Enrique Fernandez is a Staff Autonomy Developer at OTTO by Rockwell Automation. OTTO, a division of Clearpath recently acquired by Rockwell Automation, develops AMRs for Material Handling. These robots are designed to automate tasks within factories and warehouses, improving efficiency and safety.
On May 13th Enrique introduced the AMRs manufactured by OTTO, which span from flatbed differential drive platforms with different payload specs to an autonomous forklift.
His presentation focused on perception algorithms that control those robots as well as the safety concerns, autonomous navigation and fleet management.
From a perception standpoint, all vehicles are equipped with safety 2D LiDARs, which are a requirement in industrial environments. For 3D perception, multiple RGBD cameras are used with different configurations depending on the AMR. Those sensors are used to estimate the robot pose in a map built using SLAM. Dynamic object tracking is used to estimate the pose of people or other vehicles moving around the robot, so the navigation algorithms can be smarter when moving in crowded environments. Custom dock targets are tracked at the endpoints to pick up or drop off customer goods as part of the factory or warehouse workflows.
Machine Learning is used to estimate the ground plane so we can detect tiny objects and avoid colliding with overhanging objects. ML is also used to estimate the pose of pallets in order to pick them up with the autonomous forklift.
Those and other perception algorithms were described during the presentation, as well as the use cases and challenges faced by AMRs in industrial environments.