Distributed Trust-based Routing Decision Making for WSN

Abstract

This thesis describes novel approaches to deal with routing in distributed wireless sensor networks (WSNs) decision making and proposes new distributed protocols based on trust. The trust is defined as the level of belief that a sensor node has on another node for specific action, based on certain criterion that is specified according to applications. As WSNs are applications specific, the proposed trust-based solutions are mainly targeting at two types of network structures, namely, the static homogeneous network, and the network with mobile sink. The first contribution of the thesis is a multi criteria trust model called Hierarchical Trust-based Model (HTM). The model considers several criteria and evaluates the trustworthiness of a node in two levels. HTM is different from most of the existing trust models as it evaluates the trust for multiple nodes rather than a single node evaluation. The model uses the Analytical Hierarchical Process (AHP) in computing the node's trust. The second contribution is a novel distributed trust-based protocol called Adaptive Trust-based Routing Protocol (ATRP). The proposed ATRP embed the proposed HTM in its process. Four network performance metrics (energy, reliability, coverage and reputation) were considered in the forwarder selection. The reputation, which is the accumulated value provided by indirect nodes about evaluated nodes previous communication behaviours is gained using Q-learning. ATRP takes into consideration the resource constrained factors of the nodes by introducing several control mechanisms (timeliness and number of interactions). Thirdly, the thesis considers the implementation of the mobile sink and taken into consideration the relocation issue which is the main concern in existing distributed mobile sink routing. A new distributed mobile sink routing protocol called Blockchain-based Routing Protocol (BCRP) is presented where it adapts the blockchain elements in its relocation decision strategy. The decision in BCRP is determined by other mobile sinks in ensuring the relocation position is not redundantly covered. This is because the redundant coverage in some applications are unnecessary and will consume more energy. The participating mobile sinks are able to make decisions without the central entity's help but based on a set of rules that are pre-agreed by all mobile sinks. The relocation will only happen if it is agreed (verified) by a certain number of mobile sinks. In such situations, the decision making will benefit a larger number of nodes and all nodes are able to get updated information. The performances of BCRP are evaluated and compared under several simulation environments in terms of five performance metrics, i.e., energy consumption, packet delivery ratio, average delay, throughput and coverage level. Based on the simulation results, the proposed approaches outperform the other comparable protocols for all the performance metrics.