Indoor AI-based UAV installations to improve inventory, materiel, and safety inspections

A low-cost UAV with Level 4+ indoor autonomy AI software stationed on base will drive Air Force goals of responsiveness and speed by streamlining materiel and facility inspections on domestic and overseas bases. Imagine fleets of UAVs flying programmatically inside every Air Force logistics depot to count inventory 10x faster, or UAVs with machine vision AI finishing aircraft inspections in hangars 10x earlier.    Outdoor UAVs with GPS are maturing. However, UAVs flown indoors still require human intervention in single missions & have only partial user-led autonomy. What’s missing is an AI system for large-scale indoor navigation & perception that’s persistent over full mission life-cycles, while needing minimal infrastructure changes for flexible installs. This allows UAVs stationed in facility interiors to fly repeatedly without human operator. This has the potential to advance the frontiers of UAV autonomy, opening up new use cases indoors for logistics, manufacturing inspections, & facility safety. Corvus has overcome critical technical challenges integrating low-level UAV controls, timing, & power systems, with high-level AI software for persistent mapping & navigation in a warehouse environment. This Phase I focuses on the enhancement of the AI software system to account for varying lighting conditions, facility layouts, and inventory types in USAF end-user facility environments.   In Phase I, we will visit the appropriate Air Force customer to understand end-user environments. We will demonstrate a UAV with AI autonomy software and charging dock, navigating indoors to detect objects of interest. The UAV’s AI software will complete missions fully autonomously indoors (no operator), including takeoff, return and land for recharge, and data transfer programmatically. It will not require Internet connection.    One possible outcome we foresee is, if selected for Phases II or III, we will integrate our AI-led UAV onto one or more Air Force base warehouses or hangars. We will add other AI capabilities for Air Force needs. This includes new payloads (RFID, infrared, machine vision), custom-made machine vision AI, and AI to navigate safely through different Air Force settings.    Corvus Robotics is a YCombinator & Silicon Valley VC-backed AI robotics startup. We develop, program, and build our UAVs in our Boston HQ. The founding team’s backgrounds include MIT, graduate degrees in robotics, and researchers in indoor AI autonomy & AI-based perception systems. We have built heavy-lift UAVs, delivery UAVs, & UAVs for outdoor surveillance previously. We plan to grow through sales to enterprise & government customers. We’ve signed purchase contracts with Fortune 500 companies in logistics & manufacturing, and are a top contributor to PX4, the widely-used open source UAV autopilot system.