We are surrounded by autonomous machines, with drones hovering overhead, self-driving cars hitting the road, and even street-smart robots that can deliver your groceries cropping up in major cities. But will there come a day when humans are no longer needed and forage harvesters operate autonomously?

That’s the goal, according to Matt Digman, an agricultural engineer with the University of Wisconsin-Madison. He said forage harvester automation has been around for a while, beginning with the addition of metal detectors on choppers in the 1990s. Since then, other features like crop flow and moisture sensors have become fixtures on forage harvesters. But even though manufacturers have taken great strides with these developments, creating a fully autonomous machine is another endeavor altogether.

Digman broke down what it will take to achieve full automation into eight categories: harvest timing modeling, machine setup and maintenance, navigation and guidance, throughput management, performance quality assessment, coordination with wagons and trucks, safety and compliance, and connectivity and data management. He explained how each technology is progressing at a different pace.

Monitoring plant moisture and milkline doesn’t always paint an accurate picture to predict the timing of silage harvest. Therefore, current harvest timing technology compiles weather modeling data, satellite vegetation indexes, and economic modeling data to produce a more precise prognosis. While we have the means to gather this information, the challenge lies in tying it all together.

Navigation and guidance technology is more advanced. During forage harvest, farmers can record windrow locations, save that data to the cloud, and then access those maps when they come back into the field with the chopper. This essentially allows for better synchronization of different field operations.

The progress of autonomous throughput management is more difficult to justify, especially since it can be measured several ways. Digman explained that while manufacturers boast harvest efficiency per unit of time, engineers assess throughput management based on horsepower hours per ton of forage processed — and that number has stayed relatively stable over the years. With that said, chopping more forage faster is still a valuable improvement in newer models, especially for crops like alfalfa that lose quality over wilting time.

Crop quality analysis has also seen significant progress; however, the effectiveness of this feature needs some help. Digman said choppers equipped with cameras that take pictures of harvested forage use artificial intelligence (AI) to analyze images and determine kernel processing score and actual length of cut. Ideally, farmers would use this information to optimize silage processing, but there is considerable noise in the lab reference data that AI uses to generate results, and that limits the utility of this technology.

Likewise, there are some hoops to jump through for autonomous software connectivity and data management. Digman said although the industry has made hefty investments in this area, there remains a gap in interoperability.

The state of autonomous machine setup and maintenance is a mixed bag. On one hand, some regular maintenance tasks, like lubrication and knife sharpening, have been successfully autonomized in choppers that are currently on the market. On the other hand, humans are still largely responsible for overall machine condition monitoring, and many maintenance tasks still need to be done by hand.

Chopper coordination with silage wagons and trucks is also evolving, but autonomizing over-the-road vehicles is a major roadblock. Digman can envision an intermediate solution in which autonomous choppers can stop and fill wagons and trucks, but those vehicles would need to be driven by humans.

There will arguably always be room for improvement when it comes to safety and compliance. Even though researchers can test machinery safety systems in controlled trials, more work needs to be done to understand how these systems will perform in real field environments. So, how close are we to a fully autonomous forage harvester? It’s safe to say we’ve got significant ground to cover. But assessing the position we are in reveals just how much progress has already been made