Drones Now Count Warehouse Inventory
Logistics companies are now hiring autonomous drones from Corvus Robotics to handle inventory counting. The drones automate a tedious manual task, providing a real-world application of commercial robotics that relies on robust real-time control, sensor fusion, and embedded system reliability.
Mountain View-based Corvus Robotics has raised a total of $28.1 million, including an $18 million Series A round co-led by Spero Ventures. Spero partner and investor Marc Tarpenning co-founded Tesla, bringing significant experience in scaling complex hardware and software systems. The Corvus One drone operates on a proprietary "AI world model," running all computations for navigation and object recognition on its embedded systems in real-time. This onboard processing, utilizing data from 14 cameras, allows the drone to function without relying on a trunk of external GPUs, showcasing a highly efficient hardware architecture. A key technical differentiator is the drone's ability to navigate without any external infrastructure such as WiFi, GPS, or reflective beacons. This reliance on computer vision and onboard AI enables rapid deployment and operation in dynamic environments, including complete "lights-out" warehouses and narrow aisles down to 50 inches. The system has a notable local footprint, with a successful deployment at the global headquarters of Dermalogica in Carson, California. At that facility, the drones increased the frequency of inventory imaging by 600% and repurposed approximately 120 labor hours per month away from manual counting. Each drone can scan between 200 and 400 pallet positions in approximately 20 minutes, flying at a walking speed of 2-3 mph to operate safely alongside human workers. A specialized version, the "Corvus One for Cold Chain," is engineered to operate autonomously in freezer environments down to -20°F. To optimize its scanning capabilities, Corvus integrates Honeywell's SwiftDecoder barcode-decoding software into its drones. This
