A fuzzy logic-based car-following model for simulating transit vehicle movement at and around bus stop used by various users

Linking urban bus transit analysis with traffic engineering contributes to safe and efficient movement of people. A simulation model capable of modelling passenger service at bus stops and traffic conditions around bus stops is an excellent tool in transit efficiency assessment. In the paper, the analysis focused on street segments between intersections with bus stops used by regular bus service, various other transit providers (private operators), and other users (taxis, passenger cars). Large numbers of minibuses of non-urban bus operators stopping at bus stops cause significant disruptions in the traffic flow. The apparent differences in the efficiency and operation of the stops designed for use by local buses and those used by other users required additional analyses. In the probabilistic and simulation models of bus stop operation developed so far, did not use car-following models for the modelling of the movement of buses and other vehicles at and in proximity of bus stops. This theory is used in off-the-shelf traffic microsimulation software packages, but they do not allow a reliable representation of how bus stops operate under the deregulated service of passengers (oversaturated bus stops with a variable number of service channels). This study analyses the movement of buses and other vehicles at a bus stop area using a car-following model with fuzzy logic applied for parameter estimation. A microscopic simulation model was built and tested. The selection of the shape of fuzzy sets and membership functions was based on multiple simulation runs. The model simulates queuing and delays imparted on buses and traffic flow in lanes adjacent to bus stops used by city buses and other transit providers. The simulation results were compared with the real-world processes with the use of the author's two-stage method. Verification based on the PROC and SDDIST indicators showed that the simulated and observed distributions of travel time and delay were highly consistent, with maximum deviations below approximately 10%. The analysis also confirmed that the model accurately reproduces average delays caused by bus queues and vehicle interactions near shared-use stops. The analyses demonstrated that the fuzzy logic-based car-following model proposed in this study is suitable for planning, designing, and relocating urban bus stops used by various operators. The feasibility of the developed model and directions of its further development were evaluated.