Showing papers on "Sequential probability ratio test published in 1995"
TL;DR: It can be shown that the likelihood ratio is monotonic; consequently, the tests can be easily implemented in realtime and suggest that the TSPRT is more desirable.
Abstract: Studies sequential acquisition schemes of m-sequences, based on the sequential probability ratio test (SPRT) and a truncated SPRT (TSPRT), with noncoherent demodulation. Most reported results on sequential acquisition schemes assume that independent samples are available for the decision process. The assumption of independent and identically distributed (i.i.d.) samples requires the integrator in the receiver to be reset periodically. This introduces loss in the effective signal-to-noise ratio, degrading the performance. In the present paper, on the contrary, the authors' two sequential schemes use continuous integration. It can be shown that the likelihood ratio is monotonic; consequently, the tests can be easily implemented in realtime. Methods are proposed for designing the decision thresholds to achieve the desired false-alarm and miss probabilities. Performances of the proposed schemes are obtained and they suggest that the TSPRT is more desirable. The effect of slowly-varying channel fading is also investigated. Results show that fading does not affect the false alarm probabilities, but it can drastically reduce the probability of detecting the alignment of the two PN sequences, especially when the fading is severe. >
46 citations
TL;DR: Bounds on error probabilities and asymptotic expressions for the stopping time and error probabilities are given and a key result of this correspondence is a proof that the generalized MSPRT is asymptonically efficient.
Abstract: A sequential multihypothesis test known as the M-ary sequential probability ratio test (MSPRT) is generalized to account for nonuniform decision costs. Bounds on error probabilities and asymptotic expressions for the stopping time and error probabilities are given. A key result of this correspondence is a proof that the generalized MSPRT is asymptotically efficient.
40 citations
09 May 1995
TL;DR: A modified SPRT algorithm is described, using a new feature dubbed blanking variability (BV) as the basis for discrimination, and preliminary results showed that the proposed method decreases the overall error rate to 5%.
Abstract: We describe a modified sequential probability ratio test (SPRT) for the discrimination of ventricular fibrillation (VF) from ventricular tachycardia (VT) in measured surface electrocardiograms. The algorithm uses a novel regularity measure dubbed blanking variability (BV) applied to threshold crossings from the measured ECG. Blanking variability corresponds to the normalized rate of change of cardiac rate as the blanking interval is varied. The algorithm has been trained and tested using separate subsets drawn from the MIT-BIH malignant arrhythmia database. BV values are modeled using a truncated Gaussian distribution, and parameter values are derived by averaging over the training component of the database. In testing, the algorithm achieved an overall classification accuracy of 95%.
37 citations
TL;DR: This paper considers sequential tests with varying stopping bounds (NSPRT) for testing simple hypotheses on independent but nonstationary observations and states a considerable generalization of the famous Wald-Wolfowitz theorem on the (simultaneous) minimizations of the expected sample sizes for given error probabilities.
Abstract: For original paper see IEEE Trans. Inform. Theory, vol.38, p.177-82 (1992). Liu and Blostein (1992) consider sequential tests with varying stopping bounds (NSPRT) for testing simple hypotheses on independent but nonstationary observations. They state a considerable generalization of the famous Wald-Wolfowitz theorem [1948] on the (simultaneous) minimization of the expected sample sizes for given error probabilities (Liu and Blostein, Theorem 27). Unfortunately, this statement fails to be true: 1) a simple counterexample may be constructed by choosing densities f/sub 01//spl ne/f/sub 1,1/ such that 0 >
8 citations
TL;DR: The sequential probability ratio test developed by Wald is used to detect a change of the standard deviation of a Gaussian distributed white noise signal and results show that the method is very well suited for anomaly detection.
8 citations
01 Jan 1995
TL;DR: Results show that fading does not affect the false alarm probabilities, but it can drastically reduce the probability of detecting the alignment of the two PN sequences, especially when the fading is severe.
Abstract: AbstTact- In this paper, we study sequential acquisition schemes of m-sequences, based on the sequential probability ratio test (SPRT) and a truncated SPRT (TSPRT), with noncoherent demodulation. Most reported results on sequential acquisition schemes assume that independent samples are available for the decision process. The assumption of independent and identically distributed (i.i.d.) samples requires the integrator in the receiver to be reset periodically. This introduces loss in the effective signal-to-noise ratio, degrading the performance. In this paper, on the contrary, our two sequential schemes use continuous integration. It can be shown that the likelihood ratio is monotonic; consequently, the tests can be easily implemented in real time. Methods are proposed for designing the decision thresholds to achieve the desired false-alarm and miss probabilities. Performances of the proposed schemes are obtained and they suggest that the TSPRT is more desirable. The effect of slowly-varying channel fading is also investigated. Results show that fading does not affect the false alarm probabilities, but it can drastically reduce the probability of detecting the alignment of the two PN sequences, especially when the fading is severe.
2 citations
TL;DR: A modification of the single sensor parity algorithm is proposed in order to make the statistical properties of the generated residual determinable in advance.
1 citations
07 Jun 1995
TL;DR: The design of the modulation classifiers, which identify the modulation type of received noisy signals in this research, are examined, and a new approach called the sequential probability ratio test has important advantages, including low error rate, low computational complexity and flexible tradeoff between performance and speed.
Abstract: We examine the design of the modulation classifiers, which identify the modulation type of receivednoisy signals in this research. The family of modulation schemes of our concern is the quadratureamplitude modulation (QAM), which has been widely used in modern digital communication. Mod-ulation recognition has been traditionally viewed as a hypothesis test problem with a fixed samplesize. Because the amount of received data increases with time, we formulate the problem as a van-able sample size test, and propose an iterative recognition procedure. The new approach called thesequential probability ratio test (SPRT) has important advantages, including low error nate, lowcomputational complexity and flexible tradeoff between performance and speed.Keywords: modulation recognition, sequential probability ratio test, multihypothesis test. 1. Introduction For traditional fixed station communication links, the modulation scheme is usually fixed becausethe channel condition is quite stable and the transmitted data type is unchanged. However, for someapplications such as mobile communication or transmission of multimedia data, it is desirable to usemodems with variable constellation schemes to adapt to the fast changing channel and accommodatea wide range of data transmission rates, bit error rates and data types[19]. In this research we areinterested in designing receivers which can identify the QAM (Quadrature Amplitude Modulation)constellation schemes from the received signal with unknown transmitted symbol sequence.Two different approaches have been reported in the previous works. The first approach formulatesthe problem with hypothesis tests by given exact statistical descriptions of received signals [1] ,
1 citations
TL;DR: In this paper, an asymptotic optimality property of the Wald sequential test for general, possibly dependent, random variables with respect to the rth moment of observation time, under some r-quick convergence condition is derived.
Abstract: We give a generalization to the case of m hypotheses of a theorem of Lai and derive an asymptotic optimality property of the Wald sequential test for general, possibly dependent, random variables with respect to the rth moment of observation time, under some r-quick convergence condition. We also extend the definition of this convergence and give an application to sequential analysis.
1 citations
08 Nov 1995
TL;DR: Results indicate that the test statistic based on the random sequence model offers considerable savings in E[T/sub acq/] for moderate to large values of chip SNR.
Abstract: We study the mean acquisition time performance of a novel serial search acquisition scheme for chip-asynchronous direct-sequence spread-spectrum systems. The sequences arising under the out-of-phase hypothesis are modeled as random sequences, and an appropriate likelihood ratio is used in a sequential probability ratio test (SPRT) to test for the true phase of the received sequence. A different (and commonly used) likelihood ratio is obtained for the SPRT if the out-of-phase sequences are modeled as zero sequences, and we use the performance under this model as a benchmark for comparison. An appropriate verification stage for the serial search scheme is designed based on fixed sample size (FSS) tests which employ the same likelihood ratio as used in the testing stage. A general expression for the mean acquisition time Er[T/sub acq/] is derived, and the results indicate that the test statistic based on the random sequence model offers considerable savings in E[T/sub acq/] for moderate to large values of chip SNR.
02 Mar 1995
TL;DR: In this article, the authors assume that n players are represented by n reals, uniformly distributed over the unit interval, and assume that each player has a fixed number of reals.
Abstract: Assume that n players are represented by n reals, uniformly distributed over the unit interval.
TL;DR: In this article, supermartinagles are used to approximate the optimal stopping sets associated with the Bayesian sequential test for normal population means, and several conclusions are given in the paper.
Abstract: In this paper, we present a simple and direct approach in which supermartinagles are used to approximate the optimal stopping sets associated with the Bayesian sequential test for normal population means. Several conclusions are given.
TL;DR: The sequential probability ratio test or the sequential likelihood ratio test can be used to test one-sided hypotheses regarding the ratio λ = μ/σ as mentioned in this paper, which is a function of the usual t-statistic.
Abstract: Either the sequential probability ratio test or the sequential likelihood ratio test can be used to test one-sided hypotheses regarding the ratio λ = μ/σ. Both tests are functions of the usual t-statistic. I discuss and compare empirical and theoretical properties of these two tests, and provide tables of critical and truncation values for a modified version of the sequential likelihood ratio test.
TL;DR: In this article, the Bayesian sequential estimation problem for an exponential family of processes is considered and the Bayes sequential procedures are derived using a weighted square error loss and observing cost involving a polynomial function of the process.
Abstract: The Bayesian sequential estimation problem for an exponential family of processes is considered. Using a weighted square error loss and observing cost involving a polynomial function of the process, the Bayes sequential procedures are derived.