Bloomfield Robotics is looking to build on over 15 years of research at the Carnegie Mellon University Robotics Institute.
Founded by Dr. George Kantor, Tim Mueller-Sim, Eric Felix, Tanvir Parhar, and Harjatin Singh, Bloomfield Robotics is bringing our combined 30+ years of experience in designing and deploying robots in agricultural settings to the commercial space. Below you'll find a few relevant research projects our team has been a part of over the years.
The Robotanist is a high-throughput crop phenotyping platform initially developed as part of the US Department of Energy’s Transportation Energy Resources from Renewable Agriculture (TERRA) program, which targeted sorghum specifically in an effort to accelerate the breeding process and increase the rate of genetic gain from season to season. The goal of the project was to provide plant breeders, cultivators, and scientists with the high-resolution crop data they need to breed better plants, faster.
The platform will continue to be used as a research tool in a variety of agriculture research projects at CMU.
At a base weight of 140 kg, the Robotanist is capable of carrying an additional 100 kg at up to 2 m/s for more than 8 hours between charges, thanks to the devotion of one-third of the base to a Li-Ion battery. With considerably lower ground pressure than the human foot and enough torque to drive straight up a wall, we've yet to get the robot stuck in the mud.
A suite of navigation and phenotyping sensors give the robot a chance at navigating within the world. The robot comes standard with an RTK GPS, and aerospace grade inertial measurement unit, multiple navigation cameras, a Time of Flight sensor, and multiple LiDAR units. The modularity of the platform allows the inclusion of more sensors as the environment changes.
Computer vision, neural networks, and high-value crops
We've gained extensive experience over the past 8 years deploying custom-built camera systems through all sorts of high-value crops and extracting meaningful information from the captured images, including health, quality, and other important traits.
Through partnerships with growers associations, world renowned universities, and industry experts we've developed a wide range of algorithms for crops such as wine and table grapes, apples, and grain crops such as sorghum.
We then pass this high-level information back to the scientists, breeders, and cultivators so they can perform labor loading estimation for the year, determine problem areas in their fields, and help them breed better plants faster.