A Multimodal Data Collection Vehicle for AV Algorithm Development Based on Data Fusion

Autonomous vehicles still need progress towards algorithms for perception in all-environmental conditions (e.g.bad weather). Data collection in such conditions needs complex equipment hard to set-up. At UPC-CD6 (https://www.cd6.upc.edu) we have built an instrumented car with multiple imaging modes (3D+2D) for mapping and sensor testing uses, and to build a publc dataset of driving scenarios to develop DL perception algorithms based on data fusion