In the realm of agricultural robotics, the automation of harvesting processes has emerged as a critical research frontier, driven by the need for efficiency, labor reduction, and precision. Central to any fruit-picking robot is the end effector, often referred to as the robot’s “hand,” which is mounted on the wrist joint of a robotic arm. This end effector directly interacts with the target fruit, and its performance dictates the overall success and viability of the harvesting system. While numerous advancements have been made in robotic harvesting technologies, many solutions struggle with practical implementation due to challenges such as fruit damage, low success rates, high costs, control complexity, and lack of adaptability. In this context, I propose a novel end effector design specifically tailored for spherical fruits, leveraging pneumatic actuation as its core driving mechanism. This end effector addresses key limitations by integrating a two-finger clamping system and a rotating cutting blade, both powered by compressed air, to ensure gentle handling, reliable detachment, and broad applicability.
The design philosophy of this end effector stems from the biological and physical characteristics of spherical fruits, such as apples, peaches, and oranges. These fruits typically have a relatively firm yet delicate surface, a spherical geometry, and a stem (peduncle) that must be severed without damaging the fruit or the plant. Traditional end effectors often rely on complex sensing and actuation systems, such as torque-controlled twisting or vacuum-assisted gripping combined with scissors, which can increase cost and control difficulty. My approach simplifies this by utilizing pneumatic cylinders for both clamping and cutting actions. Pneumatic systems offer distinct advantages: they provide rapid and forceful motion, allow easy adjustment of force via pressure regulation, and are lightweight, clean, and cost-effective. This end effector is designed to clamp the fruit using two concave fingers lined with soft rubber, then rotate a blade around one finger to cut the stem from any radial position, eliminating the need for precise stem localization. The entire system is engineered for high structural integrity, excellent manufacturability, and standardization of components, resulting in a prototype weighing only 1.2 kg. Through laboratory experiments with apples, this end effector has demonstrated exceptional performance, achieving a 100% success rate under optimal conditions with no fruit damage, underscoring its potential for widespread adoption in automated harvesting.
The overall design of this pneumatically actuated end effector can be broken down into two primary subsystems: the pneumatic translational clamping component and the pneumatic rotary cutting component. These subsystems work in tandem to execute the harvesting sequence. The clamping component is responsible for securely grasping the fruit, while the cutting component severs the stem to detach the fruit from the plant. The integration of these functions into a single, compact end effector is key to its efficiency and versatility. The end effector is mounted via a connection plate to the robotic arm’s wrist, allowing it to be positioned accurately within the orchard environment. The use of pneumatic actuation not only simplifies the control architecture but also enhances reliability, as pneumatic components are less prone to overheating and can operate in dusty or humid conditions typical of agricultural settings.
Delving into the structural details, the pneumatic translational clamping component forms the foundation of the end effector. It consists of a base plate that serves as the main structural frame. Attached to this base plate is a double-acting pneumatic cylinder (Cylinder I) via a cylinder bracket. The piston rod of Cylinder I is connected to a cylinder push plate, which in turn is rigidly linked to a rear rack through a rear rack connection plate. This rear rack meshes with a gear mounted on a gear shaft, which is supported by deep-groove ball bearings housed in bearing seats on the base plate. The gear also engages with a front rack, which is connected to a front rack connection plate and a front reinforcement plate. The clamping fingers—left and right—are attached to finger mounting plates; the left finger is fixed to the rear assembly, while the right finger is fixed to the front assembly. Thus, when Cylinder I extends or retracts, it drives the rear rack linearly, causing the gear to rotate and translate the front rack in the opposite direction. This symmetric motion opens or closes the two fingers in a parallel translational manner, ensuring that the fruit is centered and gripped evenly. The fingers themselves are designed with a spherical concave shape to conform to the fruit’s surface, and they are lined with a layer of soft rubber to provide cushioning and prevent bruising. This translational clamping mechanism is inherently robust and avoids the collision risks associated with rotational finger designs.
The pneumatic rotary cutting component is mounted on the clamping assembly and is responsible for stem severance. It features a second double-acting pneumatic cylinder (Cylinder II) fixed to the rear rack connection plate. The piston rod of Cylinder II is attached to a cylinder push block, which is connected to a synchronous belt connection plate. This connection plate is coupled to a linear slider that rides on a linear guide rail, allowing smooth linear motion. A synchronous belt is clamped to the connection plate via a belt clamping block. The belt runs around a front arc-toothed synchronous pulley and a rear arc-toothed synchronous pulley, with the rear pulley fixed on a rear pulley shaft. The front pulley is connected to a blade holder connection plate, which in turn is linked to a bent blade holder plate. Cutting blades (two in number for redundancy) are secured to the bent plate using clamping screws and pressure plates. When Cylinder II actuates, its linear motion is transmitted via the synchronous belt to rotate the front pulley, thereby causing the blade holder plate and blades to swing in an arc around the left finger. The cutting radius is adjustable by modifying the overlap between the blade holder plates, allowing accommodation of different fruit sizes. This design enables the blades to sweep nearly 360 degrees around the finger, ensuring that the stem can be cut regardless of its orientation relative to the end effector. This eliminates the need for complex vision systems to locate the stem precisely, significantly simplifying the robotic control task.
The advantages of pneumatic actuation in this end effector are multifaceted and can be analyzed through force and motion equations. For the clamping action, the clamping force $F_j$ exerted on the fruit is directly proportional to the pneumatic pressure $P_j$ and the cylinder bore diameter $D_j$. The relationship is given by the fundamental pneumatic force equation:
$$F_j = P_j \cdot \pi \cdot \left(\frac{D_j}{2}\right)^2$$
where $F_j$ is the clamping force in Newtons, $P_j$ is the pressure in Pascals, and $D_j$ is the cylinder bore diameter in meters. By selecting an appropriate cylinder size and adjusting the system pressure, the clamping force can be precisely tuned to be sufficient to hold the fruit securely without causing damage. For instance, to achieve a target clamping force of 50 N at a pressure of 0.15 MPa (150,000 Pa), the required cylinder diameter can be calculated as:
$$D_j = 2 \times \sqrt{\frac{F_j}{P_j \cdot \pi}} \approx 0.0146 \, \text{m} = 14.6 \, \text{mm}$$
Accounting for frictional losses and efficiency, a standard 16 mm bore cylinder is selected. Similarly, the clamping stroke $S_j$ (the distance between the fingers) is determined by the cylinder stroke $L_j$, with $S_j = 2L_j$ due to the gear-rack mechanism doubling the displacement. This straightforward relationship allows easy adaptation to different fruit sizes by simply choosing a cylinder with an appropriate stroke, bypassing the need for complex position sensors or feedback loops.
For the cutting action, the cutting force $F_q$ at the blade tip must be adequate to shear through the stem. This force is derived from the pneumatic pressure $P_q$ applied to Cylinder II, its bore diameter $D_q$, and the mechanical advantage provided by the pulley system. The torque balance equation for the cutting mechanism can be expressed as:
$$F_q \cdot \frac{D_2}{2} = \frac{\pi \cdot D_1 \cdot D_q^2 \cdot P_q}{8}$$
where $D_1$ is the pitch diameter of the front synchronous pulley (e.g., 40 mm), $D_2$ is the cutting blade rotation diameter (e.g., 140 mm), $F_q$ is the cutting force at the blade tip, $P_q$ is the pressure, and $D_q$ is the bore diameter of Cylinder II. Rearranging to solve for $D_q$:
$$D_q = 2 \cdot \sqrt{\frac{F_q \cdot D_2}{\pi \cdot D_1 \cdot P_q}}$$
Assuming a required cutting force of 30 N at a pressure of 0.15 MPa, the calculation yields $D_q \approx 0.01493 \, \text{m} = 14.93 \, \text{mm}$, leading to the selection of another 16 mm bore cylinder. The rotation angle $\Delta \theta$ of the blade is governed by the cylinder stroke $L_q$ and the pulley diameter: $\Delta \theta = 2L_q / D_1$. By adjusting $L_q$, the sweep angle can be controlled to ensure complete stem cutting without over-travel. These parametric relationships highlight how pneumatic actuation simplifies force and motion control in this end effector, reducing the reliance on expensive sensors and intricate algorithms.
To further elucidate the design parameters and their interactions, the following table summarizes key specifications and calculated values for the end effector:
| Parameter | Symbol | Value | Unit |
|---|---|---|---|
| Clamping Cylinder Bore Diameter | $D_j$ | 16 | mm |
| Clamping Pressure (Optimal) | $P_j$ | 0.15 | MPa |
| Target Clamping Force | $F_j$ | 50 | N |
| Clamping Stroke (Cylinder) | $L_j$ | Variable | mm |
| Cutting Cylinder Bore Diameter | $D_q$ | 16 | mm |
| Cutting Pressure (Optimal) | $P_q$ | 0.15 | MPa |
| Target Cutting Force | $F_q$ | 30 | N |
| Front Pulley Pitch Diameter | $D_1$ | 40 | mm |
| Blade Rotation Diameter | $D_2$ | 140 | mm |
| End Effector Mass | – | 1.2 | kg |
Another critical aspect of this end effector is its manufacturability and standardization. The custom-made parts, such as the fingers, base plate, and mounting brackets, are designed with excellent processability. For example, the spherical concave fingers are turned on a lathe from aluminum alloy (grade 2024), making them easy to produce with conventional machining techniques. This contrasts with earlier end effector designs that often involved complex, non-rotational shapes requiring specialized fabrication. The use of aluminum ensures a lightweight yet strong structure, contributing to the overall low weight of 1.2 kg, which minimizes inertial loads on the robotic arm and enhances energy efficiency. Purchased components are highly standardized, including the gear-rack sets, arc-toothed synchronous belts and pulleys, linear guides, ball bearings, pneumatic cylinders, and screws. These off-the-shelf items are readily available, cost-effective, and reliable, facilitating easy assembly, maintenance, and scalability. This focus on simplicity and standardization not only reduces production costs but also improves the end effector’s accessibility for farmers and integrators, aligning with the goal of practical deployment.
The performance of this pneumatically actuated end effector was rigorously evaluated through laboratory harvesting trials using apples as the target fruit. The prototype was integrated with a robotic arm system, and a quiet air compressor paired with two-position five-port solenoid valves supplied the pneumatic power. The working pressure was varied across a range from 0.5 to 2.5 bar (0.05 to 0.25 MPa) to assess its impact on harvesting metrics. For each pressure setting, a sample of 10 apples was harvested, and key performance indicators were recorded: the average cutting angle (which reflects fruit rotation during cutting, indicating clamping security), stem cutting success rate, average harvesting time per fruit (excluding arm movement and vision time), and fruit integrity rate (absence of damage). The results are compiled in the table below:
| Working Pressure (bar) | Average Cutting Angle (°) | Stem Cutting Success Rate (%) | Average Harvesting Time (s/fruit) | Fruit Integrity Rate (%) |
|---|---|---|---|---|
| 0.5 | 120 | 60 | 3.1 | 60 |
| 1.0 | 80 | 70 | 3.0 | 70 |
| 1.5 | 20 | 90 | 2.5 | 100 |
| 2.0 | 0 | 100 | 2.4 | 100 |
| 2.5 | 0 | 100 | 2.4 | 90 |
Analysis of the data reveals that at lower pressures (0.5 and 1.0 bar), the clamping force is insufficient, causing the fruit to rotate significantly within the fingers during cutting (average angles of 120° and 80°, respectively). This rotation stems from the stem resisting the blade, leading to poor cutting success rates (60-70%) and some fruit damage due to slippage. At 1.5 bar, performance improves markedly, with minimal rotation (20°), a 90% success rate, and no damage. At 2.0 bar, the end effector achieves optimal operation: zero rotation, 100% stem cutting success, an average harvesting time of 2.4 seconds per fruit, and perfect fruit integrity. This pressure provides adequate clamping to hold the fruit firmly, allowing the blade to cut the stem cleanly without causing fruit movement or injury. However, at 2.5 bar, while success remains high, slight fruit surface bruising occurs in some samples due to excessive clamping force, and mechanical vibrations increase, potentially affecting longevity. Thus, 2.0 bar (0.2 MPa) is identified as the ideal working pressure for this end effector when harvesting apples. These results underscore the effectiveness of pneumatic pressure tuning in balancing force application for different tasks.
The versatility of this end effector extends beyond apples to other spherical fruits like peaches, plums, and tomatoes, thanks to its adjustable clamping stroke and cutting radius. The pneumatic system allows quick reconfiguration for different fruit sizes and stem strengths by merely changing cylinder strokes or adjusting pressure settings. Moreover, the end effector’s design avoids the need for stem localization sensors, as the rotating blade covers all possible stem positions relative to the fruit. This significantly reduces system complexity and cost, making it more amenable to real-world orchard environments where lighting conditions and obstructions can challenge computer vision systems. The end effector’s lightweight construction (1.2 kg) also minimizes the payload on the robotic arm, enabling faster and more energy-efficient movements, which is crucial for covering large areas during harvest seasons.
In conclusion, this pneumatically actuated end effector represents a significant step forward in agricultural robotics by addressing core challenges in fruit harvesting. Its key innovations include the use of pneumatic cylinders for both clamping and cutting, which simplifies control and force modulation; a translational finger mechanism that ensures secure and centered gripping; a rotating blade system that eliminates stem localization requirements; and a design emphasizing manufacturability and standardization for cost reduction. The end effector has demonstrated high performance in laboratory tests, achieving 100% success rates with no fruit damage under optimal pressure settings. Future work could focus on enhancing the end effector’s intelligence through integrated tactile or pressure feedback for adaptive gripping, further miniaturizing the components for improved dexterity in dense foliage, and exploring hybrid pneumatic-electric actuation for even greater control precision. Additionally, field trials in actual orchards will be essential to validate robustness under variable environmental conditions. By continuing to refine such end effector designs, the path toward commercially viable and universally applicable fruit-picking robots becomes increasingly clear, promising to revolutionize agricultural practices worldwide.
Throughout this discussion, the term “end effector” has been emphasized to highlight its pivotal role in robotic harvesting systems. This end effector, with its pneumatic drive, stands as a testament to how thoughtful engineering can overcome traditional barriers, offering a reliable, efficient, and gentle solution for automating the harvest of spherical fruits. As research progresses, the principles embedded in this end effector—simplicity, adaptability, and reliability—will likely inform future generations of agricultural robots, paving the way for sustainable and productive farming in the years to come.