Structure Inference from Mobility Encounters
Abstract
In the final year of the grant we have focused our efforts on the large-scale analysis of mobility data, such as GPS traces from vehicles. Understanding trajectory data sets, and extracting meaningful information from them, entails many computational challenges, from data set size to sensing uncertainty and trajectory heterogeneity in quality, format, and temporal support. At the same time, individual trajectories can have complex shapes, and even small nuances can make big differences in their semantics. A major tension in understanding trajectory data is between the need to capture the fine details and shape features of individual trajectories and the ability to exploit the wisdom of the collection, i.e., to take advantage of the information embedded in a large collection of trajectories but missing in any individual trajectory. This emphasis on the wisdom of the collection is one of the main novelties of the work presented. We discuss results on extracting a pathlet dictionary for a trajectory collection, on exploiting a collection to better map individual trajectories to an underlying road network, and on exploiting such a collection to derive information that helps the mobile entities.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 20, 2013
- Accession Number
- ADA599132
Entities
People
- Leonidas J. Guibas
Organizations
- Stanford University