The measurement and provision of precise and upto-date traffic-related keyperformance indicators is a key element and crucial factor for intelligenttraffic controls systems in upcoming smart cities. The street network isconsidered as a highly-dynamic Cyber Physical System (CPS) where measuredinformation forms the foundation for dynamic control methods aiming to optimizethe overall system state. Apart from global system parameters like traffic flowand density, specific data such as velocity of individual vehicles as well asvehicle type information can be leveraged for highly sophisticated trafficcontrol methods like dynamic type-specific lane assignments. Consequently,solutions for acquiring these kinds of information are required and have tocomply with strict requirements ranging from accuracy over cost-efficiency toprivacy preservation. In this paper, we present a system for classifyingvehicles based on their radio-fingerprint. In contrast to other approaches, theproposed system is able to provide real-time capable and precise vehicleclassification as well as cost-efficient installation and maintenance, privacypreservation and weather independence. The system performance in terms ofaccuracy and resource-efficiency is evaluated in the field using comprehensivemeasurements. Using a machine learning based approach, the resulting successratio for classifying cars and trucks is above 99%.
展开▼
