Agricultural robot is an artificial intelligence robot applied in the agricultural production process, and is also the focus of global agricultural equipment competition. At the beginning of this year, 17 departments including the Ministry of Industry and Information Technology issued the "Robot +" application Action Implementation Plan, proposing to accelerate the intelligent transformation of agricultural, forestry, animal husbandry and fishery infrastructure and production equipment, promote the deep integration of robots with agricultural planting, aquaculture, forestry and fishery production, and support the development of smart agriculture.
As a large agricultural country, China's agricultural robot industry since the start of the 1990s, the market size continues to grow, product applications are increasingly extensive, and the technical system is gradually improved. During the 2023 World Robot Conference, Zhao Chunjiang, an academician of the Chinese Academy of Engineering, said that China's agricultural robots are shifting from general automation equipment to highly intelligent robots. Complete technological iterations are taking place in everything from communication control to end-effector operation.
Agricultural robots are more diverse and more powerful
Field planting, fruits and vegetables, breeding scenes gradually from mechanization to intelligence
Tongxiang City, Zhejiang Province, there is a engaged in the development of edible fungi, planting of agricultural science and technology enterprises. In the past, mushroom picking was a time-consuming and laborious job in the mushroom room of the enterprise. On the one hand, the efficiency of manual picking edible fungi is limited, the labor cost is high, and the non-standard picking will also bring a certain degree of loss.
For efficiency and cost reasons, companies eventually turned to robots to pick them. The mushroom picking robot has a pair of "soft hands", which are bionic flexible suction cups with vacuum negative pressure. The suction cup is located under the robot arm. When the robot receives the instruction, it senses the location of the mushroom through the relevant system, and then gently shakes and sucks up the mushroom through the suction cup to avoid damage to the mushroom cap during the picking process. Such a robot can pick a single mushroom in just two seconds. It is not only handy, but also can move along the bed, and accurately determine whether the mushroom on the bed is mature, which greatly solves the problem of the company's many workers and recruitment difficulties.
Not only mushroom companies, but also a tea garden in Hangzhou, Zhejiang Province, has used picking robots. The tea-picking robot, developed by the Agricultural Robot and Equipment innovation team at Zhejiang University of Science and Technology, has a pair of "scissorhands". There is a small scissors at the end of the robotic arm of the tea picking robot, plus a negative pressure straw. When the robot locates the petiole, it can separate the buds and leaves from the branches through the scissors, and then the negative pressure straw can quickly suck the buds and leaves into the tea basket.
In fact, there are many different kinds of picking robots in China, whether it is melons, tea buds, or soft crops such as mushrooms... Are not difficult to pick a pair of robots "skilled hands".
In recent years, agricultural robots such as picking robots have been more and more applied in the domestic agricultural production process, and according to different agricultural needs, the functions of agricultural robots are not the same: fertilization robots, feeding robots, picking robots... There are many kinds. Behind this is the demand of the market on the one hand, and the progress of technology on the other hand. From the performance point of view of the robot, from the beginning can only be competent for a single agricultural operation, to now can complete more complex, more delicate farm work, with the development of technology, agricultural robots are more powerful, more flexible action, so there is a wider range of application scenarios.
At a grain processing enterprise in Yixing, Zhejiang province, palletizing robots have greatly improved the efficiency of the workshop. "Palletizing robots can easily palletizer according to pre-programmed procedures, assist workers to complete repeated and heavy work, and automatically carry out steps such as sorting, stacking, stacking, lifting, stacking, and so on." When the size and volume of the rice and the shape of the tray change, it only needs to be slightly modified on the touch screen." Lu Fangfang, head of the company, said, "After the introduction of palletizing robots, the palletizing speed can be reached 10-15 times per minute, reducing the damage and confusion of goods in the process of operation."
Compared with foreign countries, China's research and development in the field of agricultural robots started late, but the development momentum is very rapid. In the view of Wu Haihua, Minister of science and Technology development of the Chinese Academy of Agricultural Mechanization Science, in the future, the three major scenes of field planting, fruits and vegetables, and breeding should be developed from mechanization to intelligence. In field farming, increasing the level of digital production can reduce the rate of food loss and achieve significant efficiency gains. In terms of fruit and vegetable harvesting, due to the difficulty of traditional agricultural machinery and equipment to meet the requirements of selective harvesting, the current fresh food picking process basically relies on manual labor, and robotization is an important way to reduce costs. And in farming, robots can reduce bacterial infections in manual operations.
Jiang Wenlai, a researcher at the Institute of Agricultural Resources and Agricultural Regionalization of the Chinese Academy of Agricultural Sciences, said that in the future of smart agriculture, agricultural robots will play an increasingly important role, and agricultural robots in China have a wide range of demands and a huge market, with great potential for development.
Agricultural environment is characterized by non-standardization and uncertainty
Agricultural robots are more difficult and have a higher level of research and development than industrial robots
Compared with industrial robots, the technical difficulty of agricultural robots is greater, and the level of research and development is higher. Sun Tan, vice president of the Chinese Academy of Agricultural Sciences, said that many people think that the experience of industrial robots can be directly transferred to the agricultural field, but in fact, it is not possible. Because the factory floor where industrial robots work is a standardized and structured closed indoor scene, various environmental elements are determined and controllable. But the agricultural environment is arguably one of the most complex, unstructured and unstandardized scenarios with uncertainty.
Jiang Wenlai said that compared with industrial robots, agricultural robots have many difficulties. First of all, the scene is more complex, such as uneven fields, lighting, dry and wet degree differences, etc., to deal with this complex scene requires more new technologies and their mutual cooperation. In industrial scenarios, robots face relatively stable work sites. In addition, the work of agricultural robots varies greatly. Taking the apple picking robot as an example, there are differences in the environment where apples grow, such as the shape of the apple tree, the size of the apple, the position of the apple, some may be hidden by leaves, some bad fruits, etc. The uncertainty of these factors requires more complex technical procedures to control to complete the picking.
The difficulty of designing agricultural intelligent robots is higher than that of industrial robots. Therefore, the successful application of any kind of agricultural robot is inseparable from a large number of scientific research and innovation and strong technical support. In recent years, the number of invention patent applications in the field of agricultural robots in China has been rising, which provides a guarantee for the application of more agricultural robot achievements from the level of technological innovation.
This spring, the fifth-generation intelligent tea picking robot, developed by the Agricultural Robot and Equipment innovation team of Zhejiang University of Science and Technology, successfully completed the work of picking spring tea in tea gardens in Hangzhou, Zhejiang province. Facing the West Lake Longjing bud leaf, which is only 2 centimeters long and a few millimeters thick, the robot's recognition accuracy rate can reach 86 percent, and it takes only about 1.5 seconds to pick a bud leaf.
The tea leaves faced by the tea picking robot, young leaves and old leaves from the appearance of the difference is very small, the machine is difficult to distinguish. Also, each leaf has a different shape. In order to accurately identify young leaves and pick them off among many leaves, high visual recognition technology is required.
The fifth generation intelligent tea picking robot can achieve such accurate recognition, on the one hand, thanks to the robot's use of binocular cameras, which can realize 3D positioning through scanning to determine the location of buds and leaves. In addition, the robot is equipped with a set of recognition models called deep convolutional neural networks. In the early stage, technicians will input a large number of tea tree bud and leaf photos into the tea picking robot system, so that the robot can remember the shape and texture of the bud and leaf through learning and induction of a large number of image data.
Zhao Chunjiang said that through the development of technology, agricultural robots can now build circles on the track of farmland operations, identify and avoid obstacles, and can also carry out multi-machine communication and decision-making information communication. In addition, the agricultural robot can also carry out target identification, production prediction task decision-making, fault diagnosis. With the advancement of technology, agricultural robots will certainly have more powerful performance and can be applied in more complex agricultural scenarios.
Emerging technologies such as 5G promote the iterative update of agricultural robots
The new generation of robots can accurately sense, make autonomous decisions, and execute automatically
The change from the general automation equipment to the highly intelligent robot direction is an important trend in the development of agricultural robots in our country. Dr. Lin Yingzhi of the Institute of Digital Agriculture of the Fujian Academy of Agricultural Sciences said that compared with the relatively simple automated machinery of the past robots, intelligent agricultural robots can achieve accurate perception, autonomous decision-making, intelligent control and automatic execution, and with the continuous upgrading of performance, agricultural robots will have more functions.
In Sun Tan's view, the difference between agricultural intelligent robots and intelligent agricultural machinery equipment is that agricultural machinery equipment solves the "consistency problem", such as agricultural machinery equipment will not consider the individual differences of each wheat plant when harvesting wheat, and agricultural robots can solve the problem of "difference and individual". In other words, intelligent agricultural robots can cope with more complex operating objects.
In recent years, the rapid development of technology in related fields in China, especially the progress of emerging technologies such as 5G, the Internet of Things, and artificial intelligence, has promoted the iterative update and intelligent upgrade of agricultural robots.
In recent years, the Institute of Digital Agriculture of Fujian Academy of Agricultural Sciences has combined 5G and artificial intelligence technology, developed and launched the first 5G artificial intelligence agricultural robot "Xiaorui" in Fujian province, and has been iteratively upgraded. Compared with the traditional intelligent agricultural machinery equipment, even without considering the performance, from the appearance alone, "small Rui" is also closer to "people". It has the "ears" of two 10-megapixel cameras, the "eyes" of a 5-megapixel fisheye camera, carbon dioxide, photosynthetic radiation, wind speed and wind sensors on the head, and temperature and humidity sensors on the body, which can sense and collect ambient data in real time. Through the wheeled chassis walking, turning, "he" can smoothly along the cultivation tank automatic inspection, fixed point collection, automatic turn, can autonomously avoid obstacles around, but also can automatically return to charge.
In addition to human-like features, the most powerful place of "Xiaorui" is its intelligent "brain" - edge computing host and artificial intelligence recognition algorithm let him have a certain analysis and processing and judgment ability, and 5G communication technology solves the problem of data transmission delay, and improves the "Xiaorui" rapid response ability to external instructions and conditions.
Referring to the role of emerging technologies such as 5G on agricultural robots, Lin Yingzhi said: "In 5G agricultural robots, 5G technology is responsible for establishing communication connections between robots and cloud platforms. In the past, limited by the robot's own computing power, we hope to send the data collected by the robot back to the server of the cloud platform for calculation, and then return the results to the robot. This mode requires high communication requirements between the robot and the cloud platform, with a large amount of data, high reliability and low delay. "Because of the 5G boom, we set up 5G base stations at the test site to make the idea possible."
In addition to the "faster response" of 5G agricultural robots, with the substantial improvement of AI computing power and the iterative progress of artificial neural network models, the intelligence level of agricultural robots has also been rapidly improved. Lin Yingzhi introduced: "Before the outbreak of artificial intelligence, agricultural robots were difficult to land. For example, in the past, the inspection equipment running in the greenhouse required a lot of manual intervention, to ensure that the inspection road is smooth, lay guide rails, set road signs, and let the inspection robot walk in accordance with the designated route. If road signs are lost or guideways are damaged, the system breaks down and maintenance is heavy. The inspection robot with the support of artificial intelligence can independently plan the inspection route and adjust it according to the actual situation."
"Take the automatic rice transplanter. At first, two people were needed, one to drive and the other to plant. Later through mechanical transformation, seedlings can be automatically on the plate, but still need people to drive, and later with high-precision positioning technology, coupled with map control, the transplanter can run automatically, can be called the first generation of agricultural robots. However, this kind of agricultural robot needs GPS with high-precision maps, and the information such as pole and mound identification that may exist in the field needs to be marked on the map by the user in advance, and the procedure is complicated. In recent years, machine vision supported by artificial intelligence technology can very reliably identify fields, poles, trees, mounds and other information, can build their own maps, no longer need to manually annotate information, which is the current stage of intelligent agricultural robots." Lin Yingzhi said.
Lin Yingzhi said that the current development of agricultural robots is developing towards hand-eye collaboration, man-machine collaboration, and multi-machine collaboration, and the research field has expanded from bulk agricultural products to characteristic crops, from the intelligent upgrade stage of traditional agricultural machinery to the stage of new operation actuators, and gradually realize the full automation of production links. These trends will promote the rapid development of agricultural robots and provide new impetus for the development of the agricultural industry.
Source: Central Commission for Discipline Inspection and State Supervisory Commission Website Author: Wang Yajing