Extracting Time-Critical Situational Awareness From Resource Constrained Networks

Abstract: Overall Objective.The goal of this project is to facilitate the timely retrieval of dynamic situational awareness information from field deployed nodes by an operational center in disaster recovery or search and rescue missions, which are typically characterized by resource-constrained uncertain environments. Efficient situational awareness information retrieval under severe resource limitations is critical in applications like disaster response. Technology advances allow the deployment of field nodes capable of returning rich content (e.g., video/images) that can significantly aid rescue and recovery. However, development of techniques for acquisition, processing and extraction of the content that is relevant to the operation under resource constraints poses significant interdisciplinary challenges, which this project addresses.

Research Contributions. Towards realizing a networked system that facilitates the retrieval of time-critical, operation-relevant situational awareness, we have focused on the following research tasks.

Task A: Resource-Constrained Data Acquisition and Analysis. This task looks at how to reconfigure the network and adapt video analysis in real time to meet different (sometimes conflicting) application requirements, given resource constraints. We have specifically worked on adaptability of camera networks for efficient data acquisition, summarizing the data in a network of camera efficiently, and on adaptive algorithm selection based on the environmental conditions and available resources.

Task B: Information Fusion Under Resource Constraints. In this task, we seek to extract accurate situation awareness information (e.g., person detection) from the feeds of a set of wireless cameras, and deliver it in a timely way to an operations center, when presented with bandwidth constraints. We facilitate co-ordination among the cameras in such a network to realize significant energy savings compared to cases where there is no such co-ordination, and yet, achieve a high detection accuracy.

Task C: Cost-effective Query Formulation and Retrieval. This task addresses challenges in query formulation, refinement and retrieval, including prioritizing queries as per importance criteria, effective query dissemination in the field, and effective retrieval of the sensed information. In an experimental testbed at UCI that generates 1.2 million events per month, we are building a query computation framework which computes the query answer in a progressive manner while adhering to resource constraints.

For more information, please visit the UCR project webpage at http://www.ee.ucr.edu/~amitrc/resource-constraints.php.

Explanation of Demonstration: We design a novel framework that only transfers very limited data from distributed producers to a central summarizer supporting highly accurate detection and constructs concise visual summaries for key events. We demo it on Mininet (SDN emulator) where multiple producers are connected to a summarizer via a network of arbitrary topology. The summarizer then detects the top events and retrieves multimedia objects that represent each event.