Announcing Traptic - Strawberry Picking Robots

Centuries ago, we grew our food by hand, with farmers limited to manual tools and limited scale of production. And you could measure the world’s population in millions, not billions.

Today we've totally mechanized the production of field crops. That means our nutritional staples - wheat, corn, soybeans, etc - are grown and harvested using giant machines. These harvest quickly, efficiently, and at enormous scale. But this technology revolution hasn't made it to most fruits and vegetables due to the delicacy required to harvest them. As a result these foods are still harvested the old fashioned way - by hand.

Aside from the obvious cost and efficiency problems of harvesting manually, the supply of workers willing to work in the fields has become less predictable. Tightening immigration policy and strengthening economies in Mexico and other countries exacerbate this issue. Even a rise in wages has often been insufficient to generate the labor required.

With strawberries in particular, growers lose 20% of their crop each year because they cannot find enough people to harvest it. That means berries literally being left to rot in the fields due to lack of labor.

At the same time, the world's growing population and expanding middle class demand more fruits and vegetables than ever before. We already grow over $200 billion dollars of produce annually, worldwide. We’ll soon number 10 billion humans on Earth (7.7 billion today!) even though overall birth rates are declining. But this “slower growth” presents a real risk to global agriculture because the world's aging population is even less able to work in the fields. This leads to increasing demand for fruits and vegetables yet fewer people available to produce them. And our growers are unable to find enough labor to fill even the existing demand, while demographics conspire against them.

Quite simply, this labor crisis threatens the viability of the world's food production system.

That's what motivated us to start Traptic.

We set out to develop a smart robotic harvesting machine to relieve the world's food production system from this crippling labor crisis. So, how do you build a strawberry picking robot?

final traptic ceres non blurry.jpg

First and foremost we had to solve the grasping problem. Strawberries are among the most delicate fruits, and they require great care when touching or handling them. They grow in a crowded environment, with other berries, stems, and leaves surrounding them. On top of that, a commercial farming environment has dust, dirt, varied temperature, wind, and other factors which can change day-to-day. Finally, separating the berry from the stem requires applying great force relative to the size and weight of the berry itself, and we must be able to apply that force in a certain direction on the berry without damaging the delicate skin. Other fruits and vegetables have similar challenges so we knew that if we could pick berries, we could pick other produce as well.

With significant hours of engineering time, our team developed a unique gripper which picks these berries. The Traptic gripper grasps strawberries firmly enough to extract them from the plant, yet delicately enough to preserve the quality.

We then built a robotic harvesting platform around this gripper. Technological advances in the last decade have enabled new capabilities for sensing, understanding, and interacting with the world. We rely on these advances, and in particular we incorporate 3D cameras, neural networks, and low cost actuators into our system.

final gripper non blurry.JPG

These enabled us to build our vision system. This understands the plant structure and detects the ripe strawberries. It can distinguish between ripe and unripe berries based just on a tiny smidge of white on part of the berry. And it determines the berries' positions with millimeter accuracy. With this vision system, and the other components we use, we have the capability to act quickly and precisely in a tough environment, and we will be able to do so reliably for years on end.

During the course of 2016 and 2017 we built 3 machines of increasing size and capability. In 2017 we demonstrated that we could pick a strawberry successfully in a commercial farm here in California. Out of the lab and into the fields!

Then, in 2018, we built a machine which can work in the field for years on end. That machine, Ceres, is the product of tens of thousands of hours of engineering work by our multidisciplinary team of hardware and software engineers. It also has multiple robot arms, enabling it to harvest a larger volume of fruit than any of our previous machines.

Ceres entered continuous field testing early this year across both Northern and Southern California partners.

Our upcoming machines will help growers augment their existing workforce. This will allow them to expand their operations, reduce waste, and increase the quality of fruit they produce.

Since fruits and vegetables are harvested by hand these crops, including strawberries, peppers, oranges, and melons, all suffer from the same labor shortage. Traptic will deliver harvesting-as-a-service, helping the global food supply chain scale to the demands of a 10-billion person population, in a world where we can no longer rely on a limitless source of affordable human labor. For us, strawberries are just the beginning.

Looking back, the mechanical revolution made John Deere huge. The chemical and biochemical revolutions made Monsanto huge. The robotic revolution is next, but no existing companies can solve this. We believe we will.

final team photo 5.28.19.jpg
 
 
Lewis Anderson