We recently wrapped up a two-year project with John Deere focusing on their autonomous unload system. While much of it is still confidential, some has now been published under a provisional patent, and we wanted to share a bit of Vyn Lab’s role in the project.
This project has been a collaboration between Purdue Agronomy, Ag and Bio Engineering, Mechanical Engineering, Aerospace Engineering, and John Deere Global. Each department handled a different aspect of the project with Agronomy focusing on field setup and layout, hardware implementation, and testing and data collection. Ag and Bio Engineering headed up the hardware development and implementation and planned and executed the data collection and testing. Mechanical and Aerospace engineering developed extensive software and programming to execute main functions with the hardware setup. This collaboration helped us accomplish some impressive results in a short period of time. A lot of hard-working graduate students deserve credit for really pushing themselves and making this project a success.
So, what was this project we were working on? Over the course of two years, we were working with John Deere to develop their auto unload system for combine and grain cart. What does this mean? Let’s start with what currently happens in fields all across the country at harvest time. As a farmer harvests a field, the grain cart’s job is to relay harvested grain from the combine to the trucks at the edge of the field. The grain cart will approach the combine, and if the conditions are right, the combine operator will unload the combine onto the grain cart while still harvesting. Unloading on the go is the most efficient way to maximize the movement of grain away from the field, while still keeping harvest moving. The less time the combine is stopped the better. However, there are a whole host of logistical problems with unloading on the go. The combine operator and grain cart operator must match speed to prevent any spills. The grain cart driver needs to stay exactly parallel with the combine, and the combine operator must be able to vary their speed to fill the length of the grain cart, all while continuing to safely and accurately harvest the crop in front of them. Enter, our autonomous harvest project.
John Deere already has some tools in place that make this system easier for the combine and grain cart operators, mainly, machine sync. This tool allows the combine operator to set a “home point” for the grain cart to pair with. The grain cart operator simply approaches this home point, engages the system, and machine sync takes over and keeps the combine and grain cart in the set relative positions to each other. The combine operator also has the option to push a nudge forward or backward button, within machine sync to change the relative position and fill the length of the cart.
The autonomous project is taking the machine sync capabilities one step farther by freeing up the combine operator to focus on harvesting what is in front of them. An imaging system looks at the cart dimensions as well as the fill level in the cart and does any nudging that the combine operator was previously responsible for. In this system, the imaging system can decide to move the grain cart forward or backward based on the grain pile it sees. The combine operator still has some control over the fill by selecting how full they want the cart, and what the fill strategy should be (front to back, back to front, middle to front to back, etc.)
Implementing this kind of system reduces operator fatigue, reduces chances of operator spills, and speeds up unload time. All of these benefits can go a long way towards increasing harvest efficiency and ease.
Our role in this project was to plan and monitor the agronomic systems we would need in place to test the system. We implemented multiple hybrids of different maturities to allow for a wider harvest window to test at the critical time. Additionally, we designed and implemented a cover crop test strip on the side of the field to help reduce field compaction and preserve some soil structure from repeated test runs during the summer. Our research technician, Rachel was the grain cart operator for the project and helped with data collection, instrumentation of the equipment and grain handling. Additionally, Dr. Vyn procured all the equipment upgrades we needed for the duration of the study including a 900 bushel grain cart, a 12 row folding head, and a 9 series John Deere tractor.
Further information about this project will be published in the coming year, as well as available in filed patents. We are excited about the results of this project and its ability to reduce operator fatigue, improve harvest efficiency, and open the door for further automation of harvest systems. We hope to share more details on some of the specifics of testing in future blog posts.
Comments