Summary: | 碩士 === 國立交通大學 === 資訊科學與工程研究所 === 99 === Since traffic congestion occurs frequently in urban areas, Traffic Information Systems (TIS) have been constructed to provide road users and traffic managers with accurate and reliable real-time traffic information. Traffic monitoring based on cellular network data, also referred to as Cellular Floating Vehicle Data (CFVD), can be more cost-effective than traditional approaches, such as roadside sensors, because no field installation or maintenance is needed. Consecutive handover events of a handset on a moving vehicle can be used to estimate traffic speed. However, when traffic congestion occurs, handover events are rare, and thus there may be no traffic speed reports from CFVD. Furthermore, the location where mobile stations perform a handover procedure from one cell to anther is imprecise. Thus in this paper, we design novel effective CFVD algorithms to generate traffic information. Using Normal Location Update (NLU) events, we can obtain accurate traffic speed a freeway segment of length about 10 km. Moreover, Selected Handovers (SHOs) that exhibit no cell oscillation problem were used to estimate traffic speed for shorter road segments,. Furthermore, we use call arrival and call completion events to enhance the quality of traffic estimation using CFVD.
|