Computerized traffic speed measurement using sparse data

Abstract

A computer system for calculating traffic speed using sparse data. Sensors provide location data, over time, for a sampling of vehicles on the road network, such as a fleet of vehicles. This location data from a sampling of vehicles is sparse with respect to both road segments in the road network and time. From the location data, the computer system generates sample data associating speed of sampled vehicles on the road segments to points of time in a plurality of time slots. The computer system accesses other information that defines correlations among different road segments and among different time slots. The computer system derives at least an average vehicle speed for each road segment in the road network for at least the current time slot using the correlation data and the sparse sample data. The computer system can inter traffic volume from average vehicle speeds, and then compute environmental data.