Context awareness emergence for distributed binary pyroelectric sensors

We have built a distributed binary pyroelectric sensor network (PSN) for the purpose of multi-walker recognition and tracking. It is important to identify a region of interest (RoI) in the monitoring area in order to find any interesting targets (i.e., walkers). The prerequisite of RoI identification is to accurately extract context features (such as the target IDs and positions) from a hybrid, binary, multi-walker sensor data stream. In this paper, we present our research results on the contextual basis learning and context feature extraction through signal projection in orthogonal subspaces. Particularly, the context identification effects (from signal reconstruction viewpoint) have been investigated a signal projection scheme called non-negative matrix factorization (NMF). Our results have shown the accuracy of context feature learning under a PSN-based multi-walker monitoring scenario.