Localization Using Low-Resolution Optical Sensors

In wireless sensor networks, the data collected from numerous low end sensors can be equivalent or superior to a few high fidelity sensors. In this paper, we evaluate the validity of this concept with low-resolution optical sensors. We implemented passive localization using visual object identification techniques on a field programmable gate array (FPGA) sensor platform. The complexity, memory utilization, and accuracy of our algorithms were analyzed for performance on a resource constrained environment. We show that low-resolution optical sensors can accurately localize objects on a global two dimensional plane. We also present a model that quantifies fidelity versus scale for optical sensors. Using our model and application, we demonstrate that low-fidelity optical sensors are an effective tool in sensor networks and at a certain scale can provide superior coverage of an area over a high-resolution camera. This analysis lays the groundwork for future advances in optical sensors.