Moses W. Chan

Human-aided multi-sensor fusion

This paper discusses some fundamental requirements for a human-aided multi-sensor fusion system, and proposes an approach that implements such a system. This approach involves integration of the probabilistic argumentation system and the structural evidential argumentation system, which both are variants of the Dempster-Shafer belief function theory. An example is shown that illustrates how this integrated approach can be applied to missile defense applications.

MODELLING DEPENDENCE IN DEMPSTER-SHAFER THEORY

Belief functions can only be combined by Dempsters rule when they are based on independent items of evidence. This paper proposes a method for handling the case where there is some probabilistic dependence among the items of evidence. The method relies on compact representations of joint probability distributions on the assumption variables associated with the belief functions. These distributions are then used to compute degrees of support of hypotheses of interest. It is shown that the theory of hints is the appropriate general framework for this method.

A belief function classifier based on information provided by noisy and dependent features

A model and method are proposed for dealing with noisy and dependent features in classification problems. The knowledge base consists of uncertain logical rules forming a probabilistic argumentation system. Assumption-based reasoning is the inference mechanism that is used to derive information about the correct class of the object. Given a hypothesis regarding the correct class, the system provides a symbolic expression of the arguments for that hypothesis as a logical disjunctive normal form. These arguments turn into degrees of support for the hypothesis when numerical weights are assigned to them, thereby creating a support function on the set of possible classes. Since a support function is a belief function, the pignistic transformation is then applied to the support function and the object is placed into the class with maximal pignistic probability.

Boosting target tracking using particle filter with flow control

Target detection and tracking with passive infrared (IR) sensors can be challenging due to significant degradation and corruption of target signature by atmospheric transmission and clutter effects. This paper summarizes our efforts in phenomenology modeling of boosting targets with IR sensors, and developing algorithms for tracking targets in the presence of background clutter. On the phenomenology modeling side, the clutter images are generated using a high fidelity end-to-end simulation testbed. It models atmospheric transmission, structured clutter and solar reflections to create realistic background images. The dynamics and intensity of a boosting target are modeled and injected onto the background scene. Pixel level images are then generated with respect to the sensor characteristics. On the tracking analysis side, a particle filter for tracking targets in a sequence of clutter images is developed. The particle filter is augmented with a mechanism to control particle flow. Specifically, velocity feedback is used to constrain and control the particles. The performance of the developed “adaptive” particle filter is verified with tracking of a boosting target in the presence of clutter and occlusion.

A Java implementation of the probabilistic argumentation system for data fusion in missile defense applications

In missile defense target recognition applications, knowledge about the problem may be imperfect, imprecise, and incomplete. Consequently, complete probabilistic models are not available. In order to obtain robust inference results and avoid making inaccurate assumptions, the probabilistic argumentation system (PAS) is employed. In PAS, knowledge is encoded as logical rules with probabilistically weighted assumptions. These rules map directly to Dempster-Shafer belief functions, which allow for uncertainty reasoning in the absence of complete probabilistic models. The PAS can be used to compute arguments for and against hypotheses of interest, and numerical answers that quantify these arguments. These arguments can be used as explanations that describe how inference results are computed. This explanation facility can also be used to validate intelligent information, which can in turn improve inference results. This paper presents a Java implementation of the probabilistic argumentation system as well as a number of new features. A rule-based syntax is defined as a problem encoding mechanism and for Monte Carlo simulation purposes. In addition, a graphical user interface (GUI) is implemented so that users can encode the knowledge database, and visualize relationships among rules and probabilistically weighted assumptions. Furthermore, a graphical model is used to represent these rules, which in turn provides graphical explanations of the inference results. We provide examples that illustrate how classical pattern recognition problems can be solved using canonical rule sets, as well as examples that demonstrate how this new software can be used as an explanation facility that describes how the inference results are determined.

Multisensor fusion for 3D target tracking using track-before-detect particle filter

This work presents a novel fusion mechanism for estimating the three-dimensional trajectory of a moving target using images collected by multiple imaging sensors. The proposed projective particle filter avoids the explicit target detection prior to fusion. In projective particle filter, particles that represent the posterior density (of target state in a high-dimensional space) are projected onto the lower-dimensional observation space. Measurements are generated directly in the observation space (image plane) and a marginal (sensor) likelihood is computed. The particles states and their weights are updated using the joint likelihood computed from all the sensors. The 3D state estimate of target (system track) is then generated from the states of the particles. This approach is similar to track-before-detect particle filters that are known to perform well in tracking dim and stealthy targets in image collections. Our approach extends the track-before-detect approach to 3D tracking using the projective particle filter. The performance of this measurement-level fusion method is compared with that of a track-level fusion algorithm using the projective particle filter. In the track-level fusion algorithm, the 2D sensor tracks are generated separately and transmitted to a fusion center, where they are treated as measurements to the state estimator. The 2D sensor tracks are then fused to reconstruct the system track. A realistic synthetic scenario with a boosting target was generated, and used to study the performance of the fusion mechanisms.

Tracking low SNR targets using particle filter with flow control

In this work we study the problem of detecting and tracking challenging targets that exhibit low signal-to-noise ratios (SNR). We have developed a particle filter-based track-before-detect (TBD) algorithm for tracking such dim targets. The approach incorporates the most recent state estimates to control the particle flow accounting for target dynamics. The flow control enables accumulation of signal information over time to compensate for target motion. The performance of this approach is evaluated using a sensitivity analysis based on varying target speed and SNR values. This analysis was conducted using high-fidelity sensor and target modeling in realistic scenarios. Our results show that the proposed TBD algorithm is capable of tracking targets in cluttered images with SNR values much less than one.

Editorial: Special section on dependence issues in knowledge-based systems

Approximate-reasoning methods rely on a host of evidence and knowledge manipulation functions to produce estimates of the truth of various hypotheses of interest. Among these functions two are particular important: information propagation and information combination [1]. The special section on Dependence Issues in Knowledge-Based Systems of this issue of the International Journal of Approximate Reasoning focuses on a most important aspect of information combination, namely the role of dependence between informational bodies in their aggregation into a fused result. Intuitively, the combination of information derived from identical sources should not result in a evidential body that is more informative than either of them while, on the other hand, information derived from independent sources should reinforce areas of common agreement while eliminating conclusions that are inconsistent with any of the sources being merged. The modeling of informational dependence and the selection of appropriate algorithms on the basis of knowledge about such dependence are obviously of paramount importance in the derivation of credible fused results. Depending on the approximate-reasoning methodology being employed it is sometimes possible to explicitly represent, through various modeling structures, the nature of the relations between multiple information sources. More often than not, however, such knowledge is only available in a partial, imprecise, and uncertain manner. The objective of this special section is to present novel results relevant to the fusion of dependent information in such situations. Although the notions of dependence and independence are very important, not only in approximate-reasoning approaches—including all forms of probabilistic and possibilistic methods—but also throughout system modeling and analysis, the papers published in this special section address the problem of combining partially dependent evidential bodies in the context of the theory of belief functions [2,3] and its generalizations [4,5]. The theory of belief functions is specially suited to the representation of imprecise knowledge both for its ability to model ignorance and for its formal relations to other approximate-reasoning formalisms. This methodology, however, still lacks general mechanisms to represent dependencies between sets of descriptive variables and to merge dependent evidence on the basis of such relations. Furthermore, there are still a number of open questions about the important notion of evidential independence. The works included in this issue are representative of various approaches to the formal characterization of the notion of independence, the representation and combination of dependent evidence, and the learning of combination formulas applicable to specific situations. Cattaneo derives two combination rules for the fusion of belief functions derived from not necessarily independent sources. These formulas approximate plausibility and commonality functions of the fused evidence so as to meet a number of requirements that should be reasonably demanded from any such formulation. In the context of such derivations the author proposes a new measure of evidential conflict that has advantages over that derived from the normalization factor associated with Dempster’s combination formula. In addition, these rules are related to the minimum rule of possibility theory, Dempster’s combination formula, and Denœux’s cautious rule [6]. As Cattaneo points out, his work should be regarded as a step towards the development of better approximations for the combined evidence. Jirous̆ek and Vejnarová approach the evidence-combination problem from the perspective of the representation of multidimensional belief distributions as a composition of marginal or conditional distributions of lower dimensionality. These compositional models, introduced earlier as an alternative to probabilistic graphical Markov models, rely on new notion of conditional independence to iteratively apply a composition operator to factorize complex distributions into smaller components. Originally developed in the context of classical probability theory, these models were later extended to represent distributions in possibility theory. The present work presents a new composition operator that extends the approach into the realm of the theory of belief functions permitting the generation of compositional representations that reveal relations of independence between variables. Monney, Chan, and Romberg present a reasoning model, based on a sound combination of classical logic and probability theory, to treat automatic classification problems where the discriminating features are dependent and not fully reliable.