Proximity sensors are increasingly used in distributed monitoring applications, e.g. for surveillance purposes. However, most of existing solutions have two different, usually contrasting requirements, i.e. either they have strong directional features, or they just cover very short distances. In order to tackle such issues, in this paper a new, almost omnidirectional proximity detection technique is described. The proposed approach is explicitly conceived for wearable wireless sensor networks (WSN) nodes and it relies on the fusion of time-of-flight (ToF) and received signal strength (RSS) values. Even though it is not accurate enough for fine-grained indoor localization, the proposed solution is robust in detecting when a given threshold is crossed by a moving object. Also, it is flexible, light from the computational point of view and easy to implement using commercial off-the-shelf (COTS) components.