Localization Based on Feasible Geographic Region Approximation

Most traditional localization approaches, providing single position estimate for each sensor node, cannot evaluate localization accuracy and position uncertainty. In this study, we construct a constraint set using proximity information ,and then, compute feasible geographic region (FGR) for each sensor node by projecting high dimensional feasible region of the constraint set to a specific 2D plane. The FGR approximated by simple and expressive polygon could bound exact sensor node position and clearly reflect position uncertainty. A distributed algorithm is also proposed to compute FGR effectively. In addition, considerable improvements of localization accuracy can be made if we use polygon with more number of vertices to approximate FGR, or utilize non-convex range constraints to locate infeasible holes within the polygon.