Fruit picking robot test facility

Abstract

A fruit-harvesting robot testing facility. A simulated fruit tree for training a harvesting robot includes: a support frame and a main structure and an outer shielding structure sequentially arranged on it, wherein several simulated fruits are suspended from the main structure via magnetic hanging points. This invention utilizes a simulated fruit tree structure to train a harvesting robot, employing a highly integrated structure to simulate diverse real-world harvesting scenarios, making it highly efficient and practical. The harvesting difficulty is divided into three main categories: single-fruit harvesting and multi-fruit fully ripening (i.e., all simulated fruits are ripe and need to be harvested), and multi-fruit multi-ripening (i.e., simulated fruits include fruits of different ripeness levels, only ripe fruits need to be harvested). Each main category includes three specific difficulty levels: no interference, visual interference (fruit recognition under different lighting conditions), and visual and harvesting interference (obstacle avoidance, vibration resistance), providing a unified training and evaluation standard.

Description

Technical Field

[0001] This invention relates to a technology in the field of robot control, specifically a testing mechanism for fruit-picking robots used to train them. Background Technology

[0002] Currently, training harvesting robots is often difficult to conduct frequently in actual orchards, while existing simulation scenarios in laboratories are usually quite rudimentary. They can only use pre-taken images to train visual recognition functions or suspend foam fruit models on simple structures to train grasping functions, making it difficult to train harvesting robots efficiently. Furthermore, the evaluation of harvesting functions is inconsistent, lacking a systematic and standardized reference. Summary of the Invention

[0003] This invention addresses the aforementioned shortcomings of existing technologies by proposing a simulated fruit tree for training harvesting robots. The robot is trained using a simulated fruit tree structure, employing a highly integrated structure to simulate diverse real-world harvesting scenarios, resulting in high efficiency and practicality. The harvesting difficulty is categorized into three main levels: single-fruit harvesting, multi-fruit fully ripening (i.e., all simulated fruits are ripe and require harvesting), and multi-fruit ripening (i.e., simulated fruits include fruits at different ripeness levels, but only ripe fruits require harvesting). Each main level includes three specific difficulty levels: interference-free, visual interference (fruit recognition under different lighting conditions), and visual and harvesting interference (obstacle avoidance, vibration resistance), with unified training and evaluation standards.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to a simulated fruit tree for training a harvesting robot, comprising: a support frame and a main structure and an outer shielding structure arranged thereon, wherein: a number of simulated fruits are suspended on the main structure by magnetic hanging points.

[0006] The main trunk part includes: several connecting rods and their horizontal and vertical connectors, wherein: each connecting rod includes three connecting ends, which are rotatably connected to the horizontal connectors and / or the vertical connectors respectively to form a planar ring structure or a spatial hemisphere.

[0007] The outer shielding structure is an integrated umbrella-shaped frame structure, including: a bottom guide post part and a top umbrella-shaped structure, wherein: the end of the top umbrella-shaped structure is provided with a hanging hole for hanging static fruits that do not need to be moved, or for attaching additional branches to increase the degree of shielding.

[0008] The simulated fruit is equipped with a programmable light strip to simulate different levels of ripeness, and the top of the fruit has a rectangular opening that can be magnetically attached to the main structure. Attached Figure Description

[0009] Figure 1This is a schematic diagram of the overall invention;

[0010] Figure 2 This is a schematic diagram of the main tree trunk.

[0011] Figure 3 This is the smallest unit diagram of the main body;

[0012] Figure 4 This is a diagram of the main connecting rod structure.

[0013] Figure 5 This is a diagram of the joint structure of the main body;

[0014] Figure 6 This is a structural diagram of the connecting nails for the main body.

[0015] Figure 7 This is a schematic diagram of the structure of the part being obscured;

[0016] Figure 8 This is a simulated image of a fruit. Detailed Implementation

[0017] like Figure 1 As shown, this embodiment relates to a simulated fruit tree mechanism for training a harvesting robot, including: a support frame 5 and a main structure 4 and an outer shielding structure 1 arranged thereon, wherein: a number of simulated fruits 3 are suspended on the main structure 4 by magnetic hanging points 2.

[0018] like Figures 2-6 As shown, the main trunk part 2 includes: several connecting rods 101 and their horizontal connecting parts 102 and vertical connecting parts 103, wherein: each connecting rod 101 includes three connecting ends, which are rotatably connected to the horizontal connecting parts 102 and / or the vertical connecting parts 103 respectively to form a planar ring structure or a spatial hemisphere.

[0019] like Figure 7 As shown, the outer shielding structure 1 is an integrated umbrella-shaped frame structure, including: a bottom guide post part 201 and a top umbrella-shaped structure 202, wherein: the end of the top umbrella-shaped structure 202 is provided with a hanging hole for hanging static fruits that do not need to be moved, or for attaching additional branches to increase the degree of shielding.

[0020] like Figure 8 As shown, the simulated fruit 3 has an inner and outer layer structure. The inner layer is made of PLA material and the outer layer is made of TPU material. They are the same in shape but different in size. The outer layer wraps the inner layer to form a simulated fruit. The simulated fruit 3 is equipped with a programmable light strip 301 to simulate different ripeness levels. The top of the fruit is equipped with a rectangular opening 302, which can be movably connected to the main structure 1 through the magnetic hanging point 2.

[0021] When the bottom structure drives the main trunk, the distance between the two nodes of the main structure decreases, the radius increases, and the layer spacing also increases, allowing for changes in the height and radius of the fruit tree. The outer shading structure and the main body include magnetic fruit-hanging points, enabling repeated picking and hanging of fruit, and efficiently training various functions of the harvesting robot. Furthermore, the shading structure can be combined with "branches and vines" to increase the area of ​​fruit obstruction, thereby increasing the difficulty of visual recognition. In use, the simulated fruit tree is adjusted to the required height, and fruit is hung through the magnetic points, simulating the harvesting robot's work object. Combined with ambient light and changes in fruit color, the robot's visual recognition, path planning, positioning, and grasping functions are trained and evaluated.

[0022] Compared with existing technologies, this device can efficiently test the visual recognition and positioning functions of the picking robot in a limited space. In addition, the mechanism proposes a method to classify the picking difficulty based on factors such as fruit maturity, variety, visual interference and picking disturbances.

[0023] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A simulated fruit tree mechanism for training a harvesting robot, characterized in that, include: The support frame and the main structure and outer shielding structure are arranged on it in sequence, wherein: several simulated fruits are suspended on the main structure by magnetic fruit hanging points; The main trunk includes several connecting rods and their horizontal and vertical connectors, wherein each connecting rod includes three connecting ends, which are rotatably connected to the horizontal and / or vertical connectors to form a planar ring structure or a spatial hemisphere. When the bottom structure drives the main trunk, the distance between the two nodes of the main structure decreases, the radius increases, and the layer spacing also increases, thereby realizing the change of the height and radius of the fruit tree. The training categorizes the difficulty of picking into three main levels: single-fruit picking, multiple fully ripe fruits, and multiple fully ripe fruits. Each main level includes three specific difficulty levels: no interference, visual interference, and visual and picking interference, thereby unifying the training and evaluation standards.

2. The simulated fruit tree mechanism for training a harvesting robot according to claim 1, characterized in that, The outer shielding structure is an integrated umbrella-shaped frame structure, including: a bottom guide post part and a top umbrella-shaped structure, wherein: the end of the top umbrella-shaped structure is provided with a hanging hole for hanging static fruits that do not need to be moved, or for attaching additional branches to increase the degree of shielding.

3. The simulated fruit tree mechanism for training a harvesting robot according to claim 1, characterized in that, The simulated fruit has an inner and outer layer structure. The inner layer is made of PLA material, and the outer layer is made of TPU material. They are the same in shape but different in size. The outer layer wraps around the inner layer to form a simulated fruit. The simulated fruit is equipped with a programmable light strip to simulate different ripeness levels. The top of the fruit has a rectangular opening, which can be movably connected to the main structure through a magnetic hanging point.

Images