Advanced drone algorithm guides drones through city streets like a car
In explaining the technology, Scaramuzza said: ““With this algorithm, we have taken a step forward toward integrating autonomous drones into our ‘everyday life…Instead of relying on sophisticated sensors, DroNet only requires a single camera — very much like that of every smartphone — on a drone.”
The difference between drones using DroNet and traditional drones is that traditional drones utilize a global positioning system (GPS) for navigation, a system that can become harried and start to fray at low altitudes. So long as your drone is above a city’s buildings, GPS works great, but the minute your drone dips into the city streets, GPS can become a bit of a hazard for consumers and pedestrians.
DroNet incorporates the behavior of bicycles and cars into its algorithm to power the drone’s behavior. To get this behavioral data, Scaramuzza and his team collected data from realtime bicycles and vehicles as they navigated city streets to inform the algorithm’s development.
The DroNet algorithm could see use in a variety of consumer-facing and industrial fields, such as food delivery or emergency services. The algorithm will need to undergo further refinement before it can be deployed in a commercial setting but Scaramuzza and his team are confident they are on the right track. A recent missive detailing the project was outlined in the journal IEEE Robotics and Automation Letters.