SIMULATING A REGULARLY SAMPLED BUS LOCATION SYSTEM
operations - traffic, infrastructure - vehicle, planning - signage/information, policy - congestion, technology - geographic information systems, mode - bus
Vehicle locating systems, Transit buses, Traffic congestion, Simulation, Microsimulation, Linear referencing, Gridlock (Traffic), GIS, Geographic information systems, Geocoding, Denver (Colorado), CORSIM (Traffic simulation model), Computer simulation, AVL, Automatic vehicle location, Automatic location systems
Between 1967 and 1997, transit agencies in 20 cities installed automatic vehicle location (AVL) systems to improve safety, efficiency, and quality of service. The bus AVL system typically provides a means of tracking individual buses for fleet management. More recently, the AVL data have also been used to develop algorithms to predict bus arrival time, estimate link travel time, and detect incidents. Other traffic management and control applications are also being explored. Therefore, the availability of a simulation capable of mimicking a bus location system that reports vehicle location at regular, prespecified intervals is becoming increasingly important. CORSIM, an integrated freeway and surface street traffic simulation model, also simulates buses on prespecified routes and stations for given dwell time distributions and frequency of service. However, very little has been reported about CORSIM's bus simulation module. CORSIM's bus route simulation module and its drawbacks are examined, and the results of an effort to compare data collected from the Denver Regional Transportation District bus AVL system and the microsimulator for a test network are presented. Bus location data from the field and the simulation were collected at regular intervals under both recurring congestion and nonrecurring congestion conditions. Linear referencing in a geographic information system was used to extract bus location data for the test network from the AVL system, and external programs were written to collect the same data from the microsimulator. Based on the bus location data, space mean speeds were estimated and compared to evaluate the performance of the model. The results are encouraging. However, several obstacles remain and they are discussed in detail.
Khan, S, Hoeschen, B. (2000). SIMULATING A REGULARLY SAMPLED BUS LOCATION SYSTEM, Transportation Research Record, Vol. 1733, p. 105-114.