Loading...

Fast Centralized Model Predictive Control of Signal Splitting in Large Scale Urban Road Networks

Hosseini Najafabadi, Sajad | 2020

515 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53514 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Haeri, Mohammad
  7. Abstract:
  8. With the rapid increase in the number of vehicles and the limitations on the extension of transportation network infrastructure, the need for utilizing proper methods to reduce traffic congestion has become increasingly important. One way to reduce traffic congestion is to control the signal splitting of traffic lights. Model predictive control (MPC) has been utilized so far by using different models of traffic flow dynamics in this field. Due to the high computational effort of MPC at each step time, the real-time implementation for large-scale networks has become a big challenge. Store-And-forward (SAF) is one of the traffic flow models that, due to its simplicity, makes it possible to perform real-time MPC calculations for large-scale networks. In SAF, the traffic flow dynamics are obtained as a multivariable, first-order, linear, and time-invariant model. However, this model is not accurate enough to predict. Also, due to causing an additional constraint in the optimization problem, it is not possible to use all available green time resources for traffic lights. In this research, by making some changes in the model to increase the accuracy of predicting and also to use all available green time sources, the model will be time-variable. Then a mapping is used to transfer the time-varying factors of the model to the constraints of the optimization problem. Due to the signal-based and receding horizon nature of MPC, the new optimization problem can be solved more efficiently. The performance of the proposed method is evaluated by applying the controller on the SUMO traffic simulator. The results show that the performance of the proposed method is better than the existing methods, and its real-time implementation is possible for large-scale networks
  9. Keywords:
  10. Traffic Light Controling ; Orthogonal Signal Division Multiplexing (OSDM) ; Dynamic Modeling ; Generalized Predictive Control ; Coordinated Traffic Responsive Control ; Signal Splitting ; Traffic Flow Dynamic Modeling ; Signal Timing ; Urban Traffic

 Digital Object List

 Bookmark

No TOC