Ashraf M. Aziz

A new multiple decisions fusion rule for targets detection in multiple sensors distributed detection systems with data fusion

Currently, multiple sensors distributed detection systems with data fusion are used extensively in both civilian and military applications. The optimality of most detection fusion rules implemented in these systems relies on the knowledge of probability distributions for all distributed sensors. The overall detection performance of the central processor is often worse than expected due to instabilities of the sensors probability density functions. This paper proposes a new multiple decisions fusion rule for targets detection in distributed multiple sensor systems with data fusion. Unlike the published studies, in which the overall decision is based on single binary decision from each individual sensor and requires the knowledge of the sensors probability distributions, the proposed fusion method derives the overall decision based on multiple decisions from each individual sensor assuming that the probability distributions are not known. Therefore, the proposed fusion rule is insensitive to instabilities of the sensors probability distributions. The proposed multiple decisions fusion rule is derived and its overall performance is evaluated. Comparisons with the performance of single sensor, optimum hard detection, optimum centralized detection, and a multiple thresholds decision fusion, are also provided. The results show that the proposed multiple decisions fusion rule has higher performance than the optimum hard detection and the multiple thresholds detection systems. Thus it reduces the loss in performance between the optimum centralized detection and the optimum hard detection systems. Extension of the proposed method to the case of target detection when some probability density functions are known and applications to binary communication systems are also addressed.

A new adaptive decentralized soft decision combining rule for distributed sensor systems with data fusion

A new adaptive decentralized soft decision combining rule for multiple-sensor distributed detection systems with data fusion is proposed. Unlike previously published rules, the proposed combining rule fuses soft decisions of sensors rather than hard decisions of sensors and does not require the knowledge of the false alarm and detection probabilities of the distributed sensors. Such a fusion rule is adaptive, insensitive to the instabilities of the sensor thresholds, and has the advantage of soft decision fusion. The proposed combination rule is derived: (1) for the case where the fusion center estimates the error probabilities of the sensors and (2) for the case where the fusion center does not estimate the error probabilities of the sensors. The performance of the proposed approach is evaluated, and illustrative examples are presented in the cases of Gaussian and Rayleigh distributed observations. Comparisons with the optimum centralized fusion, the optimum soft decision fusion, a soft decision fusion approach based on fusing confidence levels, and the optimum decentralized hard decision fusion are also presented. The results indicate that the proposed approach significantly outperforms the optimum decentralized hard decision fusion, is better than the approach based on fusing confidence levels, and has a performance similar to that of the optimum soft decision fusion.

A joint possibilistic data association technique for tracking multiple targets in a cluttered environment

Abstract Multitarget tracking in a cluttered environment is a significant issue with a wide variety of applications. A typical approach to address this issue is the joint probabilistic data association (JPDA) technique. This technique determines joint probabilities over all targets and hits and updates the predicted target state estimate using a probability-weighted sum of innovations. This paper proposes a new joint possibilistic data association technique for tracking multiple targets. Unlike the JPDA technique, the proposed technique determines joint possibilities over all targets and hits and updates the predicted target state estimate using a possibility-weighted sum of innovations. The possibility weights are determined using the noise covariance matrices and the current received measurements such that the total sum of the distances between all measurements and targets is minimized. The proposed technique performs data association based on a possibility matrix of measurements to trajectories; thus, it highly reduces the computational complexity compared to conventional data association techniques. The proposed association technique is applied to examples of multitarget tracking in a cluttered environment, and the results demonstrate its efficiency.

A new multiple classifiers soft decisions fusion approach for exons prediction in DNA sequences

Prediction of exons locations in deoxyribonucleic acid (DNA) sequences is a significant issue for biologists. This paper proposes a new method to solve this problem. Unlike the published studies, in which the prediction of exons locations depends on hard decisions from a single classifier, the proposed prediction approach depends on fusion of soft decisions from two classifiers. In the proposed approach we utilize the sliding window discrete Fourier transform (DFT), which is normally used to detect exons 3-base periodicity feature, in a different manner. The novelty here depends on obtaining soft decisions, rather than hard decisions, from two classifiers using different numerical mapping schemes, and fuses them in a decision fusion center to obtain a final global decision about the prediction of exons locations. Simulation results based on real data performed on the HMR195 dataset showed that the proposed soft decisions fusion method achieves better prediction performance compared to the traditional hard decision single classifier method. Moreover the proposed method can easily be extended to more than two classifiers.

A new multitarget tracking approach based on a non-iterative fuzzy clustering means algorithm

In this paper, a new multitarget tracking approach is proposed. In the proposed approach, a non-iterative fuzzy clustering means algorithm is used to generate the association measures between the received measurements and the targets. Measurements-to-tracks associations are computed jointly across all targets and all validated measurements using the non- iterative fuzzy clustering means algorithm. For a given target, the validated measurement that has the maximum fuzzy association weight is used for updating the state of the target. The performance of the proposed approach is evaluated and compared to that of the standard nearest-neighbor association and conventional fuzzy logic data association approaches. The results show that the proposed tracking approach achieves better performance compared to the standard tracking approach and the conventional fuzzy tracking approaches.

Improved time-domain approaches for locating exons in DNA using zero-phase filtering

Accurate prediction of exons locations in deoxyri-bonucleic acid (DNA) sequences is an important issue for geneticists. Time-domain periodogram (TDP) and average magnitude difference function (AMDF) are two time-domain approaches previously proposed for this purpose. These two approaches employ a second-order infinite impulse response (IIR) resonant filter as a preprocessing stage so as to emphasize the period-3 behavior exhibited by the exonic segments of DNA strands. The major drawback of IIR filters is their non-linear phase response, which results in a delay distortion experienced by the spectral components of the genomic signal at the filter output. This type of distortion affects the exons prediction accuracy of the TDP/AMDF classifier. This paper proposes the use of zero-phase filtering technique in the preprocessing stage so as to eliminate the phase distortion introduced by the traditional filtering. MATLAB simulation conducted on the ASP67 genomic dataset shows that the proposed modified time-domain approaches using zero-phase filtering reveal better performance, compared with the traditional approaches, in terms of the receiver operating characteristic (ROC) curve, precision-recall curve and F-measure.

A new perspective on the choice of fuzzy membership functions in multitarget tracking systems

The functional paradigm for fuzzy multisenosr-multitarget tracking systems with data fusion consists of fuzzification, fuzzy knowledge-base, fuzzy inference mechanism, and defuzzification. In fuzzy system design, users start with some fuzzy rules, which are chosen heuristically based on their experience, and membership functions, which in many cases are chosen subjectively based on understanding the problem, and they use the developed system to tune these rules and membership functions. In most publications, in the area of track-to-track association in multitarget tracking systems, the fuzzy membership functions are chosen subjectively according to the underlying problem. The most commonly used membership functions are trapezoidal, triangular, piecewise linear, and Gaussian membership functions. They are chosen by the users based on their experiences. Therefore the problem of constructing optimal fuzzy membership functions is not considered in most publications. This paper addresses the critical issue of constructing optimal fuzzy membership functions for given input information in case of track-to-track association in multitarget tracking systems.

A novel and efficient approach for automatic classification of radar emitter signals

Radar emitter signal identification is a special issue of data clustering for classifying unknown radar emitters. In this paper, an efficient approach for automatic classification of radar emitter signals in multisensor systems is proposed. The proposed approach exploits measured features extracted from multiple sensors as well as the sensor accuracies for classification of unknown multiple radar targets. The proposed approach can easily be applied to any number of sensors with different accuracies, any number of emitters, and any number of measured features without exponential growing of the required computations. The performance of the proposed classification approach is evaluated in terms of percentage of correct classification and compared to other classification approaches. The results show the feasibility and the effectiveness of the proposed classification approach.

A least squares fusion rule in multiple sensors distributed detection systems

In this paper, a new least square data fusion rule in multiple sensor distributed detection system is proposed. In the proposed approach, the central processor combines the sensors hard decisions through least squares criterion to make the global hard decision of the central processor. In contrast to the optimum Neyman-Pearson fusion, where the distributed detection system is optimized at the fusion center level or at the sensors level, but not simultaneously, the proposed approach achieves global optimization at both the fusion center and at the distributed sensors levels. This is done without knowing the error probabilities of each individual distributed sensor. Thus the proposed least squares fusion rule does not rely on any stability of the noise environment and of the sensors false alarm and detection probabilities. Therefore, the proposed least squares fusion rule is robust and achieves better global performance. Furthermore, the proposed method can easily be applied to any number of sensors and any type of distributed observations. The performance of the proposed least squares fusion rule is evaluated and compared to the optimum Neyman-Pearson fusion rule. The results show that the performance of the proposed least squares fusion rule outperforms the performance of the Neyman-Pearson fusion rule.

On the use of Pseudo-EIIP mapping scheme for identifying exons locations in DNA sequences

Identifying exons locations in DNA sequences is one of the significant applications of signal processing in bioinformatics. Mapping a DNA character string into a numerical sequence is a prerequisite prior to its analysis using signal processing techniques. Recently, the Pseudo-EIIP DNA symbolic-to-numeric mapping scheme shows a promising performance with the filter-based exons prediction method, as compared to the traditional EIIP scheme. This paper investigates the performance of the Pseudo-EIIP mapping scheme with different period-3 exons prediction methods, as compared to a number of existing one-dimensional mapping schemes. We conduct MATLAB simulation on the BG570 genomic dataset for the purpose of evaluating exons prediction performance utilizing the ROC curve, precision-recall curve and F-measure. The results reveal the superiority of the Pseudo-EIIP numerical representation over other traditional one-dimensional representations, when employed with different period-3 exons prediction methods.