Self-organizing networks [Message from the Editor-in-Chief

Abstract

Wireless sensor networks and mobile ad hoc networks are considered as new communication environments where in general there is no central control on the activity of their users. In particular, users are free to move anywhere in the network, establish a connection, or move out of the network. There is no special binding such as one would see in a cellular mobile network, and as a consequence, the mobility pattern and modeling could be significantly different from those considered in legacy mobile networks. Localization techniques therefore become an important topic of research for these networks. Finding possible locations or future movement patterns of a user in these networks could result in much better resource allocation, packet delivery, and routing than when such information is not available. Typical mobility modeling, such as random walk, can be used in these network, but often they do not provide much intelligence on the operation of the network. Sophisticated mobility models specifically designed for ad hoc and sensor networks are necessary, while they have to place a minimum burden on nodes already low in resources. As an example, vehicular communications are constrained by road and traffic rules; therefore, directional mobility models would better simulate their mobility pattern. The self-organization in vehicular ad hoc networks would be better accomplished and transferred if some parameter-based mobility model is added to the traditional random mobility modeling.