Intelligent Gateways Placements Techniques for Improved Dependability of Wireless Sensor Networks

Recently, wireless sensor networks (WSNs) have received enormous attentions due to their potential use in many applications. They can be employed to enrich our understanding of natural events, such as earthquakes and volcanoes, and to increase the efficiency of surveillance operations in secure installation, border control and military reconnaissance. In most setups, miniaturized sensor nodes are deployed to probe their surroundings and send the collected data to one or multiple gateways for processing. The gateway analyzes received data and decides on appropriate actions. Therefore, proper placement of sensors and gateways in the area of interest facilities communications and allows for better operation of the network. While sensors in many WSNs applications are randomly deployed, gateways often can be placed in a controlled manner. Gateways are usually richer than sensor nodes in terms of available resources due to their nature of operation, and thus they tend to cost more. In addition, many applications require the deployment of more than one gateway to better serve surrounding sensor nodes. There is a tradeoff between deploying multiple gateways and the cost incurred in doing so. Moreover, in many wireless sensor networks setups, sensors and gateways are placed in harsh environments, e.g. fires, or combat fields. The potential closeness of the gateway to dangerous events exposes it to damage and thus risks making the network dysfunctional. Therefore, protecting the gateway by repositioning it away from safety-hazardous spots is critical for the operation of the network. In this thesis, we study the asset management problem in wireless sensor networks. Asset planning in this context includes the determination of the number and the position of gateway nodes in order to achieve certain levels of performance. We focus our attention on two variants of the gateway positioning problem. The first deals with how to place multiple gateways in order to minimize data collection latency. We employ multiple heuristics for solving such optimization problem. The second variant addresses the problem of relocating the gateway when its current position exposes it to high risk and/or makes the network operation unacceptably inefficient. We present a novel approach that balances the network performance and gateway safety goals. An evolutionary neural network model is developed to assess the safety of the gateway at the various locations. The model is then used to direct the search for a safer position that would enhance or at least maintain an acceptable level of network performance. In addition, the approach guides the gateway during the move by finding safe paths leading to the new position. In addition, we study the problem of finding the least number of gateways to be deployed in order to achieve certain upper bound on the data latency within the network. We propose a solution that uses an innovative estimation function that balances between the number of gateways to be deployed or cost and the desired latency goal or performance based on many environments and surroundings factors. The validation results demonstrate the effectiveness of the proposed techniques