A realistic approach to robotic automation

12-axis robotic arms, drones, and mobile robots – these are just a few examples of ways startups are attempting to automate the farming industry. It is possible to verify the concept in a small lab setting, however, these concepts have yet to make an impact on the way food is grown. Why?

Complexity and cost.

Unlike industrial manufacturing, where tasks are consistent and repetitive, nature provides us with a new and unique challenge every single day, and even with every crop. No matter how small or delicate a component is, a robotic arm can successfully automate tasks using it, as it is always the same exact component, and the conditions also stay identical. Drones will automate the pickup and drop-off of similar packages, for example. Now, try to develop a drone that can decide autonomously how to hold and secure different products – the complexity shifts to a whole different scale.

Farming isn’t a predictable clean-room. Every root system will look different, and every crop will need slightly different pruning. At Katif, we’re already demonstrating how software can solve these unique challenges and take on the real-world mechanical tasks required. However, to successfully and reliably perform these tasks robotically, we have to use the simplest, most robust, robots. If we just throw mechanical innovation at the problem, we will quickly find that pricing becomes completely irrelevant for the industry.

The profit margins for a Tesla plant or semiconductor facility are incomparable to those of food farms. The initial investment can never be recuperated, even with the OpEx optimization that robots promise. Some companies, vertical farm systems, in particular, are adding tech and sensors – while maintaining a large and more expensive workforce that needs to work with these systems.

All this is to say – the challenge in farming is finding a cheap and reliable robotic solution, as well as keep the CapEx as low as possible, and continuously work towards optimizing the OpEx.

CNC (Computer Numerically Controlled) machines – simple, reliable, and proven.

These machines, built as early as 1940, support all types of robotic automation – although under the main constraint that the workspace is fixed and structured. The smartphone you’re reading this on was likely crafted by such a machine. 3D printers are another popular example. As long as you don’t expect these machines to fly or drive around – they are a simple solution for automation. Motors move the “head” throughout the X, Y, and Z-axis, and some machines also control the rotation and angle of attachment. These heads change from task to task, from nozzles to inject molten plastic into 3D printers, to spinning drill bits for aluminum or wood. There are thousands of manufacturing options for these machines, all offering a wide variety of features and price points. Compared to robotic arms or other complex systems – CNC machines are a bargain.

At Katif, we’re automating farming on top of these machines. We’re designing and developing machine heads for many different farming tasks and showing that automation is possible, on the cheap.

The main drawback – automation within a small, fixed space is challenging. Especially when you consider just how much growing space you need to support the idea of feeding the world. We’ve solved this with what we’re calling “Greenhouse Rovers” – if the robot won’t come to the plant, bring the plant to the robot!

Leave a Comment

Your email address will not be published. Required fields are marked *