Novel Robot System Helps Order Fulfillment Centers Keep Up with Demand

Supply chain disruptions stemming from COVID-19 sent shock waves through industries of all types, from restaurants to manufacturing to e-commerce. Many companies had to rethink their strategies, whether that meant diversifying or reshoring their supply chains or implementing social distancing and new rules and regulations in warehouse settings.

One way to minimize economic impact while keeping products flowing out the door is to deploy advanced automation technologies such as artificial intelligence (AI) and robotics. A recent Deloitte study, “The rise of the social enterprise,” shows that even before COVID-19, many companies were already actively redesigning jobs around AI and robotics, which may not come as much of a surprise. Automating processes in fulfillment centers not only frees up human workers to add value in other areas but also frees them from repetitive jobs. Furthermore, automation technologies can work around the clock to help fill orders, whether the industry is e-commerce, pharmaceutical, grocery, electronics, or something else. In other words, these technologies quickly show their worth.

Autonomous Item Handling

Throughout the past year-plus, the RightPick™ robotic system, from RightHand Robotics (Somerville, MA, USA; www.righthandrobotics.com), has helped fulfillment centers all over the world keep the flow of essential items moving. The system handles the core task of picking and placing individual items as part of a range of fulfillment center processes.

The RightPick autonomous item-handling robot system is powered by RightPick AI software, which lets the robot handle thousands of SKUs from totes, bins, boxes, and cases with a hardware platform consisting of a multifunction smart gripper that combines sensing, suction, and fingers and provides feedback on grasp success and pick accuracy, allowing the robot to achieve low error rates and reducing manual touches for order fulfillment or replenishment.

Additionally, the system features UR5e or UR10e robot arms from Universal Robots (Odense, Denmark; www.universal-robots.com) or other industrial arms and Intel® (Santa Clara, CA, USA; www.intel.com) RealSense™ D415 3D depth cameras. In addition, the RightPick MCP API enables easy integration with existing warehouse software systems enabling a variety of typical workflows.

The control system comprises a robot arm controller, a safety controls unit, and the RightPick processor, where the 3D camera and AI software give the robot vision and decision-making abilities for picking and placing products. The AI software also leverages machine learning to continuously improve picking. At the heart of the processor sits an industrial PC customized specifically to RightHand Robotics’s requirements by CoastIPC (Hingham, MA, USA; www.coastipc.com). The scalable and reliable industrial computer is ready to work out of the box, which removed a major specification hurdle for RightHand Robotics when it came to deploying its system at scale.

In a typical workflow, RightPick AI algorithms process image data, assess which picks are easiest, and determine both grasp and arm motion plans. RightPick AI checks for obstacles to avoid bumping the tote or dislodging the item and handles exceptions, such as phantom or empty totes, with confidence and notifies the host warehouse software to take appropriate action. If a pick doesn’t succeed on the first try, RightPick AI decides whether to try again or instantly switch to a backup plan for a different item.

Related

• 3D Avatar Accurately Recreates Movements of Athletes
• Edge AI and Machine Vision: How Today’s PCs Evolve to Meet Growing Challenges
• Machine vision system inspects mixed model automotive components

Once the item is grasped, sensor feedback confirms that the robot has it throughout the transfer motion. If needed, RightPick barcode-scans items, communicates with the warehouse host to confirm each item is expected, and is told mid cycle where to place them. To place with precision, RightPick AI must understand not only the dimensions and position of the item in the gripper but also the vacant areas in the tote for possible placements. This eliminates damaging items and optimizes fill density.

Keeping the Doors Open During COVID

One relevant example of the RightPick system in action comes from Japan, where consumer packaged goods company Paltac Corporation (Osaka, Japan; www.paltac.co.jp) introduced several measures to prevent COVID-19 from spreading among employees. These measures included temperature checks, mask wearing, and regularly scheduled decontaminations. Paltac also uses several classes of robots capable of automating the fulfillment process, providing reliable throughput.

Paltac stocks more than 20,000 SKUs. Handling so many different and densely packed items required a piece-picking solution that combined advanced gripping systems with continuous machine learning and total reliability, according to the company. Paltac’s team also values the flexibility of the RightPick system. The fact that RightPick can improve from machine learning “experience” and even recognize and grasp new products that it has not seen before is important to satisfying customer expectations over time.

Paltac’s deputy manager of R&D, explains why the company chose the system after an evaluation period. “RightPick met all of our high-level requirements; we’re confident in the system’s ability to successfully pick a wide variety of items without prior knowledge or programming,” he says.

Toward Lights-Out Operation

Apologistics GmbH (Markkleeberg, Germany; www.apologistics.com), a digital pharmacy in Europe, handles more than 100,000 products. It opened a new, 220,000-square-foot facility that operates 24 hours a day, seven days a week—an important step toward a lights-out order replenishment, storage, and fulfillment operation. The company invested in a fleet of RightPick robot systems to integrate with its cube storage automation from AutoStore. The AutoStore system presents six inventory totes to each robot, which then autonomously performs inbound receiving and outbound order fulfillment tasks, often picking, scanning, and placing thousands of pharmaceutical SKUs per day.

In operation, the system uses its vision and AI capabilities to gently and predictably handle and place items of varying sizes, weights, and shapes. The RightPick system, according to Michael H. Fritsch, founder and shareholder at Apologistics, was the perfect match for the company's significant fulfillment needs. With the RightPick fleet in place, the company has no more than 20 employees working alongside the robots to achieve an output of 25,000 to 30,000 parcels per day. Competitors, says Fritsch, may have as many as 300 people working around the clock in fulfillment for the same output.

Automating the “New World”

In many industries, post-COVID businesses may look a lot different. In the order fulfillment space, companies will continue to look for ways to improve efficiency to fill customer orders—shipping everything from commonplace items like makeup to crucial medications. Companies like RightHand Robotics will continue to serve these businesses and evolve their technologies to meet the ever-growing needs of disparate markets, and automation will only become more important with passing time.

John DeWaal is President, CoastIPC (Hingham, MA, USA; www.coastipc.com).