Posted in | News | Agricultural Robotics

Robots Trained Using Silicone Raspberry to Harvest Delicate Fruit

Raspberries are a popular summer fruit with striking scarlet color and unique structure. They consist of several fleshy drupelets with a combination of sweet and mildly acidic pulp.

Robots Trained Using Silicone Raspberry to Harvest Delicate Fruit.
Kai Junge, Ph.D. student and first author of the paper that will be presented at the RoboSoft 2022 conference. Image Credit: © Anne-Muriel Brouet/ EPFL

Their delicate structure also tends to be their main weakness, as it leaves them susceptible to even the slightest bruise or scratch. Farmers are well aware that raspberries are a tough fruit to harvest, which is reflected in their market price.

But what if robots fitted with modern sensors and actuators could offer some assistance? Engineers from the Computational Robot Design and Fabrication (CREATE) laboratory at EPFL aim to find a solution.

Annually, farmers lose millions of dollars worth of yield due to exorbitant labor costs and scarcity of workers. The issue is even more severe with regard to delicate crops like raspberries. For the time being, there is no feasible substitute for harvesting the fruit manually.

It’s an exciting dilemma for us as robotics engineers. The raspberry harvesting season is so short, and the fruit is so valuable, that wasting them simply isn’t an option. What’s more, the cost and logistical challenges of testing different options out in the field are prohibitive. That’s why we decided to run our tests in the lab and develop a replica raspberry for training harvesting robots.

Josie Hughes, Professor, CREATE, EPFL

A Training Tool for Robots

For the inexperienced, harvesting raspberries by hand is not an easy task. The fruit has to be supported from below, gently grasped between the thumb and the index finger, and then pulled with care until it separates from the receptacle — the part of the fruit that stays fixed to the plant — and drops into the picker’s hand.

To help harvesting robots become familiar with this job, the CREATE team engineered and constructed a silicone raspberry that can “tell” the robot the amount of pressure being applied, both while the berry is still fixed to the receptacle and after it is been detached.

The properties of the silicone raspberry can be modified to mimic the fruit’s resistance and necessitate the robot to exert an adequate picking force. As a result of this feedback, robots can be trained to pluck the fruit without causing any damage to it.

Our sensorized raspberry, coupled with a machine learning program, can teach a robot to apply just the right amount of force. The hardest part is training the robot to loosen its grip once the raspberry detaches from the receptacle so that the fruit doesn’t get squashed. That’s hard to achieve with conventional robots.

Kai Junge, PhD Student and Study First Author, CREATE, EPFL

Under its extraordinarily uniform shape and somewhat translucent pink surface, CREATE’s imitation raspberry is a remarkable engineering accomplishment. Its tissue is composed of silicone and its receptacle was crafted using three-dimensional (3D)-printed plastic. It also comprises a fluidic sensor that has a soft silicone tube to assess the compression force put by the robot. The pulling force that grips the fruit and receptacle together is produced by two magnets.

For the moment, the EPFL harvesting robot is slightly more than a gripper having two 3D-printed fingers enclosed with a thin layer of silicone and connected to a robotic arm. Nonetheless, the engineers ended up sacrificing more than a dozen raspberries to standardize their gripper in the lab. They then performed a range of tests, first picking the imitation raspberry by hand and then employing the robotic platform.

While the CREATE team has shown the proof of concept for their innovation, the technology itself still needs a lot more fine-tuning.

It’s incredibly challenging. So far we’ve been using a very simple feedback system in our robot. The next step will be to design and build more complex controllers so that robots can pick raspberries on a larger scale without crushing them.

Josie Hughes, Professor, CREATE, EPFL

The EPFL researchers are presently designing a camera system that will enable robots to not just “feel” raspberries, but also “see” where they are hanging and whether they are set to be plucked. This summer, they propose to use their harvesting robot in the field, at the peak of the native raspberry season.

This kind of system could be used to pick other berries too for example. We’d also like to develop technology for other soft fruit and apply this physical-twin concept to more complicated tasks like other berries, tomatoes, apricots, or grapes.

Josie Hughes, Professor, CREATE, EPFL

Soft sensorized physical twin for harvesting raspberries 🍓

Soft sensorized physical twin for harvesting raspberries.Video Credit: EPFL


Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.