Traffic Paralysis Alarm System Based on Strong Associated Subnet

Urban traffic congestion is a major problem for urban transportation management all over the world. However, traditional research focuses only on detection and description of urban traffic situations, which are not enough for improving urban traffic conditions. In this paper, we distinguish two types of traffic congestion: traffic paralysis and traffic jams. The former is the state that traffic is almost stagnant in a large area and on many roads, and it will take a long time before recovering the normal traffic flow. In comparison, a traffic jam has less negative effect on traffic flow and recovers easily. According to this, we propose a traffic paralysis alarm system based on strong associated subnet to alert traffic paralysis incidents. The system orients to city road network, mines association rules between road segments, constructs the strong associated subnets and detects traffic anomalies with floating car GPS data. We analyze two parameters of our proposed system with a true dataset generated by over 2000 taxicabs in Zhuhai and explain our system with a simulation experiment on VISSIM.