During the past five years, extensive research has been conducted to develop methods to correlate and fuse data from physical sensors (“hard” data) and from human observers (“soft” data). Hard sensor data generally involves signals, images or scalar information related to the location, identification and characterization of entities (e.g., humans, vehicles, etc.), while soft data typically involves textual information (observations, inferences, and comments) from human observers. Modern Information Fusion systems are also exploiting “Contextual” information (e.g., socio-cultural data), much of which is also of a soft type. With the rapid proliferation of mobile communications devices and increased global connectivity, the need to fuse hard and soft data becomes an increasingly ubiquitous problem. Applications involve areas such as environmental monitoring, citizen science, military situation awareness/assessment, emergency response and other areas. There are numerous challenges involved in hard and soft data fusion based on issues such as; the inherent differences in level of abstraction of hard versus soft data (viz., hard data about observed entities represented by signals, images, vectors and scalars versus semantic meta-data based on human observations and inferences), challenges in characterizing the performance of physical sensors versus human observers, differences in data rates, issues in correlation and association, etc.
Topics of interest
This session is focused on research and advances in hard and soft data fusion. The session is anchored by several papers based on a five year, multi-university research initiative (MURI) funded by the U. S. Army Research Office (ARO) on the topic of Unified Research on Network-based Hard/Soft Information Fusion, led by Dr. John Lavery, ARO Program Manager). This research was conducted in collaboration with the University at Buffalo (State University of New York), The Pennsylvania State University, the Tennessee State University, IONA College, and the University of Illinois. The anchor papers submitted below provide both an overview of the five year research as well as perspectives on topics such as hard sensor fusion, soft sensor fusion, development of a collaborative cyber infrastructure for research, data discovery and visualization, and creation of a special hard/soft data set for experimentation and evaluation.
hard and soft data fusion, entity tracking and identification, situation awareness, semantic data processing, test and evaluation of fusion systems
Special Session Organizers
- David Hall: Pennsylvania State University (U.S.A.)
- James Llinas: University at Buffalo (U.S.A.)
- Rakesh Nagi: University of Illinois (U.S.A.)
Special Sessión Contact
- David Hall (firstname.lastname@example.org)