Carnegie Mellon University

Specialty Crop Automation

The Specialty Crop Automation project succeeded in bringing precision agriculture and autonomous equipment to citrus growers. The project used autonomous tractor development work by NREC and John Deere as the foundation for developing computer-controlled, tree-level precision agricultural applications.

Citrus grove workers currently operate tractors, sprayers, and mowers, detect crop disease visually and predict yield by sampling. These processes are costly and result in inefficient spraying and disease detection and in inaccurate yield estimates. The Integrated Automation for SustainableSpecialty Crops Farming project integrates a suite of precise, cost-cutting technologies around the sensing, localization, computing and data collection capabilities of autonomous tractors.

Orchard Automation  
NREC and John Deere teamed to build unmanned tractors to perform grove operations such as spraying and mowing. These tractors operate autonomously while being monitored remotely by grove personnel. They also serve as platforms for the precision agriculture technologies we are developing. 
 
GPS-Free Navigation 
NREC developed a GPS-Free Navigation system to enable autonomous operations when GPS is lost due to overhanging trees, tall structures or for other reasons. The system uses visual odometry and an inertial measurement unit (IMU) and can operate as a stand-alone system and as a supplement to GPS.  
 
Disease Detection  
NREC and the University of Florida developed automated systems that use remote sensing to detect citrus diseases. These systems make use of the localization abilities of the autonomous tractors to pinpoint the location of diseased trees, allowing targeted scouting and treatment of disease outbreaks. 
 
Yield Estimation  
NREC and the University of Florida developed improved yield estimation methods utilizing the perception sensing and localization capabilities of autonomous tractors to model tree canopies. Used in conjunction with historical yield data, these measurements can provide more accurate yield estimates.  
 
Precision Spraying  
NREC, Cornell University and the University of Florida developed tree-level precision spraying. The effort makes use of the sensing, data collection and localization capabilities of the autonomous tractors to enable each sprayer nozzle to spray only when directed towards foliage, and not air, swale banks or weeds, thereby reducing chemical usage.   
 
Precision Ag-in-a-Box  
NREC is packaging the core components of these precision agriculture systems in a single, bolt-on unit that can be used on any type of farm tractor. This allows growers to use the precision spraying, yield estimation and disease detection applications on their existing tractor fleet. 
 
Mechanical Harvesting  
The project is analyzing mechanical harvester designs that use advanced sensing and manipulation technologies to harvest fruit economically and avoid the canopy damage that's a side effect of the current generation of harvesters.  

 

Photos

Specialty crop automation at the National Robotics Engineering Center.
Specialty crop automation at the National Robotics Engineering Center.
Specialty crop automation at the National Robotics Engineering Center.