Piece picking for e-commerce order fulfillment poses several challenges for robotics engineers. Throughput, accuracy, reliability, and versatility are key requirements. While some companies have struggled to train robots to handle a wide variety of items, XYZ Robotics has developed end-of-arm tooling that it said is tolerant of sensor errors while significantly increasing productivity: a spear tool head, its JY tool changer, and a bag cup for grasping plastic bags packed items, .
At Automate 2019, the Allston, Mass.-based company introduced its XYZ Rebinning Station (XRS) and XYZ Picking Station (XPS). The XRS is an autonomous robotic turnkey system for putwall sorting, and XPS is a robotic piece-picking system for goods-to-person transport.
Tooling for speed and reliability
Using computer vision and machine learning to recognize thousands or millions of stock-keeping units (SKUs) is only part of the challenge. Many robotics suppliers are still working on training robots and end-of-arm tooling to handle items of different shapes, sizes, materials, and orientations.
“Picking for putwall sorting involves mixed items, so goods-to-person and goods-to-robot operations can’t depend on homogeneous totes,” noted Peter Yu, chief technology officer of XYZ Robotics. “Our vision system and tooling make it easier to handle challenging cases at high speed.”
For example, as robotic arms reach into bins, they can be knocked off center by objects in the way of their objectives and could even damage goods, he said.
The company’s compliant tool head can be pushed out of alignment and automatically spring back into place. This helps avoid costly downtime to manually reset the robot, Yu added.
XYZ Robotics’ JY tool changer is magnetic and has a self-centering head, allowing for swaps of less than half a second. Once XYZ’s vision system has identified the item to be picked, the robot arm can go to the tool changer and choose from a heavy suction cup, a small suction cup, or a pneumatic gripper, accordingly.
“This is good for picking up cosmetics, such as shampoo bottles and lipstick,” Yu said.
The third tooling innovation from XYZ Robotics is a patented bag cup that allows for loose polybags to be picked up via suction from any angle. This feature is especially important to process loosely bagged items such as apparel, and XYZ’s vision system can also detect transparent packages, Yu said.
Focus on components, innovation
Major consumer-goods manufacturers and retailers have expressed interest in XYZ Robotics, which is a member of the Robotic Industries Association and MassRobotics, said Yu.
“We’re also working on grocery scenarios, which are most challenging,” Yu said. “No vision processor in the world works perfectly, but we’ve tested our system to survive continuous overnight operations.”
“We also have a strong marketing team in China,” he added. “However, some large companies won’t work directly with startups, preferring trusted integrators.”
“Most AI startups needs tooling to support their vision systems, and integrators look for a vision system that can work with their tooling,” said Yu, “Our goal is to bridge the gap: scale and focus on components for both integrators and startups.”
XYZ Robotics’ tooling works with UR5 collaborative robots, and the company has developed adapters for robots from ABB and other major robot suppliers. XYZ also designs its own camera system using structured light for accurate bin picking.
During a recent site visit, the company showed The Robot Report other systems that it is working on to further accelerate materials handling.
“With CNC prototyping, we can be 50% faster in developing products,” Yu said. “We’ve built putwall and tote systems, but first we want to sell the key components, such as our end-of-arm tooling and our vision processor, which connects to controllers through Ethernet to send pick points.”
XYZ Robotics has raised Series A round funding and is currently hiring senior roles in robotics, manufacturing engineers, and business development for the U.S., European, and Asian markets.
Tell Us What You Think!