الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 138 |  |  |
| | | from | 138 | | |
Abstract
The Internet of things (IoT) is a global network of numerous connected appliances that are able to communicate with each other and with the surrounding environment. Enhancing the quality of life by improving and providing a wide range of services and applications is the main goal of this new paradigm. A key technology that enabled establishing and maintaining reliable IoT services is the wireless sensor network (WSN) technology. WSNs comprise a large number of easily deployed, cheap. small-sized and multi-functional devices, known as sensor nodes. These sensor nodes monitor the characteristics of the surroundings such as the temperature, humidity, vibration, etc., and are equipped with an on-board processor for computations. Thus, WSNs have been applied in civilian, industrial and military applications. Most of the IoT services are location based, also the developed information by the deployed sensor nodes from the gathered data will be meaningless if they are not accompanied by the locations from where they were gathered. Also, the location of sensors is important for the operation of successful routing and clustering in the WSNs. Hence, obtaining the location of sensor nodes, that is localization, is of utmost importance in WSNs. This study is concerned with the localization problem in WSNs. Instead of equipping all the deployed sensors with a Global Positioning System (GPS), which is neither a cost nor an energy efficient solution for the localization problem in WSNs, only a few location-aware nodes are deployed. These location-aware nodes assist in the localizing the rest of the sensor nodes through a localization algorithm. Range-free localization algorithms are simple and inexpensive, compared to the range-based localization algorithms. Yet, they suffer from low localization accuracy. Thus, this study aims to enhance the accuracy of range-free localization algorithms. The localization problem is mathematically formulated, and two approaches are adopted to provide efficient solutions with minimum average localization error. In the first approach, an improved variant of the Distance Vector-Hop localization algorithm is introduced. The position estimation phase is formulated as a minimization optimization problem. The optimization problem is solved by utilizing the Squirrel Search algorithm, which has not been adopted in solving any problem domain in WSNs. The performance of the proposed localization algorithm showed superiority in terms of accuracy, stability, and convergence rate when compared to the performance of existing localization algorithms that adopted the genetic algorithm, the particle swarm algorithm, and the differential evolution algorithm. In the second approach, the location of a sensor within the network is formulated as a sparse vector, hence reconstructed successfully with high probability by adopting the compressive sensing theory. The reconstruction problem is formulated as an optimization problem and solved using the ℓ1-norm minimization algorithm. The proposed algorithm outperforms several well-known algorithms, and shows competitive behavior compared to a range-based localization algorithm.