Data fusion of decentralized local tracker outputs

Lobbia, R.; Kent, M.

doi:10.1109/7.303748

Cited by 30 publications

(14 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1 We must mention that the word "fusion," which appears often in the aforementioned references, usually refers to the "combination" of classifiers for improving classifier performance using a single training data set and not necessarily to "data fusion" (combining information coming from different data sources). Commonly used methods for "data fusion" are generally based on Bayesian theory [32], [33], state estimation with Kalman or particle filtering [34]- [39], evidence theory (DS) [40], [41] and its variations [10], [42]- [45], information theoretic framework [46], neural networks [47], and evolutionary algorithms [48], [49]. The traditional application area for data fusion has long been target detection and tracking [38], [39], [50]- [54].…”

Section: B Ensemble Approaches and Data Fusionmentioning

confidence: 99%

An Ensemble-Based Incremental Learning Approach to Data Fusion

Parikh

Polikar

2007

IEEE Trans. Syst., Man, Cybern. B

101

View full text Add to dashboard Cite

Abstract-This paper introduces Learn++, an ensemble of classifiers based algorithm originally developed for incremental learning, and now adapted for information/data fusion applications. Recognizing the conceptual similarity between incremental learning and data fusion, Learn++ follows an alternative approach to data fusion, i.e., sequentially generating an ensemble of classifiers that specifically seek the most discriminating information from each data set. It was observed that Learn++ based data fusion consistently outperforms a similarly configured ensemble classifier trained on any of the individual data sources across several applications. Furthermore, even if the classifiers trained on individual data sources are fine tuned for the given problem, Learn++ can still achieve a statistically significant improvement by combining them, if the additional data sets carry complementary information. The algorithm can also identify-albeit indirectly-those data sets that do not carry such additional information. Finally, it was shown that the algorithm can consecutively learn both the supplementary novel information coming from additional data of the same source, and the complementary information coming from new data sources without requiring access to any of the previously seen data.Index Terms-Data fusion, incremental learning, Learn++, multiple classifier/ensemble systems.

show abstract

Section: B Ensemble Approaches and Data Fusionmentioning

confidence: 99%

An Ensemble-Based Incremental Learning Approach to Data Fusion

Parikh

Polikar

2007

IEEE Trans. Syst., Man, Cybern. B

101

View full text Add to dashboard Cite

show abstract

“…Equations (21) and (13) indicate that the status equation of the system was linear, and the measurement equation was nonlinear. For this nonlinear system, the most commonly used filtering method is the Extended Kalman filter (EKF) [2] . Unscented Kalman filter (UKF) is an improved method upon the EKF, and it can overcome the difficulties of the EKF, such as the large linearization error of and the likelihood of ill-conditioned covariance 12] .…”

Section: Target Tracking Filtering Frame Based On the Sequential Unscmentioning

confidence: 99%

“…Target tracking is a process of estimating the status of the target by combing the measurements by radar and infrared imaging sensors. Literature [1] carried out the Kalman filtering for each sensor, and a further Kalman filtering of the filtered results in the fusion center (KF/ KF); literature [2] implemented Kalman filtering for each sensor, and the extended Kalman filtering for the results in the fusion center (KF / EKF); literature [3] used unscented Kalman filter (UKF) to solve the nonlinear filtering problem; all these methods are limited to the situation when the measured noise is subject to Gaussian distribution. Literature [4,5] solved the temporal change problem of the parameters in the nonlinear system by building a multi-mode interaction model (IMM).…”

mentioning

confidence: 99%

A target tracking method based on high-resolution radar photoelectronic image data fusion

Lin

Shi

2014

SPIE Proceedings

View full text Add to dashboard Cite

A target tracking model and a technique for target tracking filtering based on sequential unscented Kalman filter are presented to improve target tracking performance of high resolution radar/infrared imaging sensor composite guidance system. Firstly, a measurement model for imaging sensor based of the centroid of the target is derived from images. Secondly, a measurement model for radar based of the centroid of the target is derived from traits of high resolution radar. Finally, the data fusion filtering framework for target tracking based on sequential unscented Kalman filter is presented. From the results of simulated experiments, average rate and target tracking accuracy of convergence for the technique developed are superior to those of other techniques. In conclusion, the target tracking model and filtering algorithm developed are proper for high resolution radar/infrared imagery sensor composite guidance system. Key words: data fusion；target tracking；unscented conversion；Kalman filtering 1 IntroductionAt present, radar / IR imaging compound guidance has become one of the main directions for the terminal guidance technology development of precision-guided weapons. Having integrated the performance of the target tracking system, it has greatly improved upon target tracking performance of the single guided mode. Target tracking is a process of estimating the status of the target by combing the measurements by radar and infrared imaging sensors. Literature [1] carried out the Kalman filtering for each sensor, and a further Kalman filtering of the filtered results in the fusion center (KF/ KF); literature [2] implemented Kalman filtering for each sensor, and the extended Kalman filtering for the results in the fusion center (KF / EKF); literature [3] used unscented Kalman filter (UKF) to solve the nonlinear filtering problem; all these methods are limited to the situation when the measured noise is subject to Gaussian distribution. Literature [4,5] solved the temporal change problem of the parameters in the nonlinear system by building a multi-mode interaction model (IMM). Literature [6] adopted the multi-resolution filtering (MRF) method to deal with the image measurements, making them fit in the target tracking filtering framework. Literature [7,8] made use of the particle filtering (PF) to handle the filtering of the nonlinear system, and performed well for arbitrarily distributed noise. Literature [9] constructed a data integration framework using artificial intelligence (AI), and achieved satisfying results.In these tracking systems, measurement models were constructed with the target viewed as a target point. In essence, these measurement models were based on the position measurement by radar axis and that by infrared optical axis. Literature [10] considered the problems with infrared imaging measurement model; however, its radar measurement model was constructed for the ordinary radar, and did not apply to the high-resolution radar.This paper reported a study focusing on the high resolution radar / IR im...

show abstract

“…If the distributed data fusion is said to be optimal, compared to the centralized data fusion, it often has the approximate optimal performance under some assumptions or conditions. For early results about distributed data fusion, please refer to [74][75][76][77]. BarShalom [78] proposed that there exits the same process noise in the dynamic of different sensors.…”

Section: Distributed Data Fusionmentioning

confidence: 99%

Recent progress in networked control systems — A survey

Xia

Gao

Yan

et al. 2015

Int. J. Autom. Comput.

View full text Add to dashboard Cite

For the past decades, networked control systems (NCSs), as an interdisciplinary subject, have been one of the main research highlights and many fruitful results from different aspects have been achieved. With these growing research trends, it is significant to consolidate the latest knowledge and information to keep up with the research needs. In this paper, the results of different aspects of NCSs, such as quantization, estimation, fault detection and networked predictive control, are summarized. In addition, with the development of cloud technique, cloud control systems are proposed for the further development of NCSs.

show abstract

Data fusion of decentralized local tracker outputs

Cited by 30 publications

References 13 publications

An Ensemble-Based Incremental Learning Approach to Data Fusion

An Ensemble-Based Incremental Learning Approach to Data Fusion

A target tracking method based on high-resolution radar photoelectronic image data fusion

Recent progress in networked control systems — A survey

Contact Info

Product

Resources

About