Data powers Israel’s ag robotics industry
The first phase of developing new farming technologies begins with information gathering. For artificial intelligence and machine learning systems to work, data is essential. Agricultural data from field plots, greenhouses and other growing systems are being used to help detect diseases and plant stress, as well as determining yield, nutrient levels and the appropriate variable-rate systems for irrigation and fertigation.
Even a task as seemingly simple as picking a ripe apple requires data. The information will power the sensing technologies that allow a robot to do everything from understand the apple’s location and exact position on the tree to determine whether the apple is ripe and ready to be picked. From there, the machine must know how to reach, grasp and pick the apple, completing the task without damaging the fruit.
“It takes an enormous amount of data to perform these tasks,” Bechar explains. “So nowadays, the focus in Israel is on these developments. We are creating these support systems to enable the robot system to function, but we’re thinking about all the data we collect from these systems, too.”
“Sometimes we don’t use the information after we pick the apple,” he continues. “This data, however, can bring additional knowledge to the farmer. The question we’re asking is, ‘how can this data be used to alter or improve farming techniques?’ We have all this data, but how can we expect people to go about using it?”
With the development of advancing technologies like artificial intelligence and machine learning, the collected data can be better analyzed to provide detailed insights. Bechar uses the example of a farm that has one field producing 500 tons of tomatoes, while another at the same farm only produces 200 tons. A smart system will be able to inform the farmer of hot spots or other problems that are impacting yield.
“Once we have this kind of information, we will be able to ask the questions that we are not asking now because we don't have the available data,” Bechar says. “Now, we don't know the real distribution in the field. When we can analyze this data all together, we will be able to bring the agriculture industry to a higher level. The farmer will be able to make decisions on the spot because the relevant data will be available.”
Ag robotics research projects in Israel today
Israel is experiencing an agricultural technology revolution. Although there are many problems to tackle and much progress to be made, the Institute of Agriculture Engineering has a number of exciting robotics projects in the pipeline today. This institute is one of six that operate under the Agricultural Research Organization, Volcani Institute, a governmental research entity. Together with technology companies, the Volcani Center works to develop robotics systems for the farm.
At the Institute of Agriculture Engineering, Bechar’s job is to help figure out how robots will be able to operate in farm settings. The institute selects its robotics projects based on five criteria. The project must address a real problem in agriculture, and the solution must be high impact, making a difference on farms with a lot of hectares. Robots must only be introduced if they provide a clear added value in comparison to other solutions. If there is a simpler solution to an ag-related problem, that should be used instead. Humans must be integrated into the system intelligently, and there must be modifications for the agricultural domain. Most of the current projects combine autonomous systems with human labor, as a fully automated system with a human supervisor is at least a decade away.
Here are some projects the institute is working on right now:
- A commercial sprayer that has been modified to be autonomous. This innovation is enabling workers to manage the system from afar, without being exposed to harmful chemicals.
- A robot that prunes specific branches from deciduous trees. After a worker marks the branches that need to be removed from the tree, the robot will go through the groves and remove the marked branches.
- A sonar robot that estimates yield and plant health status. This innovation can accurately predict the amount of food on the plant and the overall yield. It can scan the farm’s plots to collect data and create images. From this data, the robot can determine things like fruit mass, fruit number, leaf number, etc. This project is with Prof. Yossi Yovel from Tel Aviv University.
- A group of projects center on disease monitoring and proximal sensing. Robots, cameras and sensors are combined to determine disease and plant stress problems in their early stages. These solutions can be used in the greenhouse, packing house and field. They are also used to help farmers spray their crops more selectively and as needed. In one case, a robotic platform was combined with the manipulator to scout the rows for early disease prediction.
- A selective weeding robot that can navigate tricky crops, such as watermelon where the crop and weeds grow together on the ground. The institute also developed an electrostatic machine that went around to identify and electrify the weeds that were present in the field.
With each of these projects, Bechar says, the institute is working to collect more data than it can use right away. The data then inform the next generation of projects and solutions. Some of these current prototypes are already being developed into functional machines by technology companies that are satisfied with the proof of concept that the institute has been able to provide. Bechar believes this is just the beginning of the country’s future in robotics.
“In the next two to three years in Israel, we will start seeing some solutions in the marketplace,” Bechar says. “These solutions will be specific to a group or a certain task. But I believe that with the progress that we were experiencing now in research and in the advances in robotics and AI, the agriculture industry will be completely different in 15 to 20 years.”