, the intermittent stepping action of the conveyor belt’s stepper motor is fully controlled by an Xinje PLC, operating based on S-curve acceleration and deceleration speed planning. The integration of the physical limit structure and the PLC’s S-curve motion control eliminates mechanical impact during the conveying start and stop phases, ensuring the smooth and precise step-by-step conveying of bare-root seedlings to the picking station. Research by Yao M and other scholars on the precise positioning control of transplanter conveying devices explicitly emphasizes that an exquisite lifting configuration and underlying smooth motion planning are key to eliminating transient inertial displacement [30]. The test results of the V-shaped limit combined with PLC control in this paper are consistent with the conclusions of that literature, verifying the high stability of the design results of this conveying device.

3.2.2. Design Results of the Seedling Extraction Manipulator

The parameter planning for the motion modules of the seedling pick-up manipulator is detailed in

Table 5. Motion parameters.

Motion	Maximum Stroke (mm)	Maximum Speed (mm/s)	Acceleration/Deceleration Time (s)
Horizontal module	600	1000	0.5
Vertical module	50	100	0.1
Gripper slider	20	100	0.1

. The maximum stroke of the horizontal module is 600 mm, the vertical module is 50 mm, and the gripper slider is 20 mm. Given that the maximum diameter of the scattered stems at the gripping position of bare-root strawberry seedlings reaches up to 27 mm, and taking positioning errors into account, the maximum initial opening distance of the gripper is set to 40 mm to ensure it can completely envelop the bare-root seedlings. Existing seedling pick-up manipulators typically feature smaller, fixed-size end openings. When dealing with bare-root strawberry seedlings—which are characterized by clustered, scattered multiple stems and highly variable diameters—their enveloping space is severely insufficient. This design flaw makes the gripper highly prone to physical interference with free-standing stems during its descent, resulting in missed gripping or direct mechanical damage to the stems [31]. In this study, the large initial opening of 40 mm, combined with the end slider directly driven by a 35-type stepper motor, provides ample fault-tolerant space for pre-grasping. This ensures that the gripper can completely envelop the scattered bare-root seedlings without any physical interference during its descent. Regarding the collaborative operation of the entire system, PLC motion control and S-curve acceleration/deceleration velocity planning are utilized to ensure the high-speed and smooth operation of each module.

Combined with physical space constraints, the core linkage dimensions of the seedling gripping jaws were finally determined as follows: $L_1=90 \mathrm{mm}$, $L_2=114 \mathrm{mm}$, $L_3=80 \mathrm{mm}$, $L_4=150 \mathrm{mm}$, the frame parameters $d=155 \mathrm{mm}$ and $h=200 \mathrm{mm}$, and the half-width of the driving crossbeam $w=87.5 \mathrm{mm}$. The kinematics model established in MATLAB R2023a, as shown in Figure 15a, and the trajectory simulation of the gripping points, as shown in Figure 15b, demonstrate that the actual gripping points of the left and right grippers exhibit a significant ‘diagonally upward converging’ characteristic during the closing process [32,33].

Starting from the maximum opening state at the bottom, as the slider advances, the gripping points move toward the central axis while generating a significant vertical elevation in height, ultimately reaching the preset gripping clearance at the closing point. Most existing stem-clamping seedling pick-ups rely on pneumatic cylinders to generate a single radial horizontal direct pushing force when closing the jaws. Such a rigid and parallel linear pushing trajectory fails to effectively gather scattered petioles when squeezing strawberry stems; instead, it easily generates an outward wedge-shaped repulsive force, resulting in severe slippage, seedling dropping, or radial skin bruising during the transplanting process. Driven precisely by a 35-type stepper motor, the innovative four-bar V-shaped jaw mechanism proposed in this study utilizes a large lateral displacement with a wide span in the early stage of closure to widely touch and gather the scattered stems around. As the jaws continue to close, the vertical lifting action of the trajectory, combined with the guiding effect of the inner wall of the V-shaped jaws, smoothly gathers the stems toward the center. The actual clamping process of the gripper is shown in Figure 16.

During the control instruction execution stage, the high-speed absolute positioning action of the 35-type stepper motor is precisely triggered and executed via the rising edge of the DRVA instruction inside the PLC, which eliminates position drift of the mechanism under complex loading conditions and realizes a flexible gripping operation mechanism of “first converging, then clamping”. Ye et al. [34], in their research on a planetary non-circular gear train transplanting mechanism for potted flower seedlings, pointed out that optimizing the nonlinear composite motion trajectory can significantly reduce mechanical damage and the seedling dropping rate during transplanting. The three-dimensional spatial trajectory of “oblique upward converging” obtained from the simulation in this paper intuitively and clearly confirms this cutting-edge argument, ensuring the authenticity and high reliability of the design results for this end-effector.

Compared with previously developed seedling-picking mechanisms, which typically employ rigid pneumatic grippers designed for standardized plug seedlings, the proposed system offers substantial advantages for handling bare-root strawberry seedlings. Conventional rigid grippers apply a linear direct force, which tends to cause mechanical damage to the fragile, clustered stems of bare-root seedlings, or fails to completely envelop them due to limited initial opening. In contrast, the present mechanism employs a stepper motor and a four-bar linkage to implement a “gather-then-clamp” strategy. The large initial opening ensures high fault tolerance during the descent phase, thereby effectively avoiding physical interference. Furthermore, unlike fixed-span conveyor belts that are inadequate for accommodating variable root lengths, the system integrates a variable-span V-shaped support, which provides stable postural constraint for non-standardized bare-root seedlings and thus significantly enhances the precision and stability of subsequent gripping operations.

3.2.3. Results of the Orthogonal Experiment on Seedling Pick-Up Performance

The bench orthogonal test results are shown in

Table 6. Experimental results.

Serial Number	Picking Frequency (Seedlings/min)	Gripping Gap (mm)	Gripping Position (mm)	Success Rate of Seedling Picking ${\mathit{Y}}_1$ (%)
1	10	5	20	96
2	10	6	25	94
3	10	7	30	88
4	15	5	25	98
5	15	6	30	86
6	15	7	20	86
7	20	5	30	90
8	20	6	20	92
9	20	7	25	98

. Under the set three factors and three levels, the overall average seedling picking success rate of the system reached 92%. According to the range analysis in

Table 7. Range analysis table.

${\mathit{k}}_1$	${\mathit{k}}_2$	${\mathit{k}}_3$	$\mathit{R}$
92.67	90.00	93.33	3.33
94.67	90.67	90.67	4.00
91.33	96.67	88.00	8.67

, the range values (R) for the picking frequency, gripping gap, and gripping position are 3.33, 4.00, and 8.67, respectively. The order of influence of each factor on the success rate, from primary to secondary, is as follows: gripping position > gripping gap > seedling pick-up frequency.

The bench orthogonal test results are shown in Table 6 and Table 7. Under the set three factors and three levels, the overall average seedling pick-up success rate of the system reached 92%. According to the range analysis of the test results, the order of influence of each factor on the seedling pick-up success rate, from primary to secondary, is as follows: gripping position > gripping gap > seedling pick-up frequency.

The picking point position exerts the most significant influence on the success rate. The data indicate that the closer the picking point is to the shortened stem, the higher the seedling picking success rate. When the picking point is located 25 mm from the root, the eccentricity between the grasping point and the center of gravity is minimized. This significantly reduces the overturning moment generated during high-speed spatial transfer, thereby ensuring postural stability. The minimum closing gap is the second most influential factor affecting the success rate. An excessively large gap results in insufficient clamping force, making it difficult to overcome inertial and gravitational forces during the transfer process, which leads to dropped seedlings. Conversely, a gap that is too small may cause excessive clamping or prevent smooth closure; therefore, an optimal critical value exists. The seedling picking frequency has the most minimal impact on the success rate, indicating that the stability of the picking and dropping mechanism remains guaranteed at an operating frequency of 20 seedlings/min.

Based on the comprehensive range analysis results, the optimal operating parameter combination for the system was determined to be A3B1C2. Fifty repeated verification tests were conducted under this optimal parameter combination, yielding a final seedling pick-up success rate of 98%. However, this study has certain limitations. The validation experiments exclusively utilized the “Hongyan” strawberry cultivar. Although the mechanism is equipped with a variable-span support structure and a wide-opening gripper designed to accommodate significant morphological variations, its applicability to other cultivars with fundamental differences in stem diameter and length requires further testing. Furthermore, the single instance of picking failure indicates that extremely weak or severely deformed seedlings still exceed the current fault-tolerant limits of the gripper. Moreover, these tests were conducted on a stationary test bench, which does not account for the complexities of actual field conditions. During field operations, factors such as high-frequency mechanical vibrations, posture deflection caused by uneven terrain, and dust interference with sensors could potentially degrade spatial positioning accuracy and operational success rates. Therefore, future research will involve extensive field trials across multiple strawberry cultivars to comprehensively validate and enhance the operational robustness of the transplanter.

Damage assessment was conducted through visual observation. One day after the pick-up operation, the clamped stems of the bare-root strawberry seedlings were visually inspected for any signs of surface bruising, epidermis rupture, or crushed vascular tissues. Under the optimal parameter combination, no such mechanical damage was observed. The survival rate after transplanting reached 100% with good growth status, effectively meeting agronomic requirements.

An analysis of the single failed seedling pick-up case in the verification tests revealed that this specific plant was weak and small, with its stems in an extremely scattered state. This extreme morphology exceeded the maximum enveloping fault-tolerant limit of the current gripping jaws, resulting in a failure to effectively gather the stems during closure. This outcome also indicates that while the current system possesses extremely high adaptability to conventional and high-quality seedlings, extremely weak, small, or deformed seedlings still need to be controlled during the front-end feeding or grading and screening stages.

4. Conclusions

To address the challenges of heavy reliance on manual feeding and the poor adaptability of existing grippers for bare-root strawberry seedlings with clustered stems and exposed roots, this study developed an automatic seedling picking and feeding system integrating an intermittent conveying device with a variable-span V-shaped support and a flexible manipulator.

An automatic seedling pick-up and dropping test bench for bare-root strawberry seedlings was constructed, and bench pick-up tests were carried out. A three-factor, three-level orthogonal experiment was conducted using the seedling pick-up frequency, gripping position, and gripping gap as the test factors. The results indicate that the order of influence of each factor on the seedling pick-up success rate, from primary to secondary, is as follows: gripping position, gripping gap, and seedling pick-up frequency. Under the optimal parameter combination—specifically, a seedling pick-up frequency of 20 plants/min, a gripping point located 25 mm from the rhizome, and a jaw gripping gap of 5 mm—the seedling pick-up success rate reached 98%, and visual observation indicated no mechanical damage to the seedlings. This effectively meets the operational requirements for highly adaptable, non-destructive, and fully automated transplanting of bare-root strawberry seedlings.

This paper proposes an automatic seedling pick-up and dropping mechanism for bare-root strawberry seedlings that coordinates intermittent conveying with a flexible manipulator. The timing for the seedling pick-up and dropping process and the three-dimensional motion path were planned. Furthermore, the overall system layout of the device was determined, and the basic structures and working principles of the variable-span V-shaped intermittent conveying device and the end-effector were elaborated.

Taking ‘Hongyan’ bare-root strawberry seedlings grown for 60 to 70 days after stolon propagation as the research object, experiments on their physical appearance characteristics and the radial compressive mechanical properties of the stems were conducted. A safe threshold for non-destructive clamping was defined, bounded by a biological yield point of 28 N. The kinematic equations and the friction self-locking force model during the spatial handling and gripping processes were established, completing the design of the key parameters for the seedling pick-up end-effector.