Simulated Evaluation of Ubiquitous Sensed Situational Awareness Systems

Developing technical capability that enables the utilization of ubiquitous sensing for situational awareness will likely catalyze a paradigm shift in data fusion, especially in light of recent advances in computing and data-driven techniques. Via the use of appropriate models, the well-entrenched data fusion methods can be utilized to mitigate some of the current challenges that are hindering the progress in this direction. However, due to the inherent challenges associated with ubiquitous sensing and randomness of data acquisition, adequate evaluation of advanced detection and sensor fusion systems remains a major challenge that hampers the progress in this direction. In this paper, a simulation approach for generating a synthetic dataset for comprehensive evaluation of Situational Awareness Systems (SASs) that utilize ubiquitous sensing for data acquisition is presented. With a particular emphasis on recording and organizing spatio-temporal observations, the steps towards construction of a comprehensive dataset that takes into account various spatiotemporal parameters corresponding to location, hazards and detectors is presented. The presented approach is then utilized to illustrate a simplified evaluation of an efficient data storage and inference system designed for asynchronous detections that are characteristic of ubiquitous sensed SASs.