33 Trust and Security in Spatial Messaging: FoxyTag, the Speed Camera Case Study Michel Deriaz and Jean-Marc Seigneur Abstract. Current speed-camera alerting systems heavily rely on humans to check the trustworthiness of information sent by their users. Hence, these systems are often either expensive or suffer from drawbacks, such as incomplete information concerning mobile speed cameras. We propose an application called FoxyTag using a computational trust engine instead of human checks. FoxyTag lets any driver equipped with a Java/GPS-enabled mobile phone to post a virtual tag about a speed camera and notify other equipped drivers who can confirm or deny the (short-lived) presence of the (mobile) camera. The novel aspect of our trust engine is that it must be location and time aware to automatically compute the trustworthiness of the given tag. We have validated FoxyTag both in real-life settings and with a simulator for large-scale scenarios. The validation showed that our novel time-based trust metrics are appropriate. 1 Introduction European countries have been multiplying the number of speed cameras in order to reduce the number of car accidents. In response to that and in order to assist the drivers, more and more companies sell information systems able to warn the driver nearby a critical (speed camera) zone. Because these systems heavily rely on humans to check the trustworthiness of the information sent by their users, these systems are often either expensive or suffer from drawbacks, like for incomplete information as they do not handle mobile speed cameras. We propose a new system called FoxyTag that allows any driver equipped with a Java/GPRS-enabled mobile phone with access to a GPS (for example, we currently use affordable external Bluetooth GPS modules) to easily signal the presence of a speed camera or to signal that one has been removed. This is done by posting to our central server via GPRS virtual spatial messaging tags at the critical points, so that other drivers can be alerted on time by querying this server by affordable GPRS connections. The novelty of our system is that it does not require human checks to decide about the trustworthiness of the posted tags because our system uses a computational trust engine to automatically make this decision. Due to the fast context changing aspect of this application domain – mobile speed cameras are short lived – previous work on computational trust was challenged and we had to engineer novel time-patterned trust metrics.