Detecting unauthorized activities using a sensor network
Wireless sensor networks
| Hi-index | 0.00 |
Sensor networks are expected to bridge the gap between the physical world and the information world. The capabilities of sensor networks to monitor the physical environment can be used in almost every scientific discipline and also in defense, home, and biomedical applications. Some of the challenges in building these networks include the development of small stand-alone hardware, collaborative signal processing, and networking. The constraints specific to sensor networks are the distribution of information among nodes, the limitations of processing power, storage capabilities and power supply at each node. Also, the communication among nodes is constrained by the bandwidth available. Finally, node failures are expected due to the harsh environment in which sensor networks can be deployed. This thesis investigates the deployment of sensor networks in a geographic region for target detection. Its main focus is to consider the presence of faulty sensors in the network and to develop methods to build a system robust to these faults. Two problems are formulated, one concerning the collaboration of sensors to answer a detection query and another concerning the deployment of sensors to cover the region of interest. A model is developed as a framework to study the two problems and a survey of previous work relevant to this study is conducted. Two algorithms are developed to perform distributed target detection. These algorithms are analyzed and their performances in the absence and in the presence of faulty sensors are evaluated. To solve the sensor deployment problem, a metric is developed to measure the quality of a given deployment for detecting targets in the region of interest. That metric, namely minimum exposure, assesses the performance in detecting targets carrying out various unauthorized activities in the region. The minimum exposure can be computed for targets traveling at variable speed and in the presence of obstacles in the sensors field. Further, the metric can also be evaluated assuming that a subset of the sensors is faulty. This thesis uses minimum exposure to develop a deployment strategy in which sensors are deployed sequentially in steps until a performance level is reached.