With the impending arrival of self-driving cars, the race to build accurate maps is on. Many start-ups exclusively dedicated to building accurate maps have mushroomed and major car companies and technology giants have assembled large map teams. At Qatar Computing Research Institute (QCRI), we are working with local stakeholders and collaborating with MIT to algorithmically build accurate maps using data from multiple sources.
In this talk, I will focus on two problems: (i) map-construction: given GPS trajectories and satellite images, construct a road network graph which is “close” to the ground truth map; (ii) map-fusion: given a base map and GPS data from vehicles that ply on the road network, construct a new map which is consistent with both the base map and GPS data.
Map making turns out to be an excellent pedagogic use case to demonstrate the power of (and possibly even define) “Data Science.” This is joint work with collaborators at QCRI and MIT
Sanjay Chawla is the Research Director of the Data Analytics Group at QCRI. Before that he was a Professor in the School of Information Technologies at the University of Sydney, Australia. His research works spans data mining, machine learning and spatial data analysis. He is a co-author on the text “Spatial Databases: A Tour”