Energy-efficient Information Collection and Dissemination in Wireless Sensor Networks.

Abstract

Low-power sensors, equipped with limited capability of sensing, computation, and communication, and limited battery power, can form an ad hoc network, providing versatile scopes into both the physical world and human spaces.

For information collection and dissemination in wireless sensor networks (WSNs), a key challenge is the limited energy in each sensor node. Although the sensor hardware platforms have been designed with low-power components, achieving high application performance while minimizing energy consumption is very challenging.

To meet this challenge, this thesis proposes energy-efficient protocols and systems for WSNs. First, we develop an opportunistic in-network data aggregation scheme for WSNs, which can compensate for communication losses without incurring computation error, and achieve better energy efficiency than other approaches that rely on retransmission.

Next, we propose a data-centric approach for post-deployment performance debugging in WSNs. It focuses on the data flows a WSN application generates, and relates poor application performance to significant data losses or latencies of some data flows (i.e., problematic flows) as they go through the software modules on individual nodes and through the network. It collects debugging information from only those modules and nodes that the problematic flows go through, and therefore consumes much less energy than collecting debugging information indiscriminately from all nodes.

WSNs sometimes contain mobile data sinks. Maintaining high packet delivery ratio becomes more difficult as the network gets bigger and mobile move faster. Thus, we design a distributed location service protocol which can sustain a high packet-delivery ratio in larger networks and at higher mobile speeds, and incur lower energy consumption than other location service protocols. We also determine how to configure the protocol parameters to ensure the scalability of the location service.

Last, we explore the tradeoff between location privacy protection and energy consumption in people-centric sensing applications. The key idea of our approach is to prevent the distinguishability of a mobile user's points of interest (POIs), therefore weakening attackers' ability to infer the user's private information or mobility patterns. Our evaluation using real-world traces shows that it protects mobile users' location privacy effectively and consumes battery of their handhelds efficiently.