Activity monitoring is crucial towards enabling smart metropolitan cities. It provides the city officials with the essential information to manage resources, plan events, and help detect and stop insider threats. An automatic activity monitoring methodology that can render real-time and stable estimates of the objects while respecting all privacy concerns in a given environment is therefore of significant interest. This paper investigates the feasibility and presents the promising results of using a sub-GHz passive radio frequency (RF) wireless sensor network (WSN) to monitor activity and determine the number of pedestrians, cars, and cyclists. In particular, we describe our device-free sensing experiment and measure the received signal strength (RSS) between transceiver nodes at an intersection within the smart zone of the city of Antwerp, Belgium, where targets to be monitored do not require carrying a device or a tag. Our results reveal a correlation between the time-averaged RSS measurements and the count of target objects, which indicates the feasibility of employing a sub-GHz WSN for activity monitoring. Solving traffic monitoring using this system will still require investigating the optimal necessary number of radio links and accurate counting models.