Bridging Strategic and Microscopic ITS Modelling - a Case Study for Access Control of a New Stadium

Bridging Strategic and Microscopic ITS Modelling - a Case Study for Access Control of a New Stadium


M Fellendorf, PTV AG, DE



This paper will focus on linking a strategic private transportation model with a microscopic model in order to evaluate various ITS measures. Although the paper will be presented using access options of the a new stadium as a particular example, the applied workflow can be transferred to numerous other applications. Various techniques will be shown which are used to port a strategic transportation model to be used in detailed analysis with a microscopic traffic flow model.

Large cultural and athletic events attract masses of people putting pressure on transportation systems, whether they are road or rail based. Since the extra capacity of the transportation system is typically needed for a relative short period of time, expanding the infrastructure should be done carefully. Needless to say, enhancements to the road/rail network are subject to land use restrictions, funding constraints, environmental regulations and possible public opposition. ITS measures have a proven to increase capacity and reliability of existing high volume roads, thus reducing the need for expanding the road infrastructure.

Germany will host the next soccer World Cup in 2006 and stadium renovation planning has already begun. A new arena will be built north of Munich located next to one of Germany?s largest cloverleaf?s ? the interchange between A9 and A99. Access by public transport is guaranteed having a stop at the airport express. Access by private vehicles and buses, however, needs to be designed carefully. A parking lot with a storage of 22,000 vehicles is anticipated and should only be accessed via two motorway exits. To assure adequate traffic conditions on the motorway before and after the games different measures are considered. Among these considerations are constructional instruments like road widening with additional lanes; extension of weaving sections; an additional exit specifically serving the car park, as well as typical methods of ITS like dynamic speed control, route guidance, and dynamic use of shoulder lane. Micro simulation of traffic flow had been used to investigate various options prior to actually implementing these measures. Micro simulation is becoming well accepted as a technique for traffic feasibility studies. Traffic flow models have been looked at in great detail but other requirements have partly been neglected. Since micro simulation is especially interesting for traffic conditions near capacity, precise traffic flow is requested as input data. If networks are modelled with routes to choose, then data about the used paths and its associated traffic volumes are needed. Since path flows can hardly be measured in real life ? except costly licence plate counting ? origin-destination (OD-) matrices and assignment are widely used. 24h or 3h OD-matrices are typically calibrated in transportation planning studies. For micro simulation of peak periods the resolution should be at the most one hour and day-to-day variation should also be reflected. Thus, resolution and accuracy of OD-data has to exceed the typical output of transportation planning. As on many motorways, good data from a total of 68 measurements sites equipped with loop detectors had been made available. Speed profiles and volume counts were obtained for the last six months classified by cars and trucks and provided as five minute values. Statistical analysis was applied to generate characteristic profiles for day-of-week and time-of-year. The motorway network A9/A99 is currently used by local commuter traffic, as well as long distance travellers and freight transport. In the future these different trip purposes and their distinct time profiles will be superimposed by travellers heading for the stadium. A summary of the various OD-tables and its variation will be shown.

Hourly OD-matrices were computed using the traffic counts as mentioned and a validated 24h OD-matrix being available from a transport master plan. The historical OD-matrix was matched with the traffic counts for each time-period and time-of-day using maximization of entropy and fuzzy sets theory. As some detector values indicated higher variations than others, it was obvious that some counts had to be matched closer than others. Differences in variation of counted data is represented by calibrated slack values for each site. The OD-estimation with fuzzy sets is implemented in the transportation planning package VISUM and will be presented. Demand data for FIFA-Cup and the later use of the stadium had been estimated from existing ticket sales within Germany?s First League and UEFA tournaments. The hourly demand matrices for cars and trucks and the event driven demand data had been kept separately and added in the dynamic assignment within the micro simulation.

Key issues in modelling networks on a microscopic level are granularity of the network model (network topology), the core traffic flow model, the representation of traffic control measures and traffic demand as discussed above. To set up simulation networks of considerable size is a costly task, but network topologies meant for in-car navigation systems provide good raw data to start. These GIS-databases have to be supplemented by additional information on link characteristic, desired free flow speed, length of merging areas, location of gantries for variable message signs (VMS) and zonal connectors for traffic demand input. Network consistency has to be checked automatically, as well as visually.

Car following, lane choice and passing manoeuvres are different on motorways compared to urban traffic. In general drivers on motorways tend to plan manoeuvres like speed and lane changes well ahead. Further developments of the micro simulator VISSIM have been undertaken to mimic motorway behaviour spanning over a time period of several seconds. Lane changes at weaving sections will take place over several seconds as tactical driving manoeuvres should be considered. Advancements within this arena will be documented.

Traffic control in the Munich application contained VMS for dynamic speed control, route guidance, dynamic shoulder lane use and actuated signal control at some entries. Variable speed signs are already operated, but different algorithms are tested via simulation to find suitable thresholds for car and truck volumes and their respective speed values. Collective route guidance is tested. By placing gantries at various positions the best potential of effected movements is investigated. To increase capacity some segments on the A99 allow the use of the shoulder lane for trucks in combination with an overtaking restriction for trucks signalled via VMS and actuated by measuring traffic conditions. If this measure has to be activated over long periods road widening will be required.

As hardware becomes more powerful, GIS data becomes better available and the models become more sophisticated, micro simulation will be used more widely to investigate investments in motorway infrastructure, ITS equipment and control algorithms. Micro simulation combined with data gained from strategic GIS-based transportation models offers possibilities to check numerous options in the laboratory prior to implementation. The measures of effectiveness gained from simulation have still to be embedded in


Association for European Transport