Energy-efficient secure and anonymous communication protocols for wireless sensor networks
Large wireless sensor networks (WSNs) are distributed networks containing hundreds or even thousands of small sensor nodes (also called motes) that monitor temperature, humidity, sound, vibration, pressure, motion, pollutants and other environmental signals and communicate with one another using wireless channels and ad hoc networking protocols.
In this dissertation, we address several important research problems such as synchronization techniques, energy consumption models and network lifetime estimation models, location privacy techniques and dynamic scheduling techniques to improve response times, which need to be solved for widespread use of WSNs.
We propose a low overhead coarse-grain synchronization technique to keep nodes synchronized despite short active and long sleep durations. This technique allows nodes wake up from sleep in a staggered manner with those farther from sink waking up first. We develop detailed energy consumption and network lifetime estimation models that are validated using extensive experiments of WSNs formed from off-the-shelf sensor motes. We use these energy models to enhance the widely used TOSSIM sensor network simulator. We use the analytical and simulation models to study the impact of secure data aggregation methods on network lifetime.
We develop a probabilistic model to analyze the capture time, which is the time for an adversary to successfully identify the location of a source node. Motivated by the fact that the capture times of current anonymous communication techniques are much shorter than the theoretical upperbounds, we propose an anonymous communication protocol (ACP) to protect source node privacy against global adversaries by making all sensor nodes transmit exactly once in each cycle. Simulations show that the effectiveness of an adversary, when ACP is used, is no better than randomly picking and checking nodes in the network regardless of the resources and capabilities of the adversary. Our detailed energy consumption analysis shows that the additional transmissions of sensor nodes cost only a small fraction of the total energy consumption of sensors, and have very little impact on a sensor node's lifetime.
We extend ACP to address sink location privacy issue. We propose a sink anonymous communication protocol (SACP) to protect sink node privacy against global adversaries. The major difference between ACP and SACP is that SACP requires one more synchornization step to hide the sink among regular sensor nodes, such that all nodes, including the sink, follow a doubly staggered sleep scheduling pattern. We analyze the effectiveness of SACP using our capture attempt model as well as other existing anonymity analyses in literature. In particular, we show that SACP achieves total anonymity with only a small increase in the energy consumed by the nodes.
Finally, motivated by the recent progress in energy harvesting technology, we propose a self-synchronized dynamic scheduling protocol (SDS) to reduce event response time and ensure high delivery ratio while balancing energy consumption of all nodes in the network. To simulate SDS, we develop a realistic energy harvesting model based on time and locations of sensor nodes. Then we compare SDS with a recently proposed technique, called dynamic-cycle scheduling based on residual energy through extensive simulations. The results show that SDS ensures a much higher delivery ratio with a better end-to-end delay because it makes nodes synchronized throughout the whole network lifetime.
The combination of the completed research work will provide comprehensive solutions for anonymous communication and energy-optimized operations of sensor networks.