Regresar

Partitioning Space to Identify En-Route Movement Patterns

Abstract:

Probe data from delivery vehicles are often not taken into account when performing traffic analysis, given the additional challenges that working with it brings. It is necessary to address the complexities introduced by the nature of the business operations such as stops and speed limits on streets. However, important insights could be obtained and therefore, adequate methods are necessary. In this work, an approach is proposed for working with probe data towards discovering movement patterns in en-route operations of delivery vehicles. Data that came from five months of operations of a vehicle from the Swiss Post, the national postal service company of Switzerland, was analyzed in a segment-by-segment manner. Given the spatio-temporal nature of the dataset, geohash indexing was used to perform the data segmentation and analyze it in a coarse-to-fine granulation level. A three-phase method was defined and as a result, three (3) movement patterns were identified: parking, stop go, and slow point. This work shows that it is possible to obtain insights about what happens on the roads and streets without the need to deploy expensive equipment or a large number of vehicles. Nevertheless, further evaluation methods need to be developed to verify the validity of the results.