Motion-constrained mobile sensor networks

The addition of mobility to sensor networks gives users the ability to efficiently monitor environments about which little is known ahead of time. For example, if there is a suspected path through the forest that should be monitored, but its exact location is not known, sensors can be scattered through the area and converge on the path as people or animals are detected on it. In this work, we present several distributed algorithms for sensors to produce this convergence. The first is a robust positioning technique for open rectangular spaces. We also build on this to produce an algorithm for arbitrary connected spaces. In addition, we consider sensors that are inherently constrained in their motion, such as tethered sensors fixed to a single line of motion. We present both analytical and empirical results for a variety of cases and discuss extensions that will provide greater effectiveness