Multi-Target Localization and Tracking with Wireless Sensor Networks for Digital Storytelling

Wireless Sensor Networks (WSNs) have been considered for a wide variety of applications in areas such as medicine, industry, environmental issues or defense, due to their attractive characteristics which have motivated their rapid diffusion. Navigation, environmental monitoring, disaster prevention, health monitoring or area monitoring are some examples of their extended use. Multiple target localization and tracking using WSNs has become an interesting and challenging topic of research in recent years. A wide variety of approaches have been proposed in the literature but further work has to be done in order to provide more accurate localization and tracking algorithms. There is no unique solution for finding a particular methodology that performs optimal object detection, localization and tracking due to the excessive amount of configurable parameters in a Wireless Sensor Network (i.e. topology, sensors features, communication protocols, network architecture, etc.) or the characteristics of the scenario under study (i.e. cooperative or non-cooperative targets, indoor or outdoor areas, environmental conditions, etc.). Radio Tomographic Imaging (RTI) has been considered an interesting technique for tracking purposes. It allows an image reconstruction of the region of interest using the received signal strength (RSS) in every mote of the wireless network. Localization and target tracking is accomplished by analyzing the changes in the RSS in the set of estimated images due to the presence of targets. This technique has shown favorable results for area monitoring even in cluttered environments. In the present work, a multi-target localization and tracking algorithm has been developed for digital storytelling applications in public spaces. Radio Tomographic Imaging has been the selected methodology for target localization and tracking due to the characteristics of the area under study. Due to the lack of RF sensors in the area, environmental data (i.e. RSS changes in the area) has been generated for latter tests of the developed algorithm. From the generated data, image estimation and further processing has been performed, achieving successfully the localization and tracking tasks. Afterwards, Kalman Filter and Particle Filter are applied to the tracking results for better position estimation. A comparative study of different set-ups is also performed in order to understand the system behavior under different scenarios. Additionally, a user interface is created to make the algorithm configuration more user friendly for possible future users that may hardly understand the developed MATLAB program