This platform mainly serves the field of autonomous driving multimodal fusion perception. The multimodal perception dataset (OpenMPD) we built is committed to collecting information on highly challenging road scene in the real world and then labeling it so that openMPD can execute more accurate test on the adaptability of automatic driving related algorithms in the real world. With cutting-edge research such as single vehicle multimodal perception and vehicle road integration involved, we are expecting to see an accelerated and large-scaled commercialization of autonomous driving.

This platform provides a professional algorithm benchmark evaluation solution. You are welcome to conduct training and testing on this dataset and then submit your solution in the algorithm evaluation (hyperlink) module!

At present, we have so many types, brands, and models of Vehicle-mounted sensor, yet very less engineering experiments to find out the most suitable sensors for automatic driving. This platform will continue to release the horizontal and vertical evaluation reports of different brands and types of sensors for relevant researchers’ reference.