Development of control system for robotic apple harvesting device

. This article focuses on the control system of a robotic device designed for efficient apple harvesting with minimal fruit damage. The developed device is equipped with specialized mechanisms and sensors aimed at reducing negative impacts on fruits during harvesting. A control system for the robotic device was developed, incorporating various sensors and modules. A module for determining the position of the grip arms was designed, consisting of a polymeric magnetic strip attached to a spiral-shaped cutout. A module to track the position of the grip arms was also developed, capable of indicating the current position of the grip arms, fully open and closed grip, as well as intermediate values. A module to monitor the grip force was designed, used to control the gripping force applied to the fruit. A current sensor was connected to the winding of the linear actuator motor to measure the force. Increasing the force applied by the grip arms to the fruit results in higher current flowing through the linear actuator motor winding. By reading the current sensor data, the degree of compression on the fruit is determined. During testing of the control system modules and sensors, it was found that the grip arm angle position module has the highest sensitivity in the range of 0 to 90 degrees. The obtained data after calibration enable the control of the degree of grip arm opening. The grip arm position module allowed controlling and adjusting the grip arm position with an accuracy of up to 2 mm. A control system for a robotic grip for apple harvesting was developed. Various modules contributing to efficient apple harvesting were designed. One of the key modules is the grip force control module, which regulates the degree of fruit compression during robotic fruit harvesting, thus minimizing fruit damage.


Introduction
Modern agriculture aims to increase efficiency and optimize harvest processes.One key direction in this area is the automation of processes, particularly the robotic harvesting of fruits.Robotic systems offer a promising solution for increasing productivity, reducing labor costs, and improving harvest quality.Currently, various companies are actively engaged in producing robotic systems for apple harvesting.However, it is important to note that these system developments often pay insufficient attention to the gripping process, which is the direct contact link between the device and the fruit [1][2][3][4][5][6].Existing developments often limit themselves to basic fruit grasping, overlooking the physical and mechanical properties and peculiarities of the fruit.An effective solution to this issue involves the development of specialized modules and the use of sensors to account for the specifics of each fruit.Utilizing modern sensors allows for consideration of the individual characteristics of each fruit.These data can be crucial for optimizing the gripping process, making it more delicate and efficient.This approach contributes to increased productivity and harvest quality.The main component of robotic fruit harvesting systems is the control system, which is responsible for coordinating movements and gripping force.The effectiveness and reliability of the control system directly impact the harvesting process and, consequently, the quality of the harvested product.

Materials and Methods
At the Federal Scientific Agroengineering Center VIM, a prototype of a robotic gripper was developed for the delicate harvesting of apple fruits from trees (Figure 1).This prototype comprises various mechanisms designed to harvest fruits without causing damage.However, to effectively utilize these mechanisms, a control system needs to be developed and implemented.This system should operate automatically, continuously monitoring the functioning parameters of the gripping mechanisms.
The primary task is to design and configure a control system that takes into account the specifics of the gripping mechanisms, harvesting conditions, and the requirement for gentle fruit detachment.The utilization of modern automation methods and technical solutions has enabled the creation of a control system to optimize the harvest process.The rotation control module in the developed system is based on the use of a Hall sensor and a magnetic strip.This module ensures precise measurement and fixation of the rotation angle of the gripper.In the developed control module, the Hall sensor is positioned at a specific distance from the magnetic strip, which is attached to the rotating part of the gripper's housing.The movable base features a spiral-shaped cutout for the magnetic strip, positioned at the center of the base (Figure 3).This cutout is designed to alter the distance between the magnetic strip and the Hall sensor during the gripper's rotation [7][8].
1 -minimum cutout distance; 2 -maximum cutout distance.grip position module capable of fixing the position of the grip's claws.Due to the 360° rotation of the grip's claws, the use of contact sensors becomes impractical.To determine the position of the grip's claws, non-contact sensors are required.To register the distance between the pusher and the cover of the stationary base, a Hall sensor and a ring magnet attached to the pusher were used.The Hall sensor was installed on the cover of the external base of the robotic grip (Figure 4).A module for controlling the grip force of the gripper was developed, and a current sensor was used as a sensitive element.The current sensor is a device capable of measuring the electric current flowing through the linear actuator winding.It can register even small changes in the current, allowing us to monitor and control the forces exerted by the gripper during the apple harvesting process.The data obtained from the Analog-to-Digital Converter (ADC) is processed, and based on this data, a signal is sent to the actuator mechanisms.This enables us to record changes in current associated with the forces exerted by the gripper and take appropriate actions to ensure the necessary compression force on the fruit [9][10][11].
During the data collection from the ADC, it was observed that the obtained values represent a non-linear dependence (Figure 5).This phenomenon can be explained by the use of a driver based on the pulse-width modulation method for motor control.As a result of different motor operation modes, abrupt changes in the obtained data occur.

Fig. 5. ADC Data during grip compression
The amplitude values were used in the processing of ADC data, reflecting changes in the power consumed by the actuator motor.The amplitude of the signal obtained from the ADC is related to the electrical power consumed by the motor since increasing the force of the grip requires more energy for compression [12][13][14].

Results and Discussion
As a result of the measurements of the grip rotation angle module and the data obtained using ADC, it was found that the Hall sensor exhibits the highest sensitivity in the range from 0 to 90 degrees (Figure 6).To calibrate the compression forces resulting from the actuator's influence on the gripper's plunger, a Mark-10 Series 5 dynamometer was utilized.During the calibration of the gripping forces, these forces were measured and recorded at specific time intervals.The collected data was then transmitted to a computer for further analysis.In conducting the experiment, the dynamometer was securely mounted on a stationary platform, and its plunger was connected to the system's gripper.Thus, when the actuator compressed, the dynamometer recorded and measured the forces acting on the plunger.Due to the non-linearity of the current sensor readings, data regarding the compression force was obtained experimentally (Figure 7).In the process of measurements using the dynamometer, values of compression forces were obtained, which are used for further analysis and control of forces in the gripping system.Through the analysis of these data, we can assess the forces exerted by the actuator on the pusher depending on various factors, such as the grip position and motor power [15][16].The data obtained with the dynamometer serve as the basis for calibrating and tuning the control system, as well as for determining the optimal operating parameters of the actuator and establishing the required compression forces for efficient and gentle fruit harvesting (Figure 8).The obtained data after calibration allow for controlling the degree of opening of the grip jaws.When the distance between the pusher and the lid changes, the Hall sensor values correspond to a specific degree of jaw opening.This enables the control system to monitor and adjust the position of the grip jaws with an accuracy of up to 2 mm.Measurements were conducted while varying the actuator stroke from 0 to 30 mm.During the measurements, the sensor detected the magnetic field of the ring magnet.

Conclusion
Various modules and systems for the robotic harvesting grip were developed and applied.Different sensors were utilized to monitor position, force, and other parameters during system operation.In particular, a grip jaw position module was designed and utilized, incorporating a flexible polymer magnetic strip and a specially shaped spiral cutout.This module allows for precise determination of the grip jaw angle and rotation, ensuring accurate positioning of the grip jaws.Magnetic sensors were used to monitor the grip jaw position.These sensors provide precise determination of the grip jaw position, enabling the control system to regulate and monitor their position.Analog-to-digital converters were also employed in the control system to collect data from sensors and process signals.The ADCs provided accurate measurements and supplied data about the position, force, and other parameters of the grip system.Thus, the development and implementation of various modules and systems, along with the use of different sensors, ensured efficient and reliable operation of the robotic fruitharvesting grip.These modules and systems contribute to precise positioning, fixation, and control of the grip jaw position, ultimately enhancing fruit harvesting efficiency.

Fig. 1 .
Fig. 1.Device Model for Robotic Fruit HarvestingSensors are an important part of the control system, providing information about various parameters such as the distance from the fruit to the gripper, fruit compression force, gripper

Fig. 2 .
Fig. 2. Block Diagram of the Robotic Gripper Control System

Fig. 3 .
Fig. 3. Spiral Cutout for Attaching the Magnetic StripAs a sensing element to measure the angle of rotation of the grip, an SS49 Hall sensor manufactured by Honeywell is utilized, connected to the ADS1115 ADC.The centering of the fruit in the robotic grip is a crucial aspect of the harvesting process.It ensures the optimal positioning of the fruit inside the grip, minimizing the risk of damage.We have developed a

Fig. 7 .
Fig. 7. Current Sensor Readings at Different Levels of Force Application

Fig. 8 .
Fig. 8. ADC Data for Changes in Distance between the Pusher and the Lid of the Fixed Base