Performance Evaluation Of Detection Research Articles

Detection analysis of CA family of adaptive radar schemes processing M-correlated sweeps in homogeneous and multiple-target environments

Most radar systems need some form of Doppler processing to filter out clutter and thereby reveal faster targets. The using of a moving target indicator (MTI) as a part of detection processing attenuates the input signal power at frequencies where the clutter return is dominant. However, the presence of MTI complicates the analysis of the detection system performance since its output sequence is correlated even though the input sequence may be uncorrelated. This paper is devoted to the detection performance evaluation of the mean-level CFAR detectors processing M-correlated sweeps in the presence of interfering targets. The consecutive pulses are assumed to be fluctuating according to the Swerling I model. Exact expressions are derived for the detection probability of the conventional mean-level detector (MLD), the maximum (MX)-MLD and the minimum (MN)-MLD under Rayleigh fluctuating target model. Performance for independent sweeps can be easily obtained by setting the sweep-to-sweep correlation coefficient equal to zero. Results are obtained for both homogeneous and nonhomogeneous background environments. It is shown that for fixed M, the relative improvement over the single sweep case increases as the correlation between sweeps decreases. When the correlation coefficient approaches unity. no gain is obtained for M>1 over the monopulse detection performance. For the same parameter values, the MN-MLD gives the best performance in the presence of interfering target returns among the reference noise samples.

Evaluation of human detection performance of targets embedded in natural and enhanced infrared images using image metrics

The evaluation of human detection performance of targets embedded in natural iR images is uniquely extended to deal with enhanced images. Various image metrics designed to describe global image clutter and local target-to-background contrast are presented. Agreement between experimental results, as obtained for both natural and enhanced iR images and image metrics, is evaluated by correlating the metric values with target detection results obtained from two psychophysical experiments, which were designed to record and measure human target detection performance and image quality judgments. Positive correlation values are generally obtained between the metric values and both the target detection (rate and probability) performance and the image quality scale values. More specifically, the local contrast metric is found to be the most suitable to the data recorded from the quality judgments experiment, while the global clutter metric is found to be the most suitable to the data recorded from the detection performance experiments. Further analysis of the relationship between the image metrics and the experimental results yield empirical classification thresholds that can be used to evaluate and predict human detection performance in similar cases. A method to evaluate the classification efficiency, is introduced and used to reconfirm the metrics appropriateness to describe detection performance and quality judgments of targets embedded in both natural and enhanced images. © 2000 Society ofPhoto-Optical Instrumentation Engineers. [S0091-3286(00)00202-6]

Detection of weak random signals in IID non-Gaussian noise

This paper considers the detection of weak random signals in circularly symmetric, independent, identically distributed noise. Locally optimum detectors and ad hoc nonlinearities are considered, with asymptotic expressions provided for evaluation of detection performance. The analytical expressions are used to evaluate the robustness of detectors to mismatch in the noise models. Finite-sample Monte Carlo simulation results indicate the reliability of these asymptotic measures in cases of practical interest. The results show that, as has been found for detection of weak known signals in non-Gaussian noise, reasonably configured ad hoc nonlinearities are nearly optimum and robust to modest errors in the noise statistics.

Performance Evaluation of Error Detection and Correcting Code for Optical Disk (Long Distance Code)

The interleaved Reed Solomon code called long distance code (LDC) is used for error correction and detection in the recording format of optical disks to maintain the integrity of recorded data. Performance evaluation of the code is necessary when the code is applied to another recording channel such as higher recording density one. Hence a performance evaluator for LDC was developed. The program evaluates various LDCs by changing the input parameters of the format. The performance evaluator consists of a burst error length evaluator and an EDAC performance evaluator. An ISO/IEC 15041 (540 MB 90mm MO disk) disk was assessed with this evaluator. The results show that the raw error rate of disks in the market is low enough for data integrity when defect management is applied.

Capacity of a multibeam, multisatellite CDMA mobile radio network with interference-mitigating receivers

More and more interference-mitigating algorithms are being investigated in an attempt to increase the capacity of code-division multiple-access radio networks. At the moment, the main question about interference-resilient receivers is: do they really bring forth the capacity increase they promise on theoretical grounds? The aim of this study is to give a preliminary answer to such a question, through the capacity analysis of the forward link of a multisatellite, multibeam, radio communication network. Specifically, starting from simplified yet realistic assumptions on the beam layout, the antenna footprint shape, and the coding/multiplexing/modulation and power control strategies, it is shown that the blind, adaptive, interference-mitigation receiver, selected as an appealing representative of the lot, can indeed boost the system quality of service in terms of outage probability. The results are derived after a mix of theoretical analysis (as far as the detector performance evaluation is concerned) and simulation (to examine a number of different random system configurations) to circumvent the inherent complexity of the issue.

Functional inputs to passive correlation detectors for oscillatory transient signals

This research investigates whether passive detection of transient signals can be improved by replacing received signals with functionals of the received signals in correlation detectors. Specifically, this paper assesses the impact of using energy spectrum and autocorrelation functional inputs to the cross-correlation (energy), bicorrelation, and tricorrelation detectors. Test signals with differing autocorrelation and energy spectrum properties are used in computer simulations with independent Gaussian noise to evaluate detector performance. Detection results are presented for the case when only two channels of input data are available to form the correlations, as well as the case when three and four channels of input data are available to form the higher order correlations. In the former case, it is shown that detection performance can be improved for the narrow-band signals by using the energy spectrum and autocorrelation functional inputs rather than the original received signal. In the latter case, it is shown that detection performance can be improved by using the autocorrelation input for the broadband signal and the energy spectrum input for the narrow-band signals, and that the tricorrelation detector performs best for the signals tested.

Design of a fast neutron coincidence counter

Monte Carlo simulations were carried out to simulate the response of a boron-loaded plastic scintillator array. These detectors offer potential advantages compared to moderated 3He systems because of their fast response and high efficiency. Tallies were made of neutron capture probability as a function of time and the results were coupled to assay variance estimates to evaluate detector performance. Orders-of-magnitude reductions in count time relative to a 50% efficient thermal counter are possible, in principle, for high (α,n) samples.

The SPECTRA program library: A PC based system for gamma-ray spectra analysis and INAA data reduction

A PC based system has been developed for the analysis of gamma-ray spectra and for the complete reduction of data from INAA experiments, including software to average the results from multiple lines and multiple countings and to produce a final report of analysis. Graphics algorithms may be called for the analysis of complex spectral features, to compare the data from alternate photopeaks and to evaluate detector performance during a given counting cycle. A database of results for control samples can be used to prepare quality control charts to evaluate long term precision and to search for systematic variations in data on reference samples as a function of time. The entire software library can be accessed through a user-friendly menu interface with internal help.

Performance evaluation of frequency diverse Bayesian ultrasonic flaw detection

The performance of an ultrasonic flaw detection system is valued by the success of differentiating the flaw echoes from those echoes scattered by microstructures (e.g., grains in polycrystalline metals). Microstructure noise (clutter) is stationary and will only vary when the frequency of the ultrasonic beam is changed. In this paper, frequency diverse microstructure information is obtained through the split‐spectrum processing of broadband ultrasonic backscattered echoes. The statistical difference between clutter and flaw echoes over different frequency bands has led to the development of a Bayes classifier that is quadratic and can incorporate the correlation properties of scattered echoes. The performance of the Bayes classifier has been examined for both experimental and computer simulated data, and compared to other commonly used techniques such as mean, minimum, median, and polarity detectors. The Bayes classifier shows superior performance, and the flaw‐to‐clutter visibility has been improved in the range of 5–15 dB when the measured flaw‐to‐clutter ratio is 0 dB or less. These results confirm the feasibility of grain noise suppression by frequency diversity with a subsequent statistical processor for flaw detection.

On the performance of signal-subspace processing--Part II: Coherent wide-band systems

This is the second part of a study which deals with the performance of multiple-source direction finding via signal-subspace processing [1]. An analytical evaluation of detection (determination of the number of sources) and direction estimation performance is presented for the coherent wide-band signal-subspace processing algorithm [2], which operates on the received array signals of large relative bandwidth. The probabilities of underestimating and overestimating the number of sources are derived, under asymptotic conditions and in the threshold regions, in terms of the choice of a penalty function and signal, noise, and array parameters for the cases of at most two closely spaced sources in spatially white noise. A scalar measure is used for the evaluation of the quality of the estimated signal subspaces. It is shown that the detection and estimation performances of the practical coherent algorithm are close to those of the perfectly coherent processing, and that they are relatively insensitive with respect to a preliminary estimate of the angles used to effect an approximate coherence.

Computer graphics for quality control in the INAA of geological samples

A data reduction system for the routine instrumental activation analysis of samples is described, with particular emphasis on interactive graphics capabilities for evaluating analytical quality. Graphics procedures have been developed to interactively control the analysis of selected photopeaks during spectral analysis, and to evaluate detector performance during a given counting cycle. Graphics algorithms are also used to compare the data on reference samples with accepted values, to prepare quality control charts to evaluate long term precision and to search for systematic variations in data on reference samples as a function of time.

A QRS detection algorithm.

In this article we present the choices that the designers of any QRS detector must make and explain the constraints we adopted. We outline the signal processing that precedes and the beat analysis that follows QRS detection in our single-channel, arrhythmia-monitoring algorithm and then expound the QRS detection algorithm in detail. Finally, we present the results of a QRS detector performance evaluation and comment on their importance. This article can be read to three depths: the text affords an overview of QRS detection for on-patient, ambulatory arrhythmia analysis; the commented pseudocode documents the logic of our QRS detector; and the pseudocode "footnotes" supply technical detail.

Performance evaluation of differential and discriminator detection of continuous phase modulation

Recently, bandwidth efficient constant-amplitude digital modulation schemes have also been shown to be power efficient when coherent detection is used. Partial-response continuous phase modulation (CPM) schemes are within this class. In some applications noncoherent detection is preferred. The performance of CPM systems is analyzed for differential and discriminator detection. An additive white Gaussian channel is assumed. The detectors make symbol-by-symbol decisions. The considered schemes are M-ary with an arbitrary modulation index and pulse shaping over several symbol intervals. The performance is analyzed by means of error probability expressions. The IF filter for the detectors is optimized within a special class of filters to give good performance. The differential detector is also analyzed on a Rayleigh fading channel. The fading is assumed to be slow. The IF filter is also optimized on this channel. Simulated error probabilities for discriminator detection with a Viterbi detector are also presented both for the Gaussian and the Rayleigh fading channel. The discriminator detector making symbol-by-symbol decisions is simulated on the Rayleigh fading channel. It is shown that partial-response CPM schemes with good performance can also be obtained with noncoherent detectors.

Automatic Detection of Mesocyclonic Shear with Doppler Radar

Abstract A region of high wind shear always accompanies large intense tornadoes. Such shear is readily measured by Doppler radar and thus can be used to identify regions in storms where tornadoes are likely to form. An algorithm developed at the National Severe Storms Laboratory uses decision thresholds to discriminate between mesocyclone shear and other shears not associated with organized circulatory flow. A detailed test of these and other thresholds was made by testing the algorithm on storms with and without mesocyclones. The algorithm requires storage and calculations on two consecutive radials of data, and thus is suitable for real time implementation. We have applied the algorithm to data from sixteen mesocyclones to evaluate detection performance. The probability of false alarm was 10%, and the probability of detection was 90%. The algorithm accounts for distortion of the shear pattern caused by the large antenna beam dimensions at long ranges.

Time characteristics of inspectors

Abstract There are large individual differences in the detection performance of inspectors, particularly in their time characteristics. This study established equations for measuring the effect of time allowed for detection, defined as a time characteristic parameter m. Experiments were conducted to evaluate the effectiveness of the m value, in which 15 inspectors counted the number of letters '£>' on each of 25 test sheets. Results showed that there are very large individual differences in the effect of time allowed to inspectors, and confirm the effectiveness of the time characteristic parameter m for evaluating detection performance.

Progress in evaluation of human observer visual detection performance using the ROC curve approach

This perspective is focused on amide groups of peptides interacting with water. 2D IR spectroscopy has already enabled structural aspects of the peptide backbone to be determined through its ability to measure the coupling between different amide-I modes. Here we describe why nonlinear IR is emerging as the method of choice to examine the fast components of the water dynamics near peptides and how isotopically edited peptide links can be used to probe the local water at a residue level in proteins. This type of research necessarily involves an intimate mix of theory and experiment. The description of the results is underpinned by relatively well established quantum-statistical theories that describe the important manifestations of peptide vibrational frequency fluctuations.

Utilization of a forced‐choice technique to verify statistical performance requirements for detection systems

An automated forced‐choice method of evaluating detection performance is employed instead of the familiar fixed‐threshold technique. The latter method utilizes a number of preselected, constant thresholds to estimate both detection and false‐alarm probabilities, whereas the former incorporates a variable threshold with a precise false‐alarm probability to estimate only the detection probability. The advantages of the forced‐choice technique are a predetermined false‐alarm rate, a single parameter test, less required test time, and real‐time monitoring of test results by minimizing the amount of data reduction. A comparison of both techniques is presented for two sets of statistics which are representative of those typically encountered in detection systems under test conditions. Because of the uncertainty associated with the variable threshold, the forced‐choice results are slightly pessimistic, but are within 0.3 dB of the fixed threshold results. However, the forced‐choice technique can easily be made equivalent to the fixed‐threshold procedure by incorporating a correction factor into the test procedure.

A Technique to Verify Statistical Performance Requirements for Sonar Detection Systems

An automated forced-choice method of evaluating detection performance is employed instead of the familiar fixed-threshold technique. The latter method utilizes a number of preselected, constant thresholds to estimate both detection and false alarm probabilities, whereas, the former incorporates a variable threshold with a precise false alarm probability to estimate only the detection probability. The estimate is obtained by selecting the “greatest of” a number of alternatives—one containing signal plus noise and the remainder containing noise only—to determine the number of correct detections (selection of signal plus noise alternative). The advantages of the forced-choice technique are predetermined false alarm rate, a single parameter test, less required test time, and real time monitoring of test results by minimizing the amount of data reductions. A comparison of both techniques is presented for two sets of typical statistics encountered under test conditions. Because of the uncertainty associated with the variable threshold, the forced-choice results are slightly pessimistic, but are within 0.3 dB of the fixed-threshold results. However, the forced-choice technique can easily be made equivalent to the fixed-threshold procedure by incorporating a correction factor into the test procedure. [Work supported by Naval Ship Systems Command.]

Prediction and Evaluation of Active Sonar-System Detection Preformance

The concept of signal excess—that is, the excess of signal received over that required for detection—as commonly used in both prediction and evaluation of active sonar-system detection performance is presented. Some of the results of independent measurements of the nominal value and the variability of the various parameters in the signal excess equation are discussed in terms of the deviation of a prediction from the actual evaluation of over-all system performance. Examples are used to demonstrate that the predicted performance will equal the over-all performance as measured for a total system only after recognition and adjustment for parameter interrelationships are made. Acoustic models that have been developed, validated in system tests, and used for active sonar-system performance prediction are shown. Criteria that have been imposed in the measurement of such parameters as source level, receiving sensitivity, propagation loss, interferring noise, and others, are presented together with illustrations of the effectiveness when used. Uncertainties in predicted system performance resulting from parameters and measurements that cannot be controlled or accurately estimated are cited and discussed.

Detection of Asymmetric Sideband Signals in the Presence of Noise

The first part of this paper analyzes three methods of detection of signals contaminated by fluctuation noise: (1) Linear Envelope Detectors (2) Product Detectors (3) Exalted Carrier Detectors. A new ratio, the ``video-to-noise-error ratio'' is proposed for a more adequate quantitative evaluation of detector performance. Its use is justified by a large number of theoretical curves and experimental results. The theoretical and practical results show that an improvement in the output video-to-noise-error ratio is obtained by substitution of a product detector for an envelope detector. Improvements of 11 decibels have been measured. The second part of this paper analyzes the effects of impulse noise on the detection of asymmetric sideband signals. It is shown that product detectors eliminate the rectification of the noise envelope found in envelope detectors. This is of particular interest in asymmetric sideband systems where the noise output will be at relatively high frequencies. These can often be separated from lower frequency signals, making possible efficient impulse noise suppression.