Date of Completion

5-2-2018

Embargo Period

5-2-2019

Keywords

Opportunistic Sensing, Distributed Sensor Network, Distributed Supervisory Control, Resilient Target Coverage, Energy-efficient Sensor Network

Major Advisor

Shalabh Gupta

Associate Advisor

Peter Luh

Associate Advisor

Ashwin Dani

Associate Advisor

Thomas Wettergren

Field of Study

Electrical Engineering

Degree

Doctor of Philosophy

Open Access

Open Access

Abstract

Distributed Sensor Networks (DSN) containing a large number of sensors nodes are rapidly advancing to perform automated tasks for a variety of applications (e.g. Target Tracking). One of the main problems studied in DSN is the Target Coverage problem, where the objective is to maximize the network lifetime while ensuring that all targets are covered during all times. Therefore, the main challenges in developing a DSN are network control strategies for energy-efficiency and resilience in the event of sensor failures. This thesis presents the \emph{Prediction-based Opportunistic Sensing for Energy-efficiency} (POSE) algorithm, which aims to address these challenges by designing a DSN that minimizes energy consumption while ensuring target coverage. The first algorithm presents a distributed node-level supervisor that controls each sensor node to allow the network to self-adapt to the targets' trajectory via opportunistic sensing. This approach minimizes the network energy consumption by only enabling high power consuming devices when a target is predicted to travel within the sensor nodes coverage area, while using low power consuming devices when a target is absent. The second algorithm, \emph{POSE using Distributed Classification, Clustering and Control} (POSE.3C), extends the POSE algorithm to include classification in the control loop to opportunistically observe targets of interest, while minimizing energy consumption via distributed clustering. Finally, the third algorithm, \emph{POSE for Resilience} (POSE.R), extends the POSE.3C algorithm to incorporate Resilient Target Coverage into the DSN. This approach incorporates target and network density predictions to adapt the distributed clustering method to ensure that a target is covered in the event of sensor failures. The compilation of the three algorithms presents a distributed control strategy for DSN that performs energy-efficient and resilient target coverage.

Download
COinS