Galina Rogova

Designing ontologies for higher level fusion

The purpose of higher level fusion is to produce contextual understanding of the states of the environment and prediction of their impact in relation to specific goals of decision makers. One of the main challenges of designing higher level fusion processes is to provide a formal structure of domain-specific types of entities, attributes, situations, and the relations between them for reasoning about situations and threats. This paper presents an attempt at confronting this challenge by describing a process for building formal ontologies that combines a top-down philosophical perspective (from the most abstract levels to domain-specific levels) with a bottom-up application-based perspective (from domain-specific levels to the most abstract levels). The main focus of this paper is to provide a conceptual framework for formally capturing various sorts of complex relation-types, which can serve as a means for a more thorough decomposition of objects, attributes/properties, events, processes, and relations, necessary for higher level fusion processing.

Ontology meta-model for building a situational picture of catastrophic events

The overall goal of the research described in this paper is to design a general methodology for situation assessment to support crisis management. The purpose of situation assessment is to produce contextual understanding and interpretation of the relationships between various entities, events and behaviors of interest. One of the main challenges of designing a situation assessment process is to provide a formal structure for ontological analyses of domain-specific types of entities, attributes, situations, and the relationships between them. This paper presents an attempt to confront this challenge by utilizing formal philosophical categories and theories to design a formal ontology of catastrophic events that describe the most basic and relevant structures of objective reality. The ontology is designed from both a top-down philosophical perspective (from abstract level to domain-specific level) and a bottom-up application-based perspective (from domain-specific level to abstract level). Situations are characterized by spatial items of interest (SNAP) at different levels of granularity (objects, aggregates, combination of aggregates), temporal items of interest (SPAN) that characterize the behaviors of SNAP items, and the relations between them.

Information quality in information fusion

Designing fusion systems for decision support in complex dynamic situations calls for an integrated human-machine information environment, in which some processes are best executed automatically while for others the judgment and guidance of human experts and end-users are critical. Thus decision making in such environment requires constant information exchange between human and automated agents that utilize operational data, data obtained from sensors, intelligence reports, and open source information. The quality of decision making strongly depends on the success of being aware of, and compensating for, insufficient information quality at each step of information exchange. Designing the methods of representing and incorporating information quality into this environment is a relatively new and a rather difficult problem. The paper discusses major challenges and suggests some approaches to address this problem.

An Ontological Analysis of Threat and Vulnerability

The overall goal of this paper is to provide a formal ontological analysis of threat. In particular, this paper discusses the formal ontological structure of threats as integrated wholes possessing three interrelated parts: intentions, capabilities and opportunities, and shows how these elements stand to one another, as well as to states of vulnerability. This discussion offers a means for understanding variations of threat conditions such as potential vs. viable threats and dispersed threats. A general, metaphysical, upper-level framework for the development of a formal threat ontology (ThrO) offers a necessary foundation for designing consistent and comprehensive models for threat prediction and mitigation

Context-based multi-level information fusion for harbor surveillance

Harbor surveillance is a critical and challenging part of maritime security procedures. Building a surveillance picture to support decision makers in detection of potential threats requires the integration of data and information coming from heterogeneous sources. Context plays a key role in achieving this task by providing expectations, constraints and additional information for inference about the items of interest. This paper proposes a fusion system for context-based situation and threat assessment with application to harbor surveillance. The architecture of the system is organized in two levels. The lowest level uses an ontological model to formally represent input data and to classify harbor objects and basic situations by deductive reasoning according to the harbor regulations. The higher level applies Belief-based Argumentation to evaluate the threat posed by suspicious vessels. The functioning of the system is illustrated with several examples that reproduce common harbor scenarios.

On-line data validation in distributed data fusion

Data acquisition and data fusion systems are becoming increasingly complex, being in fact systems of systems, where every component may be a system with varying levels of autonomy by themselves. Possible changes in system configuration by entities joining or being removed from the system make the system complex. As synchronous operation cannot be expected in such a system configuration, the temporal and spatial correctness of data must be achieved via other means. This paper presents the concept of mediated interactions as a method for ensuring correctness of computation in a distributed system. The mediator associated with each computing entity is responsible for online checking of the data both before it is sent out at the sender side and before it is received at the receiver side, ensuring that only data satisfying the validity constraints of the receiver-side data processing algorithm is used in computation. This assumes that each data item is augmented with metadata, which enables online data validation. The validity and quality dimensions in use depend on the system requirements defined by a specific problem and situational context; they may be temporal, spatial and involve various data quality dimensions, such as accuracy, confidence, relevance, credibility, and reliability. Among other capabilities, the mediator is able to cope with the unknowns in the temporal dimension that occur at runtime and are not predictable, such as channel delay, jitter of clocks and processing delays. This capability becomes an especially relevant factor in multi-tasking systems and in configurations in which a computing entity may have to process a variable number of parallel streams of data. Both the architecture and a simulation case study of a distributed data fusion scenario are presented in the paper.

Information fusion approach to microcalcification characterization

Abstract The paper presents an information fusion-based approach to one of the most challenging problems in mammogram interpretation: the problem of characterizing mammographic microcalcifications as benign or malignant. There are two categories of methods typically used for designing decision aids for diagnosis of microcalcifications: computer vision methods employing intensity-based features automatically extracted from images and methods using mammogram characteristics considered by human experts. The achieved recognition accuracy of both types of methods is not yet sufficient for them to be utilized in clinical practice. The paper introduces a hybrid system combining decisions of classifiers utilizing both domain knowledge-based and intensity-based features within the framework of the Evidence theory. The system comprises a hierarchical evidential classifier employing a combination of texture features of individual microcalcifications and a neural network employing cluster features observed and described by a radiologist. The results of a pilot study have shown that a false alarm rate of the hybrid system is lower than the false alarm rate of each single classifier used in the combination as well as that of the radiologists participated in the study.

On the Integration of Cognitive Work Analysis within a Multisource Information Fusion Development Methodology

This research explored the means by which methods in cognitive engineering, namely, work domain analysis, could be used to provide input to the development of advanced information processing, or multisensor information fusion, algorithms. Specifically, a work domain analysis of an emergency management environment (in a post-earthquake context) was performed, and linked abstraction hierarchy models representing the emergency management and response system, the physical environment (e.g., buildings, transportation systems, civilians), and other goal directed agents (e.g., civilian responders and volunteers) were created. Outputs from that analysis (information requirements) were input to the design of the information processing algorithms, providing guidance as to the nature of information required by decision makers, which could be computed through fusion capabilities. This ongoing work thus presents an example of an integrated cognitive engineering/multisensor fusion methodology.

Findings of the NATO workshop on data fusion technologies for harbour protection

The NATO Security Through Science Program and the Defence Investment Division requested and sponsored the organization of a NATO Advanced Research Workshop (ARW) on the topic of Data Fusion Technologies for Harbour Protection, which was held June 27-July 1, 2005 in Tallinn, Estonia. The goal of the workshop was to help knowledge exchange between the technology experts and the security policy makers for a better understanding of goals, functions and information requirements of the decision makers as well as the way the data fusion technology can help enhancing security of harbours. In addition to presentations by experts from the research community on detection and fusion technologies as well as in practice and policy the workshop program included daily breakout sessions, in which the participants were given an opportunity to brainstorm on the topics of the workshop in interdisciplinary smaller teams. The working groups: (i) chose a scenario, including threat stages, threat types, threat methods and ranges, and response constraints due to the particular harbour environment; then (ii) identified: (a) requirements (objectives, functions and essential elements of information); (b) technologies (available and future); (c) information available and necessary through sensors and other sources, as agencies and jurisdiction; (d) methods: detection, identification, situation assessment, prediction. This paper describes the main issues and proposed approaches that were identified by the working groups.

Use of neural networks and decision fusion for lithostratigraphic correlation with sparse data, Mono-Inyo Craters, California

We explore the use of multiple artificial neural networks combined within the framework of the Dempster-Shafer Theory of Evidence to construct a hybrid information processing system for the correlation of tephra layers. The working hypothesis is that the system can correctly correlate tephra layers from one site to another even when data are sparse. The collection and analysis of data appropriate for utilization in standard statistical techniques aiding correlation is costly and time consuming. Given this state of affairs, here we employ a hybrid pattern recognition approach, which allows us to produce a recognition result with a relatively small amount of data. We used the major tephra-fall layers within one eruption sequence, the North Mono eruption, Mono Craters, CA, USA, to determine whether the system can be trained to distinguish layers on a bed-by-bed basis. The beds are distinguished in the field by the fraction of pumice, grading, zoning, thickness, and size of large pumice or lithic fragments. These same features were used to train the hybrid system. In the best case, the hybrid system was able to categorize observations correctly 93% of the time, which was markedly better than using neural networks alone. The average result for all pairwise comparisons of beds by the hybrid system is 76%, with the results for two beds that were not distinct removed. We conclude that it is possible to train the system to discriminate reasonably successfully among tephra layers with limited data. Because the system was not designed with any reliance on features specific to tephra layers, it may be possible to apply the system to the categorization of any stratigraphic units.