| With the continuous development of computer technology, the architecture of intelligent transportation system becomes more and more complex,having distributive,heterogeneous and hierarchical characteristics.The simulation environment with a single over the past can only model and simulate part of functions of the transport system, not meeting the need of modeling and simulating the entire transport system. Meanwhile, before implementing a logical system model,it is very important to ensure the logical correctness of the system. To solve the above two issues, this paper proposes a simulation approach of building transportation system in Ptolemyâ…¡software and uses DE domain to analyze the logical correctness of single layer model and use PTIDES domain to analyze its parallel execution of each node in the transportation model. The main works are as follows:1.This paper proposes a traffic modeling and execution method for single intersection, which a single intersection model is composed of vehicle detector node, traffic controller node and traffic light node. Vehicle detector node sends messages to the traffic controller node.After completing the handling of traffic event, traffic controller node sends control command of traffic signal to the traffic light node. The execution process of the entire model is managed by the DE domain, DE scheduler asynchronously handles events according to timestamp of events. DE Simulation experiments show that events are processed with the style of determinism and uniqueness.2.This paper proposes a modeling and execution method for Area Traffic Control System, which hierarchical structure system is modeled flatly and the relationship between a single node and its sub-nodes is modeled hierarchically. Area Traffic Control System characterizes its multi-levels. Among them, the top-level model uses PTIDES domain and the remaining sub-models use the DE domain. PTIDES domain improves the execution efficiency of composite actor in the top-level model and DE domain ensures execution process of all sub-models with deterministic and asynchronous style. Simulation experiments show that if the nodes that have no causality relationship in traffic system,the event can be processed not based on timestamp order, so this can improve the execution efficiency of the nodes.3.This paper introduces a fast priority queue, which is called the calendar queue. Calendar queue means that an infinite queue is divided into a finite number of small queues and all events of each queue is sorted according to timestamp of the event. Then the DAG topological sorting algorithm is introduced, which means that an actor in the model is given appropriate priority according to its position of the topology.The Smaller the value, the greater the priority. Although the result of a topological sorting is not unique, but the priority of the parent node is always higher than the priority of child nodes.Finally, this paper summarizes the research work and points out the further study. |