Growth of Artificial Intelligence in Agriculture
With the growth of artificial intelligence in everyday life, even traditional agriculture does not stand aside.
One notable development in this field is “Ecorobotix”, a two-meter robot equipped with GPS support and working on solar batteries. It is capable of efficiently eliminating weeds with an accuracy of 95%, without generating any waste. Similarly, the Energid and Universal Robots systems employ robots with multiple cameras and flexible hands for harvesting citrus fruits.
Another significant advancement is the “River Lettucebot”, which scans the crop’s geometry to optimize its growth and minimize the use of pesticides. Additionally, “PrecisionHawk” proposes the use of drones for remote monitoring and analytics.
However, among all these remarkable engineering marvels, there is one development that stands out from the rest. Scientists from the universities of Edinburgh and Sheffield are exploring visual navigation in dense vegetation, taking inspiration from ants. In their study, published in Science Robotics, the researchers emphasize the need for “low-power, but effective on-board decisions” in the field of robotic navigation.
“We drew inspiration from insects, such as ants that are able to learn and follow routes in difficult natural conditions, using relatively limited sensory and neural systems,” the scientists stated.
“In our study, we present an example of such an approach by implementing a network for memorizing visual routes on neuromorphic equipment, which is directly based on the latest achievements of insect neurons,” added the researchers.
As a result, the scientists successfully developed an artificial neural network that enables robots to navigate complex routes in densely vegetated areas, based on image analysis and a form of route memorization, similar to insects’ behavior.
The researchers conducted tests on complex routes in fields that were uneven, dirty, and overgrown with vegetation, and achieved positive results. They believe that their study highlights the potential of such systems in agriculture, forestry, and environmental monitoring.