RightHand Robotics and Intel Aim to Revolutionize Automated Order Fulfillment With RealSense Technology
RightHand Robotics is partnering with Intel to improve the machine vision of their RightPick2 Robot.
The partnership joins RightHand Robotics precision picking robot with Intel's RealSense Technology D415 to increase throughput in warehouse picking and shipping operations. With this combination, a single warehouse worker can command a fleet of ten robots picking from thousands of SKUs.
The RightPick 2 is guided by Intel RealSense D415 depth cameras and uses RealSense technology to provide data for motion planning. Image courtesy of Intel.
RightHand Robotics grew out of three robotics research powerhouses, which are MIT, the Yale "Grab Lab," and the Harvard Biorobotics Lab. Their goal is to improve supply-chain logistics by increasing filling speed and accuracy while reducing shrink from damaged goods. Their innovations have included advances in machine vision and learning, as well as picking hands and grips.
The RightPick2 is a software-driven, hardware-enabled platform that performs high-speed picking operations for warehouses and shipping departments. It incorporates the depth cameras to select individual SKUs and routing their path without collisions accurately.
To make this possible, the RightPick2 uses data from Intel RealSense D415 technology. In this system's heart is a depth camera capable of identifying and locating an object in three dimensions. The D415 has a tightly-focused field-of-view, improving the resolution of small parts. This improves the "hand-eye" coordination and allows for much tighter control of how an item is gripped and moved.
Intel RealSense computer vision technology. Image courtesy of Intel.
Robotic selection of neatly-stacked and sorted products is a relatively trivial task. The advantage to the RightPick2 is that it can pick items from a bin, where they have been placed and oriented randomly, as shown in the image.
The RealSense Technology
Rather than programming the infinite number of shapes, sizes, positions, and orientations of every product, the system must "learn" how to identify items and manage their size and shape. The RealSense technology adapts with each pick it makes. In each pick, the D415 system learns about the best way to grip and route items.
Consider a plastic bottle of water. The bottle has a shape, a distinct weight distribution, and, if thrown into a bin, it will land on whatever other objects are in the bin. The RealSense technology can identify this as a water bottle, and decide where it needs to be gripped to keep it from sliding, and how to move it without bumping into other items and equipment. Data is recorded about how to handle a water bottle for future picking.
The RightPick2 will benefit any industry that must pick and ship many products. Its high-speed selections and careful handling will reduce mispicks and damaged goods compared to human-labor intensive picking operations.
Advantages of the RightPick2
Furthermore, robotic picking reduces human handling of products, meaning there is a reduced chance of transmitting a virus from person-to-product-to-person. In the future, RightPick2 could expand beyond only warehouses, with potential inroads into grocery and pharmaceutical markets as well.
The RightPick2 can be integrated with e-commerce platforms. An order can be fulfilled as soon as the customer has committed to the purchase. With e-commerce expanding 15-20% in 2020 more growth expected.
The RightPick2 will likely see service in warehouses and shipping departments and expand into various markets due to its quick, accurate processing and careful handling of goods.