AirSim + TensorFlow = Deep Learning Fun!

Originally developed as an MAV flight simulator, Microsoft's AirSim program recently added the ability to drive a simulated  car.  Although my own interests lie on the MAV side,  students in my recent AI / Deep Learning course chose to work with the car simulator as the basis for their project. 

As the video and code show, AirSim's support for Python interaction, combined with the powerful features of the TensorFlow deep-learning package, made it a matter of a few days' work to put together a simple demonstration of visually-guided collision avoidance. 

With sophisticated simulation and machine-learning tools available to anyone for the price of a decent workstation and GPU, it promises to be an exciting year ahead!

Views: 2413

Comment by Mateusz Sadowski on December 19, 2017 at 12:46am

Very interesting results! Do you happen to have or planning to publish any videos from your lectures? I would be very interested to follow it online if possible!

Comment by Simon D. Levy on December 19, 2017 at 1:09pm

"Unfortunately our classroom was not equipped for video capture, but the lecture notes and assignments are all posted online.  For hands-on self-teaching, I can also recommend this new book about Deep Learning."

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2018   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service