Target Coverage Heuristics Using Mobile Cameras

The availability of low-cost mobile robots with sensing, communication, and computational capabilities has made feasible a new class of Cyber-Physical Systems (CPS). These miniCPSs may be used where quick, low-cost or non-lasting visual sensing solutions are required, e.g. in border protection and disaster recovery. In this paper, we take the first steps towards a fast and automated CPS. We consider the problem of low complexity placement and coordination of an unknown number of mobile cameras to cover arbitrary targets. We address this problem as an unsupervised clustering task. A set of proximal targets are clustered together, whereas the camera location/direction for each cluster are calculated/estimated individually. Our proposed solutions provide centralized computationally efficient heuristics using two clustering-based algorithms: k-camera clustering, and cluster-first algorithms. Our evaluation shows that the required number of cameras approach those obtained via near-optimal methods as the cameras’ coverage range, angles of view, or the number of targets increase.