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 sensor techniques has been a subject of considerable research interest in recent years. We propose statistical algorithms for automatic monitoring of biochemical agents in 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 a maximum likelihood estimation algorithm to find the parameters of the biochemical agent. To detect the presence of a 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.