Andrew Howard, Maja J. Mataric, and Gaurav S. Sukhatme
This paper briefly sketches a pair of algorithms for deploying and localizing a mobile sensor network. We use the term 'mobile sensor network' to describe a distributed collection of nodes, each of which has sensing, computation, communication and locomotion capabilities. For deployment, we have developed a potential-field-based approach that ensures that any compact initial configuration of nodes will spread out such that the area 'covered' by the network is maximized. For localization, we have developed an approach that makes use of the nodes themselves as landmarks. Through a combination of maximum likelihood estimation and numerical o13- timization, we can, for each node, estimate the relative range, bearing and orientation of every other node in the network. This paper sketches the basic formalism behind these algorithms, and present some experimental results.