Computationally Tractable Approaches to Modeling the Coverage of Weather Sensor Networks

Computationally tractable methods are required to identify the potential contribution of sensors. This paper presents computationally tractable methods to measure radar coverage. Two measures are introduced. One is covered voxels which measures an upper bound of the number of distinct high quality values that can be returned from the radar system for a weather feature. The other is the high quality feature coverage area percentage that defines what percentage of the weather feature is sensed by covered voxels. Five methods (2 2D and 3 3D algorithms) are presented to help calculate the data required to determine the values for these measures. The algorithms are used on two systems to test drive the framework including run-time performance measures for the algorithms themselves. For both 2D and 3D cases, the approximations helped with respect to computation time. In addition, the feature approximation/3D volume coverage algorithm (using 3D-grid) greatly reduced the computation time.