Evaluation of bus priority strategiesin coordinated traffic signal systems

Sammanfattning: Increasing congestion and environmental concerns have evoked an interest in promoting urban Public Transport (PT) the last decades. In 2012 the City of Stockholm adopted an “Urban mobility strategy” stating that public transport, cycling and walking should be prioritised over cars in central Stockholm. One of the most important factors influencing the modal choice is the travel time ratio between car and PT travel. According to earlier studies Public Transport Traffic Signal Priority (PTSP) can reduce travel times for public transport with only small negative impacts on other traffic. Conditional PTSP can also help to regulate the PT service. Thus PTSP may support drivers’ decision to change travel mode from car to PT, thus supporting adopted policy goals.Conventional control strategies for coordinated traffic signals have pre-set timings based on traffic surveys. Some traffic adaptation based on real time detector actuations can also take place within the frames of the pre-set cycle time. PTSP changes the signal timings, within pre-set limits, when a PT vehicle is detected. Self-optimising control strategies use a traffic model to predict the traffic flows from traffic counts, and determine the signal changes in real-time by minimising a cost function including delay, number of stops etc. PTSP is included directly in the optimisation by giving PT vehicles a higher weight compared to cars.In this thesis the fundamentals of signal control theory are reviewed as well as unconditional and conditional PTSP criteria and strategies. A simulation based method for evaluation of impacts of different PTSP strategies in coordinated controlled traffic signals is implemented. The simulation setup includes Software-In-the-Loop (SIL) signal controller simulators running the same control logic as used in field. Such simulation models can be useful to test and fine tune PTSP before being implemented in field. Simulations with a SIL setup also enable comparisons of signal control strategies or systems on equal terms, not practically or economically possible in field studies. The implemented SIL simulation model was used to evaluate the impacts on buses and other traffic from the different PTSP functions used in the “PRIBUSS” PTSP method. Short green time extensions showed travel time reductions for buses, with almost no travel time increase for other traffic.Long green time extensions gave somewhat larger benefits for the buses, but more delay to other traffic. Red truncation gave less travel time savings to the prioritised buses and more extra delay for cross street traffic, compared to green extensions. Double red truncation and Extra phase showed some additional travel time savings to the buses, but had the largest negative impact on other traffic. A combination of PRIBUSS functions showed the best results. Depending on the structure of the signal coordination and the location of the bus stops different PTSP functions may be needed.Based on the conclusions from the evaluation of the different PRIBUSS functions a conditional “differential on-time-status” based PTSP strategy was proposed and tested in the SIL simulation environment. The proposed method is focusing on direct travel time savings as well as on reduced bus bunching.The two self-optimising signal control systems Utopia/Spot and ImFlow were tested, and their impacts were compared to conventional control including PTSP with the PRIBUSS method in a SIL simulation environment. The aim was to test if commercially available self-optimising control systems can reduce the overall delay per person by applying more sophisticated PTSP. Both systems reduced the delay for buses, cyclists and pedestrians at a cost of increased delay and increased number of stops compared to the existing conventional control used in field. The total delay for all road users was reduced substantially.