GTC-DC 2019: Multi-Platform, Multi-Domain HADR Data Collection (Presented by Lockheed Martin) - Video

We’ll describe how a joint Lockheed Martin and NVIDIA team is using real-world assets like people, vehicles, and roads to simulate a statistically diverse series of humanitarian assistance and disaster relief (HADR) earthquake scenarios at the JIFX 19-2 event. We collected a spatiotemporally synchronized dataset using a large, diverse sensor suite spanning a variety of platforms and stimulus modalities. This dataset will be used in training AI and machine learning systems.