Controlled Mobility for Event Data Collection within Wireless Sensor Networks.

摘要

Energy/power has been a significant concern within wireless sensor networks, as normally applications require the sensor networks to operate for a long period of time while sensor nodes are powered through limited battery. As a result, various approaches have been proposed to deal with this concern including energy harvesting, energy-efficient operation, employing mobile nodes as data collectors and so on. Mobility though traditionally thought as an overhead, has been explored to bring assistance to wireless sensor networks. Besides acting as data collector, mobile nodes are possible to take other roles like data relay, common sensor nodes, energy deliverer etc.;This thesis proposes employing controlled mobility to serve for event data collection purpose in wireless sensor networks. Different from previous formulation, it formulates the problem as DVRPTW (Dynamic Vehicle Routing Problem with Time Window). We believe that DVRPTW is a more realistic formulation by taking dynamic occurrences of events into consideration. Two-step algorithm transfers DVRPTW into a series of VRPTW and existing heuristic algorithm tabu search, 2-opt*/or-opt is used to solve the VRPTW. To save energy, we decrease the size of data to be transmitted by exploring the similarity among event data and propose a straightforward modeling of event by treating each event region as a single customer. An accompanying data collection mechanism called "on the way" data collection is proposed. It means that mobile data collector not only collects data from the representative node, but also collects data from all reachable sensor nodes on its way.;To reduce the long latencies introduced by the mechanical movement of the mobile data collector, mobile data collector is equipped with both Wi-Fi interface for communicating collected data to base station and ZigBee interface for communicating with sensor nodes.;Substantial simulation results under NS2 show our formulation has advantages over random movement and multi-hop transmission in terms of energy saving, balance of energy distribution etc.