Spatiotemporal Delay Control for Low-Duty-Cycle Sensor Networks

Abstract

Abstract-Data delivery is a major function of sensor network applications. Many applications, such as military surveillance, require the detection of interested events to be reported to a command center within a specified time frame, and therefore impose a real-time bound on communication delay. On the other hand, to conserve energy, one of the most effective approaches is to keep sensor nodes in the dormant state as long as possible while satisfying application requirements. Obviously a node can not communicate if it is not active. Therefore, to deliver data in a timely manner for such extremely low duty-cycle sensor networks, communication needs to be carefully managed among sensor nodes.In this work, we introduce three different approaches to provide real-time guarantee of communication delay. First, we present a method for increasing duty-cycle at individual node. Then we describe a scheme on placement of sink nodes. Based on previous two methods, we discuss a hybrid approach that shows better balance between cost and efficiency on bounding communication delay. Our solution is global optimal in terms of minimizing the energy consumption for bounding pairwise endto-end delay. For many-to-one and many-to-many cases, which are NP-hard, we propose corresponding heuristic algorithms for them. To our knowledge, these are the most generic and encouraging results to date in this new research direction. We evaluate our design with an extensive simulation of 5,000 nodes as well as with a small-scale running test-bed on TinyOS/Mote platform. Results show the effectiveness of our approach and significant improvements over an existing solution.

Utility-Based Routing in Wireless Sensor Networks

“…However, in the low duty-cycle WSNs, the graph changes over time, which is called a "timedependant graph". G = (V, E(t)) [21] is used to represent the models, where V is a set of nodes, and E(t) is a set of edges that appears at time t. Several works have been conducted using time-dependant graph models.…”

“…In [21], Γ i = (ω i , τ ) is used to represent the working schedule of node i. ω i is an infinite binary string in which 1 denotes the active state, and 0 denotes the dormant state. τ denotes the time span for each bit.…”

“…In low duty-cycle WSNs, the sensors are scheduled to be either in active or sleep mode to achieve low-energy consumption [21,11]. When the sensors are in sleep mode, they cannot transmit the packets.…”

Multiple task scheduling for low-duty-cycled wireless sensor networks

“…In [8], Guo et al study opportunistic flooding for low-duty-cycled sensor networks with unreliable links. Gu et al propose a method for increasing duty-cycle at individual node, and a scheme on placement of sink nodes to provide real-time guarantee of communication delay [7]. A dynamic data forwarding (DSF) scheme is presented in [6] with experiments conducted on low-duty-cycled WSNs.…”

“…To bridge the gap between limited energy supplies of the sensor nodes [6] [8] and the system lifetime, many research studies suggest the WSNs operated in low-duty-cycle mode [8] [7] [6]. In low-duty-cycled sensor networks, a sensor node keeps its radio on for a small percentage (e.g., less than 5%) of time during each working period.…”