High-level Information Fusion Research Articles

An integrated approach to high-level information fusion

In today’s fast paced military operational environment, vast amounts of information must be sorted out and fused not only to allow commanders to make situation assessments, but also to support the generation of hypotheses about enemy force disposition and enemy intent. Current information fusion technology has the following two limitations. First, current approaches do not consider the battlefield context as a first class entity. In contrast, we consider situational context in terms of terrain analysis and inference. Second, there are no integrated and implemented models of the high-level fusion process. This paper describes the HiLIFE (High-Level Information Fusion Environment) computational framework for seamless integration of high levels of fusion (levels 2, 3 and 4). The crucial components of HiLIFE that we present in this paper are: (1) multi-sensor fusion algorithms and their performance results that operate in heterogeneous sensor networks to determine not only single targets but also force aggregates, (2) computational approaches for terrain-based analysis and inference that automatically combine low-level terrain features (such as forested areas, rivers, etc.) and additional information, such as weather, and transforms them into high-level militarily relevant abstractions, such as NO-GO, SLOW-GO areas, avenues of approach, and engagement areas, (3) a model for inferring adversary intent by mapping sensor readings of opponent forces to possible opponent goals and actions, and (4) sensor management for positioning intelligence collection assets for further data acquisition. The HiLIFE framework closes the loop on information fusion by specifying how the different components can computationally work together in a coherent system. Furthermore, the framework is inspired by a military process, the Intelligence Preparation of the Battlefield, that grounds the framework in practice. HiLIFE is integrated with a distributed military simulation system, OTBSAF, and the RETSINA multi-agent infrastructure to provide agile and sophisticated reasoning. In addition, the paper presents validation results of the automated terrain analysis that were obtained through experiments using military intelligence Subject Matter Experts (SMEs).

A new approach to higher-level information fusion using associative learning in semantic networks of spiking neurons

This paper presents a new approach to higher-level information fusion in which knowledge and data are represented using semantic networks composed of coupled spiking neuron nodes. Networks of simulated spiking neurons have been shown to exhibit synchronization, in which sub-assemblies of nodes become phase locked to one another. This phase locking reflects the tendency of biological neural systems to produce synchronized neural assemblies, which have been hypothesized to be involved in binding of low-level features in the perception of objects. The approach presented in this paper embeds spiking neurons in a semantic network, in which a synchronized sub-assembly of nodes represents a hypothesis about a situation. Likewise, multiple synchronized assemblies that are out-of-phase with one another represent multiple hypotheses. The initial network is hand-coded, but additional semantic relationships can be established by associative learning mechanisms. This approach is demonstrated by simulation of proof-of-concept scenarios involving the tracking of suspected criminal vehicles between meeting places in an urban environment. Our results indicate that synchronized sub-assemblies of spiking nodes can be used to represent multiple simultaneous events occurring in the environment and to effectively learn new relationships between semantic items in response to these events. In contrast to models of synchronized spiking networks that use physiologically realistic parameters in order to explain limits in human short-term memory (STM) capacity, our networks are not subject to the same limitations in representational capacity for multiple simultaneous events. Simulations demonstrate that the representational capacity of our networks can be very large, but as more simultaneous events are represented by synchronized sub-assemblies, the effective learning rate for establishing new relationships decreases. We propose that this effect could be countered by speeding up the spiking dynamics of the networks (a tactic of limited availability to biological systems). Such a speedup would allow the number of simultaneous events to increase without compromising the learning rate.

Real-time use of Kohonen's self-organizing maps for threat stabilization

Threat evaluation is a high-level information fusion function of great importance in the command and control decision-making process. Due to highly dynamic changes in the environment, it becomes more and more difficult to assign a stable threat value to an entity. The present work aims at stabilizing the threat value assigned to a moving entity having a manoeuvring behaviour. To this end, a special type of neural networks, the self-organizing maps, is used in order to extract some important features from the kinematics of the moving entity. Its direction of advance is computed based on fusing the extracted features. The stabilization the direction of advance allows the stabilization of the threat value assigned to such an entity, especially when this value is based on the closest point of approach concept. The proposed approach is tested over a set of simulated data.

Hypothesis management for information fusion

The efficient management of large collections of fusion hypotheses presents a critical challenge for scaling high-level information fusion systems to solve large problems. We motivate this challenge in the context of two Alphatech research projects, and discuss several partial solutions. A recurring theme is the exploitation of space-efficient, factored representations of multiple hypotheses to enable an efficient search for good hypotheses.

A Data Fusion Algorithm for Estimating Link Travel Time

The growing demand for real-time traffic information brought about various types of traffic collection mechanisms in the area of Intelligent Transport Systems (ITS). There are, however, two procedures in making various traffic data into information. First, a robust information-making process of utilizing data into the representative information for each traffic collection mechanism is required. Second, the integration process of fusing the “estimated” information into the “representative information” for each link out of each source is also required. That is, both data reduction and/or data-to-information process and a higher-level information fusion are required. This article focuses on the development of an information fusion algorithm based on a voting technique, fuzzy regression, and Bayesian pooling technique for estimating dynamic link travel time in congested urban road networks. The algorithm has been proposed and validated using field experimental data—GPS probes and detector data collected over various roadway segments. It has been found that the estimated link travel time from the proposed algorithm is more accurate than the mere arithmetic mean counterpart from each traffic source. The limitations of the algorithm and future research agenda have also been discussed.