Autonomous Robot Monitors Hard-to-Reach Crops
Purdue University researchers have created a small autonomous robot to help farmers monitor crops and regularly collect physical samples — saving time and effort. Watch this video to learn more.
“IoT4Ag’s mission is to utilize internet-of-things technologies to digitize the agriculture industry in order to meet the growing demand for food, energy, and water security as the population grows,” said Aarya Deb , PhD student in mechanical engineering at Purdue.
Transcript
00:00:00 -Our project is to navigate inside corn rows autonomously, and with the robotic arm sample the corn leaves to monitor nitrogen levels. -The farmer would usually have to get samples on his own. But we're trying to automate that using the robot. Our robot is pretty small, so it doesn't compact the soil like a tractor would. So it can fit inside rows of corn plant which has like 10-12 inches of spacing, where a human it might be difficult to go in there and do the daily tasks. -GPS signals are really unreliable under the canopies. So we are trying to use just the LIDAR to navigate in the rows, understand the robot's location exactly without any GPS assistance. Designing our custom end-effectors that can chop the leaf or cut it from the plant. I'm working on a deep learning based physical sampling module that can use the features of neural networks and all this data we have collected over the last two or three summer seasons to
00:01:05 train the neural network and identify a leaf. And we want to get all the poses of the leaf, like the angles of the leaf, the exact position in XYZ coordinates. Because as you can see behind us, the leaves are not really straightforward, they're poking out from every different direction. So we want to capture them in a 3D space and tell the robot exactly how it is laid out in front of it, so the arm can grasp it and sample it. So the robot can traverse acres and acres, get all those metrics that the farmer needs to make informed decisions.