Soft Robots Get a Grip on Fruit Picking

Improvements in gripping technology are fueling the race for agricultural and logistics automation, as multiple companies and research teams develop sensors, software, and grasping solutions.

Cambridge Consultants has combined low-cost sensors, image-processing algorithms, commodity hardware, and soft grippers for flexible grasping of irregular objects. The technology could ease robotic sorting and picking on farms and in warehouses.

“Traditional robots struggle when it comes to adapting to deal with uncertainty,” said Chris Roberts, head of industrial robotics at the Cambridge, England-based company, which is part of Altran.

Among the challenges are recognizing which fruits or vegetables to pick and handling them without damaging them. “The system is capable of handling objects for which no detailed computer-aided design model exists — a necessary step to using a robot with natural objects which, although they share some characteristics, are not identical,” Roberts said.

Most industrial robots are designed to operate within narrow tolerances and on identical objects. Fruit growers have complained about labor shortages for human pickers, and they increasingly view robots as necessary for productivity and profitability.

“I would imagine this is the sort of thing where it can replace some of the really low-skill jobs … but there are still plenty of jobs that humans are just much better at,” Roberts said.

Cambridge Consultants combined its mechanical engineering, electronics, and programming expertise to develop “soft controls” enabling robots to pick and sort irregular organic items, such as weeds or fruits. In a demonstration using fruit randomly stacked in a bowl, “our robot [uses] machine vision and some smart software to identify which piece of fruit is on top,” Roberts said.

“It translates this information into real-world coordinates and positions the hand to pick the required fruit, whilst avoiding other objects,” he explained. “The custom-made hand adapts to the shape of the fruit and securely grips it without damaging it. Once picked, the fruit can also be sorted by color so that, for example, red apples can be separated from green apples.”

Cambridge Consultants has been seeking to double its staff to 100 employees in the Boston area and from 450 to 700 people worldwide.

RightHand raises money

RightHand Robotics Inc., which is developing warehouse automation for order fulfillment, has raised $3.32 million in venture capital. However, the market for manipulator robots is already crowded.

Cambridge, Mass.-based RightHand was spun out of the Harvard Biorobotics Laboratory and the Yale Grab Lab. It has demonstrated a three-fingered gripper, but it is not yet commercially available.

Co-founder Yaro Tenzer declined to say exactly how RightHand plans to spend its investment funds, but the company said it is working on picking systems for e-commerce. Other robotics makers in that space include Amazon Robotics (formerly Kiva Systems Inc.), Fetch Robotics, Harvest Automation Inc., and Rethink Robotics Inc.

Hitachi dives in with two hands

For instance, Hitachi Ltd. is developing a two-armed mobile robot for the logistics industry. The Japanese company’s software powers hardware made by Seiko Epson Corp. The autonomous robot is designed to be able to pick a variety of objects.

“Because its arms can hold objects of many different shapes and sizes, the robot could be used to replace human staff doing repetitive picking work in high-mix, low-volume warehouses, where it would not be cost-effective to use dedicated robots for each different product,” said the IDG News Service.

Hitachi’s two-armed robot could eventually be twice as fast as humans in certain logistics industry functions. The company is also working on the related autonomous controls and vision systems and said it expects the robots to be available in the next few years.

A packed field

WinterGreen Research predicts that the market for agricultural robots will grow from $817 million in 2013 to $16.3 billion by 2020, so it’s no surprise that many companies are working on farm automation.

Soft Robotics Inc., which won a Game Changer Award at last month’s RoboBusiness 2015 in San Jose, has released a soft robotic hand for fruit handling. RBR50 company Schunk recently released its EGL 70 mechatronic gripper for industrial use.

Scientists at the Massachusetts Institute of Technology are working on models that could enable Robotiq’s robotic hands to grasp objects more easily by adjusting position. They hope to make more affordable systems using “extrinsic dexterity.”

MIT is also using “deep learning” techniques, involving large data sets and as many as 700 repeated grasping attempts, to train a Rethink Baxter robot to be better at grasping. In addition, MIT is developing soft robotic grippers for fragile objects.

Quebec-based Robotiq has also released a new two-fingered gripper for material handling and logistics functions. It’s already in use at a wide range of applications, including handling circuit boards at an automotive electronics manufacturer, unpacking bottles for a consumer goods provider, and sorting boxes and tubes at a pharmaceutical distribution center.

In addition, esearchers at Carnegie Mellon University are working on flexible grippers with multiple sensors and stretchable optical sensors. These grippers could be closer to the sensitivity of human hands and could be also be relevant to prosthetics.

Shadow Robot Co. is working with the European Space Agency on remotely controlled grippers for space exploration.

Human hands aren’t the only biological model for fine manipulation. Engineers at Cornell University have created 3D-printed soft actuators replicating the function of an octopus tentacle. Motion Control Robotics is among the companies using vacuum grippers to grasp irregularly shaped objects.