1980 · 10 citations · 0 references
EngineeringSystems EngineeringQueuing TheoryTransportationTransportation EngineeringQueueing AnalysisQueueing TheoryTraffic EngineeringTraffic Flow AnalysisTraffic Flow TheoryTraffic Signal ControlMinimum Variance EstimationPoint Process AnalysisPoint Process FilteringQueue DynamicsQueue Length EstimationTraffic Flow PredictionTraffic Signal Simulation
The report describes some improved algorithms for the estimation of traffic flows based on detector data. Since the modern approach to filtering and prediction begins with the development of models for the generation of the data, the report begins by developing three models for the formation and dispersion of queues at a traffic signal and two models for traffic headway statistics. Recent results in the theory of point process (the type of statistics found in traffic) are then used to develop the nonlinear minimum error variance filters/predictors, corresponding to these models, for queue length, platoon size and platoon passage time. It is then demonstrated that these estimators can be implemented in a microprocessor. Finally, the estimators were tested against the UTCS-1 traffic simulator and performed very well.