The use of floating car data is an interesting method tomonitor traffic. Vehicles act as local traffic sensors anddata from individual vehicles is aggregated into higher-levelinformation. We propose a number of reusable organizationabstractions and a software architecture to support amulti-agent approach for floating car data. The abstractionsare based on the idea of hierarchical organizationswhich are used as units of data aggregation. In this approach,an agent is deployed on each vehicle. At the lowestlevel, nearby vehicle agents collaborate to aggregate individualtraffic data and distribute it to local clients such astraffic light controllers. At higher-levels, organizations arebuilt up from lower-level organizations and represent specificaggregation interests such as the total congestion levelin a specific area. This decentralized approach avoids thebottleneck of a centralized control center and makes the systemmore robust and scalable. A prototype was built, supportinga two-level organization structure, and is used in asimulated traffic environment as initial validation.1 Introduction