Large-scale online environmental mapping: from theory to implementation

Abstract

Within the context of environmental studies, environmental mapping addresses the monitoring and analysis of physical processes appearing in the nature, such as atmospheric or oceanic transformations. In such studies, often monitoring is fulfilled by deployment of sensor infrastructures and analytics are legitimized by the existance of statistical models known to represent the unknown nature of the physical process. Despite a great number of researches existing in the field of environmental mapping, development of analytical frameworks that well-utilize the advanced sensor platforms along with theoratically proven novel mathematical ideas to address the existing research gaps is a promising area of effort. Therefore, the objective of this dissertation is evolved around development of novel frameworks to address the analytical shortcomings of the previous researches in environmental mapping by employing mathematical and statistical techniques on proper sensing platforms. Our proposed frameworks incorporate the stochasticity of the model underlying the physical processes in an online intelligent manner to progressively and adaptively build an accurate representation of the physical process based on modern sensing techonologies in a cost-effective practice. Our research asserts the validity of the proposed frameworks through both theory and implementation.