Detection and Estimation of Biochemical Sources in Arbitrary 2D Environments

We develop algorithms for biochemical detection and estimation in arbitrary two-dimensional (2D) environments using integrated sensor arrays. The development of biochemical sensors techniques has been a subject of considerable research interest in recent years. In this paper we propose statistical algorithms for automatic monitoring of biochemical agents in a realistically shaped 2D environments using multi-sensor measurements. We derive models for the concentration distribution using the diffusion equation and finite element approximations. Using these results we develop parametric statistical models and maximum likelihood estimation algorithm to find the parameters of the biochemical agent. To detect a presence of the source, we develop a generalized likelihood ratio test. We demonstrate the applicability of our techniques through numerical examples. Our results are potentially useful for national security, environmental engineering, food industry, oil industry etc.