Recently wireless sensor network has emerged as a promising technology that could induce an innovation wave in the field of (infra)structures monitoring because of its fast deployment, little interference with the surrounding, self-organization, flexibility and scalability. A key factor for the proliferation of this revolutionary technology is designing effective protocols to meet the quality of service requirements of the application considering deployment properties and characteristics. Structural condition monitoring using wireless sensor networks can be used for many (infra)structures such as bridge, railways, tunnel, pipelines and highways. These applications exhibit strong similarity in their deployment properties and the way that sensor nodes collect and disseminate their data. Monitoring condition, and operational performance of such large (infra)structures often requires wireless sensor network deployment to long stretch of narrow and elongated spreads which features a linear sensor arrangement and thus its topology resembles a chain. Moreover, ensuring quality of services has been put forward as an essential consideration for wireless sensor networks which are (i) often deployed in unattended and open environments and (ii) characterized by their limited resources and high unreliability. Quality of service in a wireless sensor network can be affected by several constraints out of which (i) the relative position of the node to the base station and other nodes, (ii) the internal reliability state of the network, (iii) the internal reliability state of individual sensor nodes, and (iv) the nodes’ available power, are the most important ones. Quality of service support and guarantees in wireless sensor networks especially for linear wireless sensor networks, is an emerging area of research. In this context, the main focus of this thesis is the design and development of solutions to guarantee combination of four important quality of service parameters, i.e. coverage, long-lifetime, reliability and timeliness for chain-based topology data collection and dissemination. To this end, first we ensure quality of service to some extent at the topology level. However, quality-aware topology control alone is not sufficient to ensure quality of services for disseminating data of many applications whose packets may convey different types or amount of information. Therefore, we concentrate on using dynamic error control schemes which are allocating the correctional power in an on-demand manner based on both the packet-level constraints and channel state. In this way and for the sake of efficiency, we put the amount of information a packet carries or the time-constrained a sensory data imposes and the state in which the channel is in, into perspective with the amount of effort (in terms of energy expenditure) that is required to reliably transmit the given packets. The main contributions of this thesis can be summarized as follows: Trust-based probabilistic coverage: We investigate and address the coverage problem to determine a schedule based on which a selection of the sensor nodes are kept active to efficiently cover the whole monitoring area, using a probabilistic coverage model. By efficient coverage of monitoring area we mean ensuring long network lifetime as well as maintaining sufficient sensing coverage and reliable sensing. Moreover, assuming a probabilistic coverage model we aim to capture the real world sensing and transmitting characteristics of the nodes. In this regard, we propose a trust-based probabilistic coverage algorithm, which leverages the trust concept to tackle the time-varying uncertainties introduced by the sensor nodes and the environment they operate in. QoS-aware Cluster-head/Chain-leader Selection in a Two-tier Architectural model: We propose a well-balanced quality of service aware approach to deliver data packets collected by the sensor nodes to the base station, respecting application requirements in addition to coverage. We address three quality of service parameters, i.e., (i) long-lifetime, (ii) reliability, (iii) delay or data freshness. More specifically in this contribution we (i) introduce a two-tier architecture model in order to energy efficiently, reliably and fast aggregate and disseminate sensed data toward the base station, (ii) integrate the three quality of service parameters (long-lifetime, reliability, and delay) with the possibility to adjust their priorities according to the specific application requirements. QoS-aware Dynamic Chain-Cluster Forming: In order to relax some assumptions we made before regarding communication capability of the nodes to communicate directly with other nodes or with the base station as well as the fixed-size of the chain-cluster, we propose two solutions which make the size/shape of the clusters adaptive regarding the state of the nodes and links. The proposed solutions well-incorporate energy, delay and transmission reliability together to construct clusters and to select proper cluster heads in each cluster. Reliable Dissemination of Time-Constrained Data: Meeting the Time-To-Live (TTL) constraint of the sensory data which should reliably be transmitted toward the base station in a low duty-cycle network that suffers from short-term burst errors is the main focus of this contribution. By short-term burst errors we mean the errors which are localized in short-term and occurs in burst forms. In this respect, we propose a runtime adaptive packet-link-local error control scheme that operates based on the links’ qualities, packets’ TTL, and duty-cycle and is able to counteract periodic short-term burst-errors in a chain topology. Information-link-aware Data Dissemination: In the same line of the previous contribution which considered the TTL as one of the packet-level indicator or constrains to ensure quality of service, in this contribution we concentrate on the information-value or amount of information a packet carries as another packet-level indicator. In this way, we propose a Run-time Adaptive FEC-based data dissemination protocol. In the proposed approach, each node decides which error control code to use abiding to the computational constraints of embedded sensors, the information-value of the packet, and the statistical properties of the observed errors for the upward link. This adaptation gives the possibility to vary the code strength and complexity on-demand and on the fly.
展开▼