Feature object extraction - a fuzzy logic approach for evidence accrual in the Level 1 Fusion classification problem

Stubberud, Stephen C.; Pudwill, R.

doi:10.1109/cimsa.2003.1227224

Cited by 10 publications

(9 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The FKF of the FOX system is based upon a development of Watkins [7] and modified in [2]. This approach was selected to allow a wide variety of measurement types and uncertainty models to be used with ease.…”

Section: Evidence Accrual Using Feature Object Extractionmentioning

confidence: 99%

See 1 more Smart Citation

Estimation of Target Maneuvers from Tracked Behavior Using Fuzzy Evidence Accrual

Stubberud¹,

Kramer²,

Stubberud³

2019

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

While the Kalman filter, including its many variants, has been the staple of the tracking community, it also has been shown to have drawbacks, particularly when tracking through a maneuver. The most common issue is a lag in the position of the target track compared to the true target position as the target performs its maneuver. Another more problematic issue can occur where the filter covariance collapses, requiring the filter to be reinitialized. Techniques exist to compensate for maneuvers, but generating their response relies on detection of error between the estimated trajectory and the measured target position. In this effort, a maneuver detection routine is developed that can be used in conjunction with more standard maneuver compensation approaches. This routine is able to validate the existence of a maneuver more quickly than use of the inherent detection relied upon in the other methods. Maneuver detection is performed by an evidence accrual system that uses a fuzzy Kalman filter to incorporate new information and provide a level of evidence that maneuver is occurring. The input data uses behavior characteristics of the Kalman gain vector from the tracking algorithm.

show abstract

Section: Evidence Accrual Using Feature Object Extractionmentioning

confidence: 99%

“…Feature objection extraction [1,2] is an evidence accrual technique that can be applied to classification problems. An undirected tree [3] with various levels of information is used to connect evidence nodes.…”

Section: Introductionmentioning

confidence: 99%

Estimation of Target Maneuvers from Tracked Behavior Using Fuzzy Evidence Accrual

Stubberud¹,

Kramer²,

Stubberud³

2019

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

show abstract

“…To fuse this information together, a fuzzy-logic based evidence accrual system is proposed to fuse the reports over multiple events. This technique was first applied to Level 1 classification problem [2,3]. It was then advanced to the applications of Level 2 fusion or situational assessment [4,5].…”

Section: Introductionmentioning

confidence: 99%

Impact assessment fusion using fuzzy evidence accrual with a novel pattern recognition technique

Stubberud¹,

Kramer

2010

2010 5th IEEE International Conference Intelligent Systems

View full text Add to dashboard Cite

The problem of impact assessment, also referred to as Level 3 fusion, is addressed using a fuzzy evidence accrual system. This technique injects data using a fuzzy Kalman filter, allowing for numeric and linguistic information from various sensors and reporting systems to be fused together. The solution provides not only an estimate of any threats or changes in mission parameters but also a measure of the uncertainty in the computed estimate. This technique works well with data over multiple events that accrue into evidence over time. One of the reporting mechanisms to be included in the fusion system is pattern recognition technique based on image correlation. This technique provides the ability to fuse individual scans of data from multiple sensors into a correlation score to known patterns that then can be incorporated into the evidence that drives the evidence accrual system. As with the evidence accrual system, the image correlation technique can work with multiple data sources and formats simultaneously.

show abstract

“…A significant difference from previous work in [9] is that information cannot traverse through the root node. This allows the values for each class to be updated independently of those for the other classes.…”

Section: Implementation Of Evidence Accrualmentioning

confidence: 75%

“…The root node indicates the target number and the leaf nodes are for four potential classes: tank, APC, truck, TEL. Unlike in [9], each state/node has evidence injected independently, without effect on the others.…”

Section: Implementation Of Evidence Accrualmentioning

confidence: 94%

Evidence accrual for the level 1 fusion classification problem

Stubberud¹,

Kramer

Geremia³

CIMSA. 2005 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications, 2005.

View full text Add to dashboard Cite

Target classification is an important part of the Level 1 Fusion problem. This is particularly true for the military application of Level 1 Fusion. With precision warfare, localization of a target is often not enough and the use of sensor information to provide an estimate to what a potential target is has grown increasingly in value. A good estimate of the classification of a potential target can determine whether it will be prosecuted. To provide this estimate, a fusion system must combine pieces of information provided by various sensors and other sources over a period of time. The most common automated techniques for the classification problem provide a probability measure of the possible classes. Another concept in classification is the use of evidence accrual. As opposed to the creation of scoring techniques that represent a random variable representation of the classification, the evidence accrual technique builds scores based on the information that can be compared to other scores or thresholds of the decision process. Since evidence affects the various potential classes differently, the technique developed is based on decoupled fuzzy-logic-based Kalman filters, similar to the concept of first-order observers. The proposed technique addresses three key problems in the classification problem. It is designed to incorporate both numeric and nonnumeric sensor reports. It incorporates measurement uncertainty and provides a level of uncertainty for each class. Finally, the technique allows for evidence to be applied independently to each potential class.

show abstract

Feature object extraction - a fuzzy logic approach for evidence accrual in the Level 1 Fusion classification problem

Cited by 10 publications

References 6 publications

Estimation of Target Maneuvers from Tracked Behavior Using Fuzzy Evidence Accrual

Estimation of Target Maneuvers from Tracked Behavior Using Fuzzy Evidence Accrual

Impact assessment fusion using fuzzy evidence accrual with a novel pattern recognition technique

Evidence accrual for the level 1 fusion classification problem

Contact Info

Product

Resources

About