As an emerging area, underwater acoustic networks have attracted rapidly growing interests in last several years. On the one hand, underwater networks enable a wide range of aquatic applications, such as oceanographic data collection, pollution monitoring, offshore exploration, disaster prevention, and tactical surveillance applications. On the other hand, the adverse underwater environments pose grand challenges for efficient communication and networking.

In underwater acoustic networks, nodes are usually powered by battery, which is hard, if not impossible, to be replaced in practice. For long-term applications such as environment monitoring, the node is expected to work continuously for a long time, from several days to several years. Thus, energy efficiency becomes one of the most important design considerations. In addition, underwater acoustic channels feature long propagation delays and low available bandwidth. Communication efficiency is also of paramount importance to a practical underwater network. In this dissertation work, we tackle the efficiency problem for underwater acoustic networks from three important aspects: Medium Access Control (MAC), reliable data transfer and distributed localization. Specifically, three research thrusts are included in this dissertation work: 1) Efficient MAC protocols; 2) Efficient reliable data transfer schemes; 3) Efficient distributed localization protocols.

For the efficient MAC protocols, we take a novel multi-channel approach to combat the severe underwater environment. We model and analyze two generalized multichannel MAC protocols: multi-channel access with Aloha and multi-channel access with RTS/CTS on a dedicated control channel. Based on our analysis, for the first time, we identify the triple hidden terminal problem which is special to the multichannel underwater acoustic network. We propose a new Cooperative Underwater multi-channel MAC protocol (CUMAC), which can effectively solve the triple hidden terminal problem and improve the system efficiency.

Secondly, delay critical applications, such as submarine detection, have strict requirements on time delay and end-to-end reliability. Conventional methods such as retransmission-upon-failure cannot satisfy both requirements effectively. We propose a novel efficient end-to-end transmission scheme, called Multi-path Power-control Transmission (MPT). MPT can guarantee certain end-to-end packet reliability while achieving a good balance between the overall energy efficiency and the packet delay. Since no retransmission is allowed in MPT, it is efficient in terms of both energy and bandwidth. Simulation results show that MPT can greatly improve the system efficiency.

Thirdly, we investigate distributed localization protocols for underwater acoustic networks. Since localization service is indispensable for many applications and networking functions, it affects the network performance greatly. By investigating the temporal and spatial correlations of the mobile underwater objects, we propose a new scalable hierarchical localization scheme, called Scalable Localization with Mobility Prediction (SLMP). SLMP is the first protocol which utilizes the mobility pattern of underwater objects and employs an advanced mobility prediction algorithm. Simulation results show that SLMP can greatly improve the localization accuracy with much less communication costs.