Autonomous Navigation in a Warehouse with a Cognitive Micro Aerial Vehicle

Micro aerial vehicles (MAVs), such as multirotors, are envisioned for autonomous inventory-taking in large warehouses. Fully autonomous operation of MAVs in such complex 3D environments requires real-time state estimation, obstacle detection, mapping, and navigation planning. To this end, we employ a cognitive MAV equipped with multiple sensors including a dual 3D laser scanner, three stereo camera pairs, an IMU, an RFID reader, and a powerful onboard computer running the ROS middleware. Tasks with hard real-time requirements such as attitude control and state estimation are processed on a Pixhawk Autopilot, which communicates with the main computer via the MAVLink protocol. In this chapter, we describe our integrated system for autonomous MAVbased inventory in warehouses. We detail the involved components and evaluate our system with the real autonomous MAV in a realistic scenario. We also report lessons learned during field testing.