The reliability of traffic data and useful short-term traffic forecasts play a very important role in the management of motorway traffic, especially in connection with real-time traffic control systems and the provision of traffic and travel information to road users regarding unexpected incidents. What follows are learning based models which use typical traffic patterns that have different traffic flow characteristics in relation to time and space, and which we refer to as “profiles of traffic flow characteristics” or in short “traffic profiles”. Their significance is related to distinct historical events, which are current until the known traffic demand is determined, which will be determined by the arrival of a particular vehicle. Therefore, we analysed various known clustering methods and proposed the one we considered most suitable to define specific groups of typical and representative time-based profiles of traffic flow characteristics with an emphasis on the daily traffic flow profile of all vehicles and heavy vehicles, regardless of the location of the measurement station and the social environment. Finding representative profiles of traffic flow characteristics is one of the steps by which we can determine any mathematical traffic-state estimation model for predicted road and environmental conditions. A traffic model cannot be defined without knowing the lane flow distribution, the speed distribution and variability of other traffic flow characteristics across the traffic lanes of the motorway in both directions, which we refer to as “lane distribution of motorway traffic” or “cross-sectional traffic profile”. From both profiles we can learn the peak values of traffic flow characteristics in terms of time and space. Based on these and their relationship to environmental factors, we can determine, review and update traffic control algorithms and criteria. Accordingly, an empirical study was carried out to determine the effect of weather conditions on the lane distribution of motorway traffic. Using a micro-simulation traffic model, we showed the importance of precise traffic characterisation for the needs of the traffic control system on motorways. Delays, calculated with the PTV Vissim traffic simulation model, were compared with the results obtained using reference scenarios without traffic control, but under different road and weather conditions and in the same traffic conditions. If delays in a certain area of a motorway network are a measure of the effectiveness of traffic control, and at the same time an indicator of the degree of congestion, then it is important to know the result that the same value of the measure of effectiveness is achieved at different traffic flow density values under different weather conditions.
|
