Hierarchical Modeling of Road Traffic Networks

Sammanfattning: The main objective of the thesis was to develop and applymethodologies for analysis and evaluation of the applicabilityof different types of road traffic simulation models for urbantraffic conditions. For this purpose a number of models ofmacroscopic, mesoscopic and microscopic type were applied,calibrated, validated and analyzed using extensive field datacollection to obtain traffic flow, travel time and queue-lengthdata for major road networks in Stockholm and Uppsala.The thesis is composed of four parts. Part I contains thebackground statement, objectives of the study, overview ofmodeling techniques, synthesis and recommendations for furtherresearch. Part II to Part IV document separate studies withdifferent traffic models including objectives, modeldescription, data collection, analysis and conclusions. Thetitles convey the specific type of traffic facility selectedand investigated using various transport modelingtechniques.Part II: Urban Arterial Study treats the subject ofcalibration and validation of the stochastic microscopictraffic simulation model TRAF-NETSIM for Swedish road trafficconditions. The method is described and the sensitivity of thedistributions of free flow speed, start-up lost time, queuedischarge headway and accepted left turning gaps is described.It was shown that the quality of the model improved andapplicable for local traffic conditions.Part III: Urban Traffic Corridor Study describes theimplementation; calibration and analysis performed on a numberof high-resolution mesoscopic and microscopic (CONTRAM andCORSIM) transport models, a prerequisite for the need to modelfast emerging ITS systems. Field surveys feed the models andexpose their ability to reproduce the actual conditionsobserved in the field. Both models reproduced the field datasatisfactory. The extent of the network imposed limitations onthe operation of the microscopic model.Part IV: Urban Road Network Study investigates thehypothesis that the use of macroscopic transport planningmodels may cause significant errors if applied on congestednetworks. An alternative method to analyze congestion impactsis the use of dynamic meso- or microscopic simulation models.Two macroscopic static assignment models, DSD and TransCAD, andthe dynamic assignment model CONTRAM were tested andcomparative studies performed based on comprehensive fielddata. The macroscopic models were shown to be poor indicatorsof network travel time. Better correlation was obtained fromthe mesoscopic model.