Real-Time Safe and Efficient Navigation in Windy Urban Environments

Urban winds are complex spatial and temporal fields that pose a critical barrier for autonomous urban aircraft operations. We present a way to predict surrounding wind patterns in real time relying on onboard sensors.

Urban winds introduce risky and inefficient routes for aircraft. We formulate a receding horizon planner exploiting local wind information to find routes through these complex flow conditions, in real time, that are safe and efficient.

We are developing a novel laboratory environment to study autonomous urban aerial operations with a particular emphasis on understanding how wind affects autonomous flight in the urban canopy layer. The facility features a 40x70x14-ft motion capture space with a free flight wind tunnel with 1134 programmable fans for spatio-temporal flow field generation. Planned experiments will attempt to validate concepts above using real world winds and identify critical gaps between simulation and reality.