Agricultural robots are an advanced form of agricultural machinery equipment, and it takes a considerable amount of time to enter ordinary farmers. Smart agriculture is not a game between humans and machines, but rather an empowerment between them. Traditional agricultural machinery has transformed farmers from "manual laborers" to "machine operators", while smart agricultural robots will upgrade farmers to "robot managers" and "data decision makers". Yushu Technology recently announced the application of its quadruped robot Go2 in the field of smart agriculture, exploring new paths for robot empowerment in agriculture; The first winter jujube picking robot in China, Lanxia, developed by Shenzhen Lanxia Robot Co., Ltd. and Harbin Institute of Technology (Shenzhen), was recently put into use in Dali, Shaanxi; The Institute of Genetics and Developmental Biology, Chinese Academy of Sciences has successfully developed the world's first intelligent breeding robot "Jier" that can automatically cruise cross pollination; The tomato precision operation intelligent agricultural robot developed by the Robotics and Autonomous Unmanned Systems Laboratory of Fudan University has made its debut... More and more subjects are running on the hot track of agricultural robots. In the context of aging rural labor force and high labor costs, agriculture has a deep desire for intelligence. Unlike traditional agricultural robots, intelligent agricultural robots can perceive environmental data in real-time and achieve autonomous decision-making and action through image recognition and deep learning. For example, in the past, people needed to "mark" the map and manually plan the route. Now, with the help of Beidou+5G technology and intelligent algorithms, it is possible to independently plan the path and automatically avoid obstacles. With the breakthrough of AI technology, agricultural robots can now carry out tasks such as inspection, fertilization, spraying, and picking. Taking Yushu Technology as an example, it is equipped with agricultural specific sensors and AI vision systems, which can be used to monitor seedling growth and collect real-time field data. Neither Yushu Technology, the Chinese Academy of Sciences, Harbin Institute of Technology, nor Fudan University are traditional agricultural research institutions or agricultural technology enterprises. The reason for joining this track is to have a positive outlook on the technology and application prospects of agricultural robots. The reason why agricultural robots are capable is not only due to the rapid development of general technologies such as machine vision, knowledge graphs, and decision models, but also because agricultural information acquisition technologies such as satellite remote sensing, low altitude technology, ground Internet of Things, and agricultural sensors are becoming increasingly mature. With the large-scale development of agriculture, there have been significant improvements in agricultural infrastructure and management systems, and new things have found their place. This has led many business or research entities to focus their innovation goals on the blue ocean of agriculture. In fact, agricultural robots are more difficult to develop and promote than industrial robots. Analyzing the reasons, firstly, agriculture is a non standardized environment where the objects are living organisms, and there are no two identical leaves in the world, requiring experts in fields such as optics, sensing, mechanics, and algorithms to collaborate in solving problems. Secondly, agriculture has relatively low efficiency, and the technology may not necessarily be high, but the cost should be as low as possible. Without efficiency, farmers will not buy it. Thirdly, some technological products sound plump and bone like when used. Taking picking as an example, the staff admitted that the most difficult thing is to distinguish mature fruits and avoid branches and leaves to achieve non-destructive picking. These factors combined have resulted in agricultural robots still being limited to "bonsai" and only being applied in a very small number of business entities and scenarios. The charm of technology lies in its powerful iterative evolution ability, and in the increasingly down-to-earth situation, the cost is expected to become lower and lower. The highlight of Yushu Technology Go2 Agricultural Edition is the ability of "AI vision+edge computing". The basic price is 1600 dollars, which is cost-effective for high-value agriculture. The second generation winter jujube picking robot prototype has entered the research and development stage, focusing on solving problems such as decreased recognition and grasping efficiency under high temperature conditions. The third generation also has a research and development schedule. This inspires the industry to start from reality, constantly iterate and optimize, especially in finding breakthroughs in reducing costs and increasing benefits, creating diverse solutions, and increasing attractiveness. Developing agricultural robots requires historical patience. We often say that agricultural machinery and equipment liberate farmers from physical labor. But even today, in many fields and links of agriculture, agricultural machinery and equipment still face difficulties in landing and usability. Agricultural robots are an advanced form of agricultural machinery equipment, and it takes a considerable amount of time to enter ordinary farmers. It should be noted that smart agriculture is not a game between humans and machines, but rather an empowerment between them. Traditional agricultural machinery has transformed farmers from "manual laborers" to "machine operators", while smart agricultural robots will upgrade farmers to "robot managers" and "data decision makers". According to the plan, by the middle of this century, a strong agricultural country will be fully established. Will agricultural robots at that time be as popular as tractors and harvesters today? Worth looking forward to. (New Society)