Publications

Refine Results

(Filters Applied) Clear All

Blind clustering of speech utterances based on speaker and language characteristics

Published in:
5th Int. Conf. Spoken Language Processing (ICSLP), 30 November - 4 December 1998.

Summary

Classical speaker and language recognition techniques can be applied to the classification of unknown utterances by computing the likelihoods of the utterances given a set of well trained target models. This paper addresses the problem of grouping unknown utterances when no information is available regarding the speaker or language classes or even the total number of classes. Approaches to blind message clustering are presented based on conventional hierarchical clustering techniques and an integrated cluster generation and selection method called the d* algorithm. Results are presented using message sets derived from the Switchboard and Callfriend corpora. Potential applications include automatic indexing of recorded speech corpora by speaker/language tags and automatic or semiautomatic selection of speaker specific speech utterances for speaker recognition adaptation.
READ LESS

Summary

Classical speaker and language recognition techniques can be applied to the classification of unknown utterances by computing the likelihoods of the utterances given a set of well trained target models. This paper addresses the problem of grouping unknown utterances when no information is available regarding the speaker or language classes...

READ MORE

Improving accent identification through knowledge of English syllable structure

Published in:
5th Int. Conf. on Spoken Language Processing, ICSLP, 30 November - 4 December 1998.

Summary

This paper studies the structure of foreign-accented read English speech. A system for accent identification is constructed by combining linguistic theory with statistical analysis. Results demonstrate that the linguistic theory is reflected in real speech data and its application improves accent identification. The work discussed here combines and applies previous research in language identification based on phonemic features [1] with the analysis of the structure and function of the English language [2]. Working with phonemically hand-labelled data in three accented speaker groups of Australian English (Vietnamese, Lebanese, and native speakers), we show that accents of foreign speakers can be predicted and manifest themselves differently as a function of their position within the syllable. When applying this knowledge, English vs. Vietnamese accent identification improves from 86% to 93% (English vs. Lebanese improves from 78% to 84%). The described algorithm is also applied to automatically aligned phonemes.
READ LESS

Summary

This paper studies the structure of foreign-accented read English speech. A system for accent identification is constructed by combining linguistic theory with statistical analysis. Results demonstrate that the linguistic theory is reflected in real speech data and its application improves accent identification. The work discussed here combines and applies previous...

READ MORE

Airbus 320 performance during ATC-directed breakouts on final approach

Published in:
MIT Lincoln Laboratory Report ATC-265

Summary

An evaluation of Airbus 320 (A320) performance during ATC-directed breakouts was conducted in a two-part study during 1995. Phase 1 tested the combined effect of proposed ATC phraseology, pilot situational awareness training, and an A320-specific breakout procedure on performance. Pilot training included a briefing and viewing a videotape, but no simulator practice. Turn performance statistics from the Precision Runway Monitor Demonstration Program were used as the test criteria. Pilot preferences regarding procedures and the training material were also elicited. Three conclusions were: (1) breakout performance given the tested combination of pilot training and proposed ATC phraseology did meet the test criteria; (2) breakout performance given existing procedures did not meet the test criteria; and (3) the tested breakout procedure should be refined because it conflicted with other cockpit procedures and increased the transition time to a positive climb rate. Based on the results of this study, it is recommended that a combination of pilot situational awareness training, A320 breakout procedure, and modified ATC breakout phraseology equivalent to that tested in Phase 2 be employed for simultaneous parallel approach operations in instrument meteorological conditions.
READ LESS

Summary

An evaluation of Airbus 320 (A320) performance during ATC-directed breakouts was conducted in a two-part study during 1995. Phase 1 tested the combined effect of proposed ATC phraseology, pilot situational awareness training, and an A320-specific breakout procedure on performance. Pilot training included a briefing and viewing a videotape, but no...

READ MORE

Sheep, goats, lambs and wolves: a statistical analysis of speaker performance in the NIST 1998 speaker recognition evaluation

Summary

Performance variability in speech and speaker recognition systems can be attributed to many factors. One major factor, which is often acknowledged but seldom analyzed, is inherent differences in the recognizability of different speakers. In speaker recognition systems such differences are characterized by the use of animal names for different types of speakers, including sheep, goats, lambs and wolves, depending on their behavior with respect to automatic recognition systems. In this paper we propose statistical tests for the existence of these animals and apply these tests to hunt for such animals using results from the 1998 NIST speaker recognition evaluation.
READ LESS

Summary

Performance variability in speech and speaker recognition systems can be attributed to many factors. One major factor, which is often acknowledged but seldom analyzed, is inherent differences in the recognizability of different speakers. In speaker recognition systems such differences are characterized by the use of animal names for different types...

READ MORE

A dual-band circularly polarized aperture-coupled stacked microstrip antenna for global positioning satellite

Author:
Published in:
IEEE Trans. Antennas Propag., Vol. 45, No. 11, November 1997, pp. 1618-25.

Summary

This paper describes the design and testing of an aperture-coupled circularly polarized antenna for global positioning satellite (GPS) applications. The antenna operates at both the L1 and L2 frequencies of 1575 and 1227 MHz, which is required for differential GPS systems in order to provide maximum positioning accuracy. Electrical performance, low-profile, and cost were equally important requirements for this antenna. The design procedure is discussed, and measured results are presented. Results from a manufacturing sensitivity analysis are also included.
READ LESS

Summary

This paper describes the design and testing of an aperture-coupled circularly polarized antenna for global positioning satellite (GPS) applications. The antenna operates at both the L1 and L2 frequencies of 1575 and 1227 MHz, which is required for differential GPS systems in order to provide maximum positioning accuracy. Electrical performance...

READ MORE

Techniques for improved reception of 1090 MHz ADS-B signals

Published in:
17th DASC: Proc. of the 17th. Digital Avionics Systems Conf., 31 October - 7 November 1998, Vol. 2, pp. G25-1 - G25-9.

Summary

The recent development of ADS-B (Automatic Dependent Surveillance-Broadcast) is based on the use of the Mode S transponders now carried by all air carrier and commuter aircraft. ADS-B aircraft broadcast aircraft positions, identity, and other information via semi-random Mode S transponder squitters. Other aircraft or ground facilities receive the squitters and the associated position and status. Squitter reception includes the detection of the Mode S 1090 MHz waveform preamble, declaration of the bit and confidence values, error detection, and (if necessary) error correction. The current techniques for squitter reception are based upon methods developed for use in Mode S narrow-beam interrogators and for ACAS. In both of these applications, the rate of Mode NC fruit that is stronger than the Mode S waveform is relatively low, nominally less than 4,000 fruit per second. Extended squitter applications now include long range (up to 100 nmi) air-air surveillance in support of free flight. This type of surveillance is sometimes referred to as Cockpit Display of Traffic Information (CDTI). In high density environments, it is possible to operate with fruit rates of 40,000 fruit per second and higher. Operation of extended squitter in very high ModeNC fruit environments has led to the need to re-evaluate squitter reception techniques to determine if improved performance is achievable. The purpose of this paper is to provide a summary of work in progress to investigate improved squitter reception techniques. Elements of improved squitter reception being investigated include (1) the use of amplitude to improve bit and confidence declaration accuracy, (2) more capable error correction algorithms, and (3) more selective preamble detection approaches.
READ LESS

Summary

The recent development of ADS-B (Automatic Dependent Surveillance-Broadcast) is based on the use of the Mode S transponders now carried by all air carrier and commuter aircraft. ADS-B aircraft broadcast aircraft positions, identity, and other information via semi-random Mode S transponder squitters. Other aircraft or ground facilities receive the squitters...

READ MORE

Vulnerabilities of reliable multicast protocols

Published in:
IEEE MILCOM '98, Vol. 3, 21 October 1998, pp. 934-938.

Summary

We examine vulnerabilities of several reliable multicast protocols. The various mechanisms employed by these protocols to provide reliability can present vulnerabilities. We show how some of these vulnerabilities can be exploited in denial-of-service attacks, and discuss potential mechanisms for withstanding such attacks.
READ LESS

Summary

We examine vulnerabilities of several reliable multicast protocols. The various mechanisms employed by these protocols to provide reliability can present vulnerabilities. We show how some of these vulnerabilities can be exploited in denial-of-service attacks, and discuss potential mechanisms for withstanding such attacks.

READ MORE

AM-FM separation using shunting neural networks

Published in:
Proc. of the IEEE-SP Int. Symp. on Time-Frequency and Time-Scale Analysis, 6-9 October 1998, pp. 553-556.

Summary

We describe an approach to estimating the amplitude-modulated (AM) and frequency-modulated (FM) components of a signal. Any signal can be written as the product of an AM component and an FM component. There have been several approaches to solving the AM-FM estimation problem described in the literature. Popular methods include the use of time-frequency analysis, the Hilbert transform, and the Teager energy operator. We focus on an approach based on FM-to-AM transduction that is motivated by auditory physiology. We show that the transduction approach can be realized as a bank of bandpass filters followed by envelope detectors and shunting neural networks, and the resulting dynamical system is capable of robust AM-FM estimation in noisy environments and over a broad range of filter bandwidths and locations. Our model is consistent with recent psychophysical experiments that indicate AM and FM components of acoustic signals may be transformed into a common neural code in the brain stem via FM-to-AM transduction. Applications of our model include signal recognition and multi-component decomposition.
READ LESS

Summary

We describe an approach to estimating the amplitude-modulated (AM) and frequency-modulated (FM) components of a signal. Any signal can be written as the product of an AM component and an FM component. There have been several approaches to solving the AM-FM estimation problem described in the literature. Popular methods include...

READ MORE

1.5-um Tapered-Gain-Region Lasers with High-CW Output Powers

Published in:
IEEE Photonics Technol. Lett., Vol. 10, No. 10, October 1998, pp. 1377-1379.

Summary

High-power diode lasers consisting of a ridge-waveguide section coupled to a tapered region have been fabricated in 1.5um InGaAsP-InP multiple-quantum-well material. Self-focusing at high current densities and high-intensity input into the taper section has been identified as a fundamental problem in these devices that has to be dealt with. To date, continuous-wave output powers>1 W with=80% of the power in the near-diffraction-limited central lobe of the far field have been obtained through a judicious choice of device parameters.
READ LESS

Summary

High-power diode lasers consisting of a ridge-waveguide section coupled to a tapered region have been fabricated in 1.5um InGaAsP-InP multiple-quantum-well material. Self-focusing at high current densities and high-intensity input into the taper section has been identified as a fundamental problem in these devices that has to be dealt with. To...

READ MORE

Total lightning as a severe weather diagnostic in strongly baroclinic systems in Central Florida

Published in:
19th Conf. on Severe Local Storms, 14-18 September 1998, pp. 643-647.

Summary

Severe weather is defined by specific thresholds in wind. hail size and vorticity. All of these phenomena have close physical connections with vertical drafts in deep convection, which are themselves not directly measured with scanning Doppler radars of the NEXRAD type. Cloud electrification and lightning are particularly sensitive to these drafts because they modulate the supply of supercooled water which is the growth agent for the ice particles (ice crystals, graupel and hail) believed essential for electrical charge separation. For these reasons, one can expect correlations at the outset between total lightning activity and the development of severe weather which may aid in the understanding and prediction of these extreme weather conditions. The exploration of these ideas has historically been impeded by lack of good quantitative observations. A recent review of results on severe storm electrification (Williams, 1998) indicates a general absence of cases for which total lightning activity is documented over the lifetime of a severe storm. The recent development of LISDAD (Lightning Imaging Sensor Data Application Display) (Boldi, et aI., 1998) has largely remedied this problem. This paper is concerned with the use of LISDAD to quantify the behavior of total lightning in all types of severe weather, with a focus on a pair of extraordinarily electrified supercells in the Florida dry season.
READ LESS

Summary

Severe weather is defined by specific thresholds in wind. hail size and vorticity. All of these phenomena have close physical connections with vertical drafts in deep convection, which are themselves not directly measured with scanning Doppler radars of the NEXRAD type. Cloud electrification and lightning are particularly sensitive to these...

READ MORE