Real-time traffic surveillance is essential in today's intelligent transportation systems. This paper reports on the design of a novel, cost-effective, and intelligent real-time traffic monitoring systems using wireless smart sensor networks. Reliable and computationally efficient algorithms were developed for vehicle counting and speed estimation. The integration of system components and distinctive algorithms is discussed in detail. A preliminary experimental study composed of various scenarios conducted under highway conditions resulted in 99.95% counting accuracy and 96.11% speed estimation accuracy. Estimated cost of a single sensor node is less than $40.