Optimal Message Bundling with Delay and Synchronization Constraints in Wireless Sensor Networks

Abstract

Energy efficiency and end-to-end delay are two of the major requirements for the monitoring and detection applications based on resource-constrained wireless sensor networks (WSNs). As new advanced technologies for accurate monitoring and detection—such as device-free wireless sensing schemes for human activity and gesture recognition—have been developed, time synchronization accuracy becomes an important requirement for those WSN applications too. Message bundling is considered one of the effective methods to reduce the energy consumption for message transmissions in WSNs, but bundling more messages increases the transmission interval of bundled messages and thereby their end-to-end delays; the end-to-end delays need to be maintained within a certain value for time-sensitive applications like factory monitoring and disaster prevention, while the message transmission interval affects time synchronization accuracy when the bundling includes synchronization messages as well. Taking as an example a novel WSN time synchronization scheme recently proposed for energy efficiency, we investigate an optimal approach for message bundling to reduce the number of message transmissions while maintaining the user-defined requirements on end-to-end delay and time synchronization accuracy. Formulating the optimal message bundling problem as integer linear programming, we compute a set of optimal bundling numbers for the sensor nodes to constrain their link-level delays, thereby achieving and maintaining the required end-to-end delay and synchronization accuracy. Extensive experimental results based on a real WSN testbed using TelosB sensor nodes demonstrate that the proposed optimal bundling could reduce the number of message transmissions about 70% while simultaneously maintaining the required end-to-end delay and time synchronization accuracy.

Energy-Efficient Message Bundling with Delay and Synchronization Constraints in Wireless Sensor Networks

“…If node i is not a leaf node and has descendant nodes, as shown in Figure 1 , the period for generating a packet bundling messages should be from the conservation of traffic flows [ 17 ]. Here, we assume that, at each non-leaf node, the messages of incoming packets are unbundled first and bundled again with the measurement messages generated at the node before being transmitted via outgoing packets.…”

“…While minimizing the energy consumption, we also need to meet the E2E delay and time synchronization accuracy requirements for time-sensitive applications and the proper operation of WSNs. In [ 17 ], the joint constraints of E2E delay and time synchronization accuracy are formulated as follows: where is the E2E delay of node i given by is a user-defined E2E delay requirement, and is another E2E delay requirement translated from the minimum time synchronization interval through the function , i.e., …”

“…In [ 17 ], for the first time, we took into account time synchronization accuracy and E2E delays in optimal message bundling, where the energy consumption is indirectly minimized by maximizing the total number of bundled messages in a WSN through computationally feasible integer linear programming (ILP). Unlike most of the existing work, in this study, we purely focus on the optimal message bundling, independently from other protocols, such as routing and scheduling.…”

“…Though the indirect minimization of the energy consumption dramatically simplifies the formulation of the optimal bundling, it cannot capture the negative effect of message bundling on the energy consumption, i.e., a higher probability of retransmissions due to the increased packet error rate (PER) [ 6 ]. In this paper, we formulate the optimal bundling problem with a new objective function, thus directly modeling the total energy consumption of a WSN, to address the issue of the optimal message bundling scheme proposed in [ 17 ], which takes into account both the positive and negative effects of message bundling on the energy consumption and, thereby, strikes the optimal balance between the number of bundled messages and the number of retransmissions given the link quality. We solve the resulting integer nonlinear programming (INLP) problem using a self-adaptive global-best harmony search (SGHS) [ 18 , 19 ].…”

“…Second, we carry out a comparative analysis of the optimal bundling schemes, where we compare the energy consumption of the three different schemes—i.e., the proposed scheme, the ILP model of [ 17 ], and the baseline of no bundling—based on the Contiki-NG and its Cooja emulator with eight different network topologies. We also calculate the power consumption of the three schemes based on a numerical analysis for the verification of the emulation results.…”