Improving geometric distance estimation for sensor networks and unit disk graphs

Distance measurement between nodes in wireless sensor networks is a prerequisite for a variety of applications and algorithms. However, special hardware allowing such measurements is expensive, especially if dealing with hundreds or thousands of nodes. Fekete et al. presented an approach on distance estimation based on only the neighborhood information available to all nodes in the network. We improve this algorithm, such that it does no longer rely on uniformly distributed nodes. For our approach, it is sufficient that the second derivative of the probability distribution function is a constant.