A Bilevel Traffic Data Extraction Procedure via Cellular Phone Network for Intercity Travel

Speed prediction from mobile sensors using cellular phone‐based traffic data

“…Furthermore, it was assumed that each vehicle has at least a switched-on CP aboard (100% penetration rate). Vehicular-based CP data were extracted in a previous research endeavour by a model that did not use a secondary reference system [3] from which SMS was estimated for 8 h at aggregation intervals of: 1, 3, and 5 min [4].…”

“…Traffic speed extracted from cellular phone (CP) data is a probe-based method that has the advantages of: (i) wide coverage, (ii) very high penetration rates, and (iii) minimal cost tag. Thus the potential of deploying prediction models based on CP data is very high [3].…”

“…As the CP LA is changed (by the CP transition from one location to another), an LA update is exerted, by a process known as 'handover', to preserve the received signals strength by the CP. Spatiotemporal data of the CPs can be collected from cellular networks from which several types of traffic and mobility data can be extracted by using various data mining algorithms [3,4].…”

“…Moreover, CP data have different geo-referencing accuracies, frequencies, and data streaming. Nonetheless, the environment of wireless network systems has many uncertainties and variables that include: possible noise at handovers' locations, distance between successive handovers' locations, and the number of handovers at each handover location (penetration rate) [3,5].…”

“…The vehicular-based traffic data were extracted from augmented CP data by using a clustering technique (modified Data Swarm Clustering). The full details of the estimation procedure can be found in [3,4]. The clustering model was able to eliminate the noise in CP-based data presenting a more realistic dataset of the traffic stream.…”