Improved Navigation on the Job
At Agility Robotics, we're building robots that are made for work. As our robot Digit enters more job sites, we must have a state-of-the-art navigation stack that’s focused on real-world deployments and creating value for our customers.
We recently celebrated the one-year anniversary of Digit's full-time deployment at GXO. In this bustling facility, autonomous mobile robots (AMRs) arrive at Digit's station carrying totes of product from the order picking area. Digit efficiently unloads these AMRs and loads the totes onto a conveyor, sending them to pack-out stations. To be efficient in dynamic warehouse facilities like this one, robots face a number of challenges. They must navigate tightly constrained spaces, carry heavy objects, and frequently stop and turn in place to successfully handle materials.
To increase Digit’s efficiency on the job, our navigation team completely redesigned the robot’s navigation stack. The robot now solves for the globally optimal, minimum-step path to the goal. This footstep path is then precisely tracked by a sophisticated model-predictive controller, ensuring Digit makes no unnecessary movements. This improved approach has several benefits for the robot:
One impressive demonstration of Digit's adaptability is its ability to handle workflow fluctuations. When downstream stations are backed up or employees are on a break, Digit intelligently stacks totes to the side instead of placing them directly on the conveyor. This ensures the flow of AMRs through the facility continues uninterrupted. Once the conveyor is clear, Digit seamlessly loads the queued-up totes and switches back to its primary task of unloading arriving AMRs.
The improvements to Digit's navigation capabilities directly translate to faster cycle times in these critical workflows, leading to higher throughput and greater value for our customers.
If you want to get deeper into the details of this work watch the full Tech Talk below.