Exploration and mapping of a geographical area is performed according to a specific predictable strategy. Mobile hosts should coordinate their moving patterns in such a way that connectivity is preserved. One simple example is to have a group of hosts moving all in one direction away from a home-base host. Before the communication signal between the group and the base degrades (due to the distance), one mobile host stops and serves as a communication bridge to the base. The rest of the group keeps exploring in the same direction and this procedure is repeated as needed. When desired, the sequence of mobile hosts can start sweeping the complete region around the base, by moving coordinately around the base. By employing multi-hop technology along with predictable moving patterns connectivity can be preserved.
The distance between hosts is important to preserve connectivity, however in many situation the explicit location of a host is essential to perform some task. For instance, a mobile host could inform other hosts to avoid walking towards a hazardous region, such as a cliff or a hill. Ad-hoc local positioning systems can be implemented by strategically placing hosts in a triangular fashion. Positioning relative to the LPS can then be obtained. Multiple LPSs may be required to cover a relatively large area.
In a sensor network, the primary goal is to send reports (images, analyzed data) to a control center. In such system we could assume that all mobile hosts are able to upload the acquired information to the central control. However, it is more likely to have a heterogeneous population of mobile hosts, where some of them, called transmitter hosts, are specialized in storing data and uploading collective data to a central controller, while others, called sensor hosts, perform chemical and physical experiments and report their findings to the transmitter hosts. One interesting problem is how to find the closest transmitter host to a sensor host. Again, the exploration strategy can be used to assist in this matter. Actually, transmitter hosts should be an intrinsic part of the exploration strategy. Transmitter hosts can also coordinate among themselves so that some of them could be idle while others would be active, in a tentative to save power.
Sensor and transmitter hosts should be able to assign and process packets with differing priorities. Hosts can either have a pre-determined knowledge of each information's importance, or they can dynamically learn to give higher priority to certain information such as alarm and hazardous conditions, or new discoveries (such as finding water or gold in Mars). High priority information should be assgned express service and should be sent to the central control station as fast as possible.
Power hosts can be strategically placed to absorb solar energy and to serve as power sources to mobile hosts. Hosts would periodically recharge themselves and the life-time of the sensor network can be extended. Locating power hosts is a similar problem to locating the closest transmitter host. This problem becomes really interesting when we assume that each mobile hosts is completely autonomous and the moving startegy of the mobile population cannot be predicted. In such situation, preserving connectivity is the primary factor determining how mobile hosts move.