Restricted Access

Reason: Access restricted by the author. A copy can be requested for private research and study by contacting your institution's library service. This copy cannot be republished

Reliable data delivery protocols for underwater sensor networks

thesis
posted on 22.02.2017, 03:08 authored by Nowsheen, Nusrat
Underwater Acoustic Sensor Networks (UASNs) are becoming increasingly promising to monitor aquatic environment. The network is formed by deploying a number of sensor nodes and/or Autonomous Underwater Vehicles (AUVs) to support diverse applications such as pollution monitoring, oceanographic data collection, disaster recovery and surveillance. These applications require transmission of data packets from the source to a sink or gateway in a multihop fashion and eventually to a message ferry or an onshore station. In such a network, reliable data delivery becomes challenging as delivery is hindered due to long propagation delay and high error-rate of underwater acoustic channel, limited energy and inherent mobility of sensor nodes. Existing data delivery protocols have certain limitations such as predefined communication architecture, dependency on periodic communication with neighbors or flooding-based data forwarding. These restrict their capabilities to achieve reliable delivery in many application scenarios. They mainly consider location, depth and energy information of sensor nodes to improve energy efficiency of underwater networks. However, to improve reliability, protocols should consider factors such as link quality, node movement, packet delay and traffic flow so that they can deal with the disruptive nature of the communication link, and harsh and dynamic context of the network. In this thesis, we focus on developing data delivery protocols to provide reliable data transfer for various underwater applications and network scenarios considering underwater communication characteristics that have direct impact on reliable data transfer. In the first protocol, an anchor-based architecture is considered to support underwater monitoring applications and mobile message ferry nodes are used to collect stored data from one or more nodes, called gateways. A strategy for gateway selection is devised to maximize their lifetime. Data delivery reliability is significantly improved by selecting the next hop forwarder on-the-fly based on its link reliability and reachability to gateways through probabilistic estimation. Data forwarding solution is coupled with delay tolerant networking paradigm by accumulating data packets for a fixed amount of "hold time" to improve delivery and reduce communication overhead simultaneously. Besides high error probability and dynamic nature of underwater acoustic channel, node movement also makes data forwarding process difficult. Moreover, consideration of network condition in the calculation of "hold time" captures more recent network status. The first protocol is therefore further extended by (i) estimating/predicting node movement pattern and using it to determine neighbors' coverage probability, and (ii) calculating data "hold time" dynamically at each node. Consideration of the above factors increases the reliability of data forwarding solution. The above protocols have been designed to improve reliability of data transfer for underwater monitoring applications. However, in many situations like disaster management and emergency rescue operations in ocean, deployment of an ad hoc network consisting of AUVs is of paramount importance to support mission critical applications. To accomplish this, an effective data delivery protocol has been devised which ensures reliability by delivering data within a given deadline to the onshore station using a set of AUVs. A modified Breadth-First Search (BFS) technique has been employed to find a path towards an AUV that meets the data delivery deadline as per a given surfacing schedule. A command-based node movement technique has been incorporated to ensure delivery in an energy efficient manner. The proposed data delivery protocols have been evaluated and compared with existing relevant and contemporary underwater protocols. Simulation results show that the proposed protocols achieve performance improvement over competing protocols in terms of successful packet delivery, communication overhead and energy.

History

Campus location

Australia

Principal supervisor

Gour Karmakar

Additional supervisor 1

Joarder Kamruzzaman

Year of Award

2015

Department, School or Centre

Gippsland School of IT

Course

Doctor of Philosophy

Degree Type

DOCTORATE

Faculty

Faculty of Information Technology