Reliable Wireless Sensor Networks using Multiple Sinks and Degree Constrained Shortest Path Trees.

摘要

Wireless Sensor networks (WSN) have gained attention in both industry and academia due to their versatile fields of application. One of the main characteristics of a sensor node is its limited energy supply. The network needs to be reliable in the sense that it can deliver the data to sink with the presence of multiple link failures. Having more than one sink provides alternative paths to route packets in the presence of link failures and helps load balancing. Degree Constrained Shortest Path Trees (DCSPT) can be used as routing trees to limit the communication of sensor nodes thus conserving energy. In this thesis, we consider the problem of designing logical topologies for routing in WSNs with multiple sinks. We design reliable logical topologies most of which are based on DCSPT in grid and random graphs. We also design scheduling and routing algorithms for the logical topologies and evaluate the reliability of the designs using simulation. We demonstrate that with moderate link failures our schemes can reliably transfer data with low delay and the performance improves as the load of the network decreases.