Pushing the Limits of Wireless Sensor Networks - WSNs 20,000 Leagues Under the Sea

In recent years, research in wireless sensor networks (WSNs) has made significant strides, facilitating a variety of applications in the Internet of Things, health, environmental monitoring, and military. Advances in wireless routing and delay tolerance design allow WSNs to achieve data acquisition in networking scenarios with intermittent link and time varying bandwidth. Creative power saving schemes were also developed to allow WSN nodes periodically power down. These advances allow researchers to deploy WSNs in a variety of challenging environments that have limited communication bandwidth with harsh channel dynamics, long delay, and high deployment costs. One such environment is under the ocean, where unconventional types of propagation medium such as acoustics and optics are the only viable types of communication medium.This dissertation describes the design, implementation, and various other aspects of an underwater WSN system using software defined networking (SDN) principles that can be deployed on the ocean floor for purposes ranging from ocean exploration and oil drilling needs to search and rescue missions and military interests. The system consists of a flock of Autonomous Underwater Vehicles (AUVs) acting as mobile WS nodes and primarily relies on a centralized SDN controller doubling as the information hub and recharging station for the AUVs. We address systematic issues related to AUV deployment and networking, including power consumption, data transfer, and channel contention. To limit power consumption, we design two neighbor discovery methods that allow AUVs to turn off their radios and thus save power. For data transfer among nodes and channel contention, we explore the idea of delay-tolerant networking (DTN) and study various acoustic media access control schemes. We also develop a new underwater acoustic MAC protocol that maximizes throughput in good channel conditions and data transfer reliability in bad channel conditions. Simulations and experiments show that our protocol greatly outperforms existing underwater acoustic protocols in both speed and reliability. Along the process, we create and expand a public testbed allowing researchers to run underwater networking experiments in a water tank, in an emulator, and even in the open water for more realistic settings