Using robots to unload cargo at docking bays is not a new idea. A quick YouTube search will return numerous results. However, many of these previous approaches have been suppressed due to high costs and the inability to handle an increasing variety of packages.

Of course, robotics companies are still working on this problem, which is one of the major issues for warehouse operators. Memphis-based robotics startup Dextrous Robotics is developing its Chopstick system, hoping to achieve greater success in unloading at only a fraction of the cost of previous solutions. Co-founder Evan Drumwright discussed this approach at last week's RoboBusiness Direct conference session, "Advances in Robot Picking, Grasping, and Manipulation."

Now, Boston Dynamics is putting its innovation hat to work. It showcased its latest robot, Stretch, a mobile manipulator designed to move boxes out of trucks and warehouses. When it goes on sale in 2022, Stretch will initially focus on truck unloading, and later, given its price point, will add palletizing to its repertoire. Currently, some partners are testing Stretch. Boston Dynamics (Boston Dynamics) is seeking customers to try using Stretch for truck unloading tasks.

Next-Generation Handle

Stretch is the next-generation Handle, a robot combining wheels and legs launched by Boston Dynamics in 2017. Stretch does not have legs, but it has an omnidirectional mobile base with four independently controlled wheels, a custom 7-degree-of-freedom industrial robotic arm capable of lifting 50 pounds, a custom suction cup gripper, and more. The pneumatic system for the grip paper tooth is located on the mobile base. Stretch has a battery life of 8 hours, but a 16-hour battery option will be available, with the ability to plug in Stretch for continuous power.

Stretch uses the Pick vision system, which Boston Dynamics acquired in 2019 when it acquired Kinema Systems in month 4. Pick uses high-resolution 2D and 3D vision along with machine learning algorithms for robotic depalletizing. Kevin Blankespoor (Kevin Blankespoor), Vice President of Product Engineering at Boston Dynamics, stated that Stretch can process 800 cases per hour. Its arm has a seven-foot reach and can access boxes up to 10 feet high. Stretch weighs 2, 650 pounds.

Blankespoor states that Stretch is semi-autonomous, with the level of autonomy depending on the application. For example, when unloading a truck, a person is still needed to open the truck doors, verify items, and push the Stretch Robot into the truck using a joystick to position it. Blankespoor said, "At that point, you press the 'Go' key, and Stretch will fully automatically complete the rest of the work, unloading the boxes."

Building pallets will be the second application for Stretch, and Blankespoor envisions greater autonomy for this task. Stretch will operate in warehouse aisles. It will go to different pallets of individual SKU SKUs and grab several boxes to build outbound pallets. For this task, Stretch will take on more positioning and navigation, similar to what you see in other AMR (Autonomous Mobile Robots), which can navigate warehouses with greater autonomy.

Blankespoor said that collaborative robotic arms lack the speed and strength to meet Boston Dynamics' target for the number of heavy boxes handled per hour. Therefore, it built a custom industrial robotic arm. For safety, Stretch uses speed and separation monitoring. He said, "As people get closer to the robot, it will slow down. If they get even closer, it will stop. This allows us to move heavy boxes quickly with the arm without any risk of injuring anyone."

Hello Robot also has a mobile manipulator named Stretch, designed as a research platform. According to IEEE Spectrum, Hello Robot co-founder Aaron Edsinger (Aaron Edsinger) was not thrilled; he had worked on robotics at Google when Google owned Boston Dynamics. "We were disappointed that they chose our name for their robot. We are paying close attention to this and considering our options."

Key to Handling Variety

You may notice in the Stretch video from Boston Dynamics that all the boxes inside the truck and on the pallets look identical and are stacked perfectly. In the real world, most, if not all, trucks are not like that. Therefore, the success of Stretch, like other solutions attempted previously, will be determined by its ability to handle a wide variety of packages.

Blankespoor said, "That is a keen observation. It is easy to detect a box sitting alone. However, detecting a box when they are tightly packed together or highly varied is a challenging task. But this is where our 'Pick Box' detection software comes in. This is one of the things you will see in Stretch. We will start with simpler trucks, but we will incorporate more complex box types and different kinds of boxes over the next year."

Aaron Prather (Aaron Prather) of FedEx (FedEx) spoke about the greatest challenge in solving the truck unloading problem—the high degree of package diversity required of the solution—in a recent keynote address at RoboBusiness Direct. Prather began his career unloading trucks for FedEx while still in college.

"When unloading bulk trucks, you might start with boxes of a certain size, then encounter a larger or smaller box, then a layer of carpets (since people now buy carpets online), and then return to the same size boxes you started with. This is why it is difficult to automate this process."

Is the Stretch Robot Just Getting Started?

Boston Dynamics (Boston Dynamics) obtained a large number of warehouse automation patents from Google's R&D organization, X Development LLC, in 2020. Boston Dynamics (Boston Dynamics) was owned by Google from 2013 to 2017. Of course, certain IP can help Stretch, but it may indicate broad interest in the logistics automation sector.

Google acquired the Industrial Perception company in 2013, which conducted truck unloading research in partnership with Wynright. Here is a video of the partnership.

"Well, we have certainly developed an interest in the warehouse sector since Atlas. By that time, we recognized it as a large market and understood the potential for mobile robots in warehouses. I don't want to discuss too many patents, but we have a rapidly growing warehouse robotics team, which I lead. This includes developing the Stretch robot, possessing the Pick vision system, and having fleet management software systems that can bring different robots together for coordinated activities. Thus, this truly aligns with the broader prospect of Boston Dynamics striving to enter the warehouse sector."

Learning from the Field

In month 2 of 2021, Boston Dynamics launched the Spot Arm, a robotic arm for the Spot quadruped robot, with a maximum lifting capacity of 24.3 pounds. Although the Spot Arm looks significantly different from the industrial robotic arms of Stretch, Blankespoor stated that the core technical components are the same.

"If you look at the wrist joint of Stretch, it is actually the same as the hip joint on Spot. Therefore, it uses the same motors, the same gearboxes, the same sensing, and even the same software," he said. "We use the same cameras and depth sensors across different platforms. More importantly, we use the same underlying software to truly understand what those sensors are seeing—everything from box detection to obstacle detection. Therefore, even though Stretch looks quite different from Atlas or Spot, it is built using mostly the same technology. This allows us to prototype quickly."

Atlas remains an R&D platform, while Boston Dynamics (Boston Dynamics) has sold more than 400 Spot robots since the quadruped robot first went on sale last year. Boston Dynamics (Boston Dynamics) won the 2020 RBR50 Innovation Award for launching the commercial quadruped robot market.

Fleet management, WMS integration

Boston Dynamics and OTTO Motors released a video in 2020 on month 3, demonstrating their vision for the future of logistics automation. In the video, the Handle robot from Boston Dynamics picked up boxes and built pallets on top of the 1500 heavy-duty AMR from OTTO. Blankespoor stated that the two companies will continue to collaborate and play a significant role in the development of the Stretch fleet manager.

“A key part of the demonstration with OTTO Motors is enabling heterogeneous robots to work together. Many AMR companies control their own robots, and they may need a fleet manager to achieve this. However, there are not many fleet managers that can control various different types of robots and coordinate them to perform different tasks. One of the reasons we partnered with OTTO is to use different robots to coordinate pallet movement and box movement.”

Blankespoor also stated that integration with the Boston Management System is on Boston Dynamics' roadmap. However, he said this will be crucial later when Boston Dynamics begins handling order building.

“Truck unloading works very well because, in terms of integrated tasks, it is more isolated. Once the robot moves to the front of the truck and says ‘go,’ our box detection system tells Stretch where all the boxes are and what their sizes are. We actually weigh the boxes during the first lift, so we can move them as quickly as possible. This means we do not have to perform extensive integration work.”

“The second task we will complete is order building. For this, you truly need warehouse management system integration, but it should be similar to the work done with human workers, correct? It involves an instruction such as ‘build me a pallet with five of this box and four of that box, then stack them all.’ We need to convey this information to our robots so they can execute these tasks.”

Evolution of Box-Handling Robots

You might think logistics robots are outside Boston Dynamics's domain, but the company has shown strong interest in this area. In 2016, one of its first videos of the next-generation Atlas robot showed a humanoid bending down, picking up a box, and placing it on a shopping cart. Then in 2017, Boston Dynamics launched Handle, a hybrid robot combining wheels and legs. Stretch is the next generation of Handle.

Blankespoor stated, “We received significant interest in [Atlas] from warehouses,” “We knew we could design a simpler robot, and this is the true origin of Handle. Handle is a branch of Atlas.”

“For several reasons, we branched off and designed Handle. One reason was that we wanted to do something more specialized for warehouses. Another reason was that we had always wanted to combine wheels and legs. Handle provided an opportunity to explore both. Therefore, we built several versions of Handle and began performing warehouse tasks.”

“Initially, we were performing pallet building, and it worked very well. Then, we began using Handle for truck unloading. At that time, it was truly an eye-opening experience—Handle could handle the job. It could grasp boxes and move them, but it took too long. To build cost-effective robots for customers, you need to move boxes quickly. Stretch can move boxes 5 times faster than Handle, which is one of the reasons we made the next leap.”

Early Customers for Hyundai?

Hyundai Motor acquired a majority stake in Boston Dynamics from SoftBank for approximately 8.8 hundred million USD in 2020 on month 12. SoftBank's subsidiary retained the remaining 20% share. The Spot Arm is an important announcement because it effectively turns Spot into a mobile manipulator, not just a data collection platform. However, the launch of Stretch is the biggest announcement from Boston Dynamics since the acquisition. Will Hyundai become an early test customer for Stretch?

Blankespoor stated, “It is definitely possible.” “We are excited to collaborate with Hyundai. Obviously, they are a world-class manufacturer. Hyundai has its own logistics company in South Korea, and they will certainly become our customers in the future.”