

/cdn.vox-cdn.com/uploads/chorus_image/image/54299301/actual_1364388007.0.jpg)


As anyone who's played the game can attest, it features enough details-like roaming pedestrians and animals, a variety of weather conditions, and complex traffic patterns-to roughly simulate a normal day of real-life driving. Over at Bloomberg Tech, Dana Hull reports that scientists from Germany's Darmstadt University of Technology and Intel Labs have devised a way to feed "visual information" from GTA V into the self-driving software being tested. Self-driving car developers need to put their software in unpredictable real-world situations, sharing the road with unpredictable human drivers, to help the systems anticipate unusual situations.Īnd at least one company is turning to the hit video game Grand Theft Auto V for help. The addition of GTA V is a significant improvement over racing Flash games and opens the door to computer vision and autonomous car researchers.The trickiest thing with autonomous cars is testing them in real-world situations. OpenAI’s Universe came about late last year to address some of this market need, launching with Atari 2600 games, 1,000 flash games and 80 browser environments to help democratize access to training resources. Companies like Udacity have been steadily releasing real-world driving data, but community needs to continue to grow in volume and specificity. Large tech companies and startups alike are digging around for nifty training methods that can reduce the barriers to getting in the self-driving car game and make annotations less of a pain. Training AI frameworks on synthetic data is all the rage these days. The collaboration will cut the once day-long setup time down to just 20 minutes. But today, the work is being integrated and open-sourced with OpenAI’s Universe. The DeepDrive project has made it possible to use Grand Theft Auto to train self-driving cars for some time now.
