Optimal Sleep/Wake Frequency Assignment for Data Gathering in a WSN

Abstract

Power conservation of sensor nodes is one of the most important issues for data gathering in Wireless Sensor Networks (WSNs). The most general and effective method for power conservation is the use of the periodic sleep-wake cycles with long periods of sleep followed by short wake-up periods. This dissertation proposes two different wake-up frequency assignment algorithms that assign a different wake-up frequency to each node with two objectives: the minimization of the total power consumption and the maximization of the network lifetime, respectively. The first optimal wake-up assignment algorithm, OWFA_MINP, minimizes the total power consumption of the data gathering tree while limiting the total delay for delivering data from any node in the tree to the sink node. Analysis is used to prove that OWFA_MINP assigns optimal wake-up frequencies for minimizing the total power consumption of all nodes. A variant, MP_OWFA_MINP is also proposed for multiple-path trees. Simulation results showed that the wake-up frequencies assigned by OWFA_MINP and MP_OWFA_MINP resulted in less average power consumption than two alternative methods under all simulated conditions. The second optimal wake-up frequency assignment algorithm, OWFA_MAXL, maximizes the network lifetime of a data gathering tree while limiting the total delivery delay. The optimization problem for maximizing network lifetime is formalized as a general convex optimization problem. Then, OWFA_MAXL solves the problem using a convex optimization problem solver. The simulation results showed that the wake-up frequencies assigned by OWFA_MAXL and MP_OWFA_MAXL resulted in longer network lifetime than three other alternatives. Coordination of sleep-wake cycles requires time synchronization of the nodes in the WSN. Thus, in addition to the wake-up frequency assignment algorithms, a new time synchronization algorithm, termed Fast Fault-tolerant Time Synchronization (FFTS), is presented. Analysis is used to quantify the resynchronization interval required for a desired level of time synchronization precision and the protocol to be used to achieve a desired level of fault-tolerance. Experimental results show that the proposed algorithm works better than the best previously proposed algorithms in terms of the level of time precision achieved.