The project considers using a variety of sensors to monitor an unknown signal field over a spatial domain with unknown geometry, which may bear elevational variations. The main objective of this project is to develop mathematical algorithms that can integrate sensor readings from a large number of distributed devices and reconstruct both the signal field and the terrain. Specifically, the sensors involved in the project include the Light Detection And Ranging technology (LIDAR) and networked wireless sensors, for their complementary capabilities. LIDAR uses a constant number of powerful sensors, each takes an image from a distance. On the other hand, distributed wireless sensors use a large number of inexpensive sensors (possibly piggybacking on cellular phones), each takes a density reading at its position. The approach in this project is to use conformal and hyperbolic geometry, discrete curvature flows, topology and numerical analysis to develop novel algorithms to generate accurate density map. From the networking perspective, this project addresses the problems of LIDAR image fusion, network localization, sensor deployment, and integration of LIDAR and sensor data.
The project is motivated by the potentially devastating chemical terror attacks. Biological and chemical agents, used by adversaries, could potentially spread to a large region in a short time. This project focuses on development of algorithmic tools and techniques for gathering timely, accurate and useful information about the chemical or biological spread in case of such an attack. By exploiting the geometric properties of LIDAR data and sensor network data, the project provides critical observations on fundamental questions of how to organize such a large scale network; how to manage sensor data; and how to use the network for acting on the environment and the users.
