PI: Greg Lowry
Co-PI(s): George Kantor
University: Carnegie Mellon University
Agriculture is currently inefficient, unsustainable, and unable to meet future demands. A rapidly changing climate exacerbates this problem as hotter, wetter climates lead to greater crop losses from plant diseases. Sensor technology, robotics, and artificial intelligence have the potential to vastly reduce crop losses by providing autonomous early detection and treatment of diseased plants. This would significantly lower crop losses, and would lower pesticide use rates, worker exposure to pesticides, and labor costs. The proposed PITA project converges expertise in robotics, computer science, and environmental engineering at CMU, with expertise in agriculture and sensing at two PA agriculture companies to develop novel senor technologies and deep learning neural networks needed to revolutionize pest management and lower the environmental impacts of agriculture.
The specific goals of this PITA are to:
- determine which multispectral detectors can best identify affected plants before they become symptomatic
- determine if combinations of chemical and spectral sensors can detect diseases earlier than spectral sensors alone
- collect the preliminary proof of concept data needed to win a USDA SCRI and NSF Cyber-Physical Systems grant to ultimately develop fully automated robotic platforms to detect and treat crop disease before it manifests.