Cardei, Mihaela Florida Atlantic University
Wireless Sensor Networks (WSNs) are recognized as a new frontier in networking. Deployed close to the phenomenon, a WSN provides a global view of the state of the phenomenon based on local sensor measurements. Sensor nodes have size, weight, and cost restrictions that limit resource availability, such as battery, CPU, and storage resources. These constraints severely affect application capabilities and performance.
This research project seeks to establish theoretical and practical foundations for fundamental optimization problems along the following research directions: sensor coverage, topology design and control, and efficient data gathering. In sensor coverage, this research will focus on new theoretical and practical advancements for the maximum set covers problem, as well as new concepts regarding connected sensor coverage and coverage problems for heterogeneous sensor networks and tiered network architectures.
Energy-efficient topology design and control is an important operation in WSN design. This research will investigate theoretical and practical mechanisms for range assignment in WSNs that minimize power consumption. For a general architecture with multiple sinks, this project will address mechanisms for deterministic sink locations and energy-efficient anycast topology-control. In support for applications that require high reliability, the k-degree anycast topology control problem will be investigated.
An important optimization problem in the design of energy-efficient data gathering involves selecting a minimum set of data forwarding nodes. This research work will investigate this problem when each forwarding node is connected to at least one of the deployed sinks. This project will also study the problem of reliable data delivery to one of the sinks in case of failure of at most k sensor nodes. Activity scheduling is another research direction that will provide mechanisms of efficiently rotating the set of forwarding nodes in order to balance the energy consumption. This research will also investigate efficient mechanisms for cluster-based data gathering using multiple sinks.
This project is expected to have a lasting impact on the theory and practice of wireless sensor research, by providing integrated theoretical and practical foundations in sensor coverage, topology design and control, and data gathering. This research work will address these problems on two fronts. First, this project seeks to design and analyze solutions with provably good theoretic results, such as approximation algorithms and PTAS. Such results are very important and they will help understand system performance limitations through worst-case guarantees. Second, for practical implementations, this project will focus on the design of low-complexity distributed and localized algorithms. In addition, this project will provide extensive simulations for performance evaluation.
This program will integrate research and education through student participation in research projects, student mentoring, research seminar, and course development. Florida Atlantic University is a minority serving institution. This project will benefit minority students in performing research in the area of theoretical and algorithmic WSNs, offering them the opportunity to stay competitive with peers at other institutions, locally and nationally. The results from this research will be included in class materials and will be disseminated via web pages, invited talks, conference presentations, and journal publications.
Wireless sensor networks (WSNs) is an emerging research field, with many potential applications. Our research work contributes important aspects from WSNs such as (1) energy efficient design for the coverage of a number of targets such that to maximize network lifetime, (2) support for both coverage and connectivity using a power constrained WSNs, (3) study the benefits of heterogeneous WSNs, (4) study topology design and control in WSNs, (5) study mechanisms to balance energy consumption in a WSN by adding additional relay nodes, (6) design mechanisms for sensor repositioning such that to avoid energy holes around sinks, (7) study mechanisms for sinks repositioning such that to avoid energy holes, (8) study of connectivity and coverage under composite event detection in sensor networks, (9) energy-efficient data gathering with QoS support, and (10) application of WSNs to hurricane evacuation. We address the optimization problems on multiple fronts, as follows. First, we are concerned to determine whether these optimization problems are P or NP-complete problems. For NP-complete problems, we seek to design polynomial time approximation algorithms that guarantee the worst-case behavior. We designed approximation algorithms for the bottleneck Steiner tree problem (placing additional relay nodes to balance energy consumption) and for the problem on topology control using cooperative communication. For practical applicability of our algorithms on large scale WSNs, we are concerned with designing low-complexity algorithms that are distributed and localized. We contributed localized algorithms for most of the sensor coverage and topology-control related optimization problems. After the design and theoretical analysis of our algorithms, we analyzed their performance through simulations.
