Publications
Operational evaluation of runway status lights
June 1, 2006
Journal Article
Published in:
Lincoln Laboratory Journal, Vol. 16, No. 1, June 2006, pp. 123-146.
Topic:
R&D area:
R&D group:
Summary
To maintain safe separation of aircraft on the airport surface, air traffic controllers issue verbal clearances to pilots to sequence aircraft arrivals, departures, and runway crossings. Although controllers and pilots work together successfully most of the time, mistakes do occasionally happen, causing several hundred runway incursions a year and, less frequently, near misses and collisions in the United States. With this rate of incursions, it is imperative to have an independent warning system as a backup to the current system. Runway status lights, a system of automated, surveillance-driven stoplights, have been designed to provide this backup function. The lights are installed at runway-taxiway intersections and at departure points along the runways. They provide a clear signal to pilots crossing or departing from a runway, warning them of potential conflicts with traffic already on the runway. Existing FAA-installed radar surveillance is coupled with Lincoln Laboratory-developed algorithms to generate the light commands. To be compatible with operations at the busiest airports, the algorithms must drive the lights such that during normal operations pilots will almost never encounter a red light when it is safe to cross or depart from a runway. A minimal error rate must be maintained even in the face of inevitable imperfections in the surveillance system used to drive the safety logic. A prototype runway status light system has been designed at Lincoln Laboratory and installed at the Dallas/Fort Worth International Airport, where Laboratory personnel have worked with the FAA to complete an operational evaluation of the system, demonstrating the feasibility of runway status lights in the challenging, complex environment of one of the world's busiest airports.
Summary
To maintain safe separation of aircraft on the airport surface, air traffic controllers issue verbal clearances to pilots to sequence aircraft arrivals, departures, and runway crossings. Although controllers and pilots work together successfully most of the time, mistakes do occasionally happen, causing several hundred runway incursions a year and, less...
READ MORE
Surveillance accuracy requirements in support of separation services
June 1, 2006
Journal Article
Published in:
Lincoln Laboratory Journal, Vol. 16, No. 1, June 2006, pp. 97-122.
Topic:
R&D area:
R&D group:
Summary
The Federal Aviation Administration is modernizing the Air Traffic Control system to improve flight efficiency, to increase airspace capacity, to reduce flight delays, and to control operating costs as the demand for air travel continues to grow. Promising new surveillance technologies such as Automatic Dependent Surveillance Broadcast and multisensor track fusion offer the potential to augment the ground-based surveillance and controller-display systems by providing more timely and complete information about aircraft. The resulting improvement in surveillance accuracy may potentially allow the expanded use of the minimum safe-separation distance between aircraft. However, these new technologies cannot be introduced with today's radar-separation standards, because they assume surveillance will be provided only through radar technology. In this article, we review the background of aircraft surveillance and the establishment of radar separation standards. The required surveillance accuracy to safely support aircraft separation with National Airspace System technologies is then derived from currently widely used surveillance systems. We end with flight test validation of the derived results, which can be used to evaluate new technologies.
Summary
The Federal Aviation Administration is modernizing the Air Traffic Control system to improve flight efficiency, to increase airspace capacity, to reduce flight delays, and to control operating costs as the demand for air travel continues to grow. Promising new surveillance technologies such as Automatic Dependent Surveillance Broadcast and multisensor track...
READ MORE
The mixer and transcript reading corpora: resources for multilingual, crosschannel speaker recognition research
May 24, 2006
Conference Paper
Published in:
Proc. Int. Conf. on Language Resources and Evaluation, LREC, 24-26 May 2006, pp. 117-20.
Topic:
R&D area:
Summary
This paper describes the planning and creation of the Mixer and Transcript Reading corpora, their properties and yields, and reports on the lessons learned during their development.
Summary
This paper describes the planning and creation of the Mixer and Transcript Reading corpora, their properties and yields, and reports on the lessons learned during their development.
READ MORE
A scalable phonetic vocoder framework using joint predictive vector quantization of MELP parameters
May 14, 2006
Conference Paper
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, Speech and Language Processing, ICASSP, 14-19 May 2006, pp. 705-708.
Topic:
R&D area:
Summary
We present the framework for a Scalable Phonetic Vocoder (SPV) capable of operating at bit rates from 300 - 1100 bps. The underlying system uses an HMM-based phonetic speech recognizer to estimate the parameters for MELP speech synthesis. We extend this baseline technique in three ways. First, we introduce the concept of predictive time evolution to generate a smoother path for the synthesizer parameters, and show that it improves speech quality. Then, since the output speech from the phonetic vocoder is still limited by such low bit rates, we propose a scalable system where the accuracy of the MELP parameters is increased by vector quantizing the error signal between the true and phonetic-estimated MELP parameters. Finally, we apply an extremely flexible technique for exploiting correlations in these parameters over time, which we call Joint Predictive Vector Quantization (JPVQ).We show that significant quality improvement can be attained by adding as few as 400 bps to the baseline phonetic vocoder using JPVQ. The resulting SPV system provides a flexible platform for adjusting the phonetic vocoder bit rate and speech quality.
Summary
We present the framework for a Scalable Phonetic Vocoder (SPV) capable of operating at bit rates from 300 - 1100 bps. The underlying system uses an HMM-based phonetic speech recognizer to estimate the parameters for MELP speech synthesis. We extend this baseline technique in three ways. First, we introduce the...
READ MORE
SVM based speaker verification using a GMM supervector kernel and NAP variability compensation
May 14, 2006
Conference Paper
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, Speech and Language Processing, ICASSP, Vol. 1, 14-19 May 2006, pp. 97-100.
Topic:
R&D area:
Summary
Gaussian mixture models with universal backgrounds (UBMs) have become the standard method for speaker recognition. Typically, a speaker model is constructed by MAP adaptation of the means of the UBM. A GMM supervector is constructed by stacking the means of the adapted mixture components. A recent discovery is that latent factor analysis of this GMM supervector is an effective method for variability compensation. We consider this GMM supervector in the context of support vector machines. We construct a support vector machine kernel using the GMM supervector. We show similarities based on this kernel between the method of SVM nuisance attribute projection (NAP) and the recent results in latent factor analysis. Experiments on a NIST SRE 2005 corpus demonstrate the effectiveness of the new technique.
Summary
Gaussian mixture models with universal backgrounds (UBMs) have become the standard method for speaker recognition. Typically, a speaker model is constructed by MAP adaptation of the means of the UBM. A GMM supervector is constructed by stacking the means of the adapted mixture components. A recent discovery is that latent...
READ MORE
Evaluation of proposed changes to the ACAS modified tau calculation
May 1, 2006
Conference Paper
Published in:
Int. Civil Aviation Organization Aeronautical Surveillance Panel Working Group, 1 May 2006.
Topic:
R&D area:
R&D group:
Summary
Modified tau is a parameter computed by ACAS to estimate the earliest time at which a collision could occur should an intruder aircraft accelerate toward the own aircraft. A concern with the modified tau calculation has been raised in a class of encounters where intruders are already close and converging slowly. In these problem cases, ACAS may induce a Near Mid-Air Collision by generating RAs with inappropriate timing or initial sense or failing to reverse sense when necessary. Performance in some problem encounters is greatly improved when using several proposed changes to the modified tau equations. These changes are outside CP112E, which focuses only on RA reversals. Although changes to modified tau resolve some problem encounters, aggregate risk-ratio results do not support implementing the existing proposals. There remains a concern about mid-air collision risk due to vulnerability in the existing modified tau equations, yet a robust solution to the problem has not been developed.
Summary
Modified tau is a parameter computed by ACAS to estimate the earliest time at which a collision could occur should an intruder aircraft accelerate toward the own aircraft. A concern with the modified tau calculation has been raised in a class of encounters where intruders are already close and converging...
READ MORE
Update on the analysis of ACAS performance on Global Hawk
May 1, 2006
Conference Paper
Published in:
Int. Civil Aviation Organization Aeronautical Surveillance Panel Working Group, 1 May 2006.
Summary
Initial results are presented from a Lincoln Laboratory study of ACAS performance on the Global Hawk UAV. The study has been applying the process outlined in the ICAO ACAS Manual which involves developing UAV airspace encounter models and running fast-time Monte Carlo simulations of encounters. ACAS performance was examined in conventional aircraft vs. conventional aircraft, conventional aircraft vs. non-ACAS Global Hawk, and conventional aircraft vs. ACAS-equipped Global Hawk cases. The existing ICAO and ACASA encounter models were modified to reflect Global Hawk flight characteristics. ACAS performance on Global Hawk was also assessed parametrically across reaction latencies from 0 - 20 s. Global Hawk flight characteristics were shown to have a small but measurable negative impact on collision risk. Assuming no system failures or visual acquisition effects occur, performance with ACAS on Global Hawk is significantly better than without ACAS if response latencies (from the moment an RA is issued to the moment maneuvering begins) are less than 10 s. Performance drops off rapidly at latencies greater than 10 s. The needs for improved airspace models and a more in-depth study of the interaction between visual acquisition and ACAS are noted.
Summary
Initial results are presented from a Lincoln Laboratory study of ACAS performance on the Global Hawk UAV. The study has been applying the process outlined in the ICAO ACAS Manual which involves developing UAV airspace encounter models and running fast-time Monte Carlo simulations of encounters. ACAS performance was examined in...
READ MORE
Support vector machines using GMM supervectors for speaker verification
May 1, 2006
Journal Article
Published in:
IEEE Signal Process. Lett., Vol. 13, No. 5, May 2006, pp. 308-311.
Topic:
R&D area:
Summary
Gaussian mixture models (GMMs) have proven extremely successful for text-independent speaker recognition. The standard training method for GMMmodels is to use MAP adaptation of the means of the mixture components based on speech from a target speaker. Recent methods in compensation for speaker and channel variability have proposed the idea of stacking the means of the GMM model to form a GMM mean supervector. We examine the idea of using the GMM supervector in a support vector machine (SVM) classifier. We propose two new SVM kernels based on distance metrics between GMM models. We show that these SVM kernels produce excellent classification accuracy in a NIST speaker recognition evaluation task.
Summary
Gaussian mixture models (GMMs) have proven extremely successful for text-independent speaker recognition. The standard training method for GMMmodels is to use MAP adaptation of the means of the mixture components based on speech from a target speaker. Recent methods in compensation for speaker and channel variability have proposed the idea...
READ MORE
Multifunction phased array radar pulse compression limits
April 25, 2006
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-327
Summary
An active phased array radar with distributed low-peak-power transmit modules requires pulse compression to provide high sensitivity and fine range resolution. A long transmitted pulse, however, has accompanying problems such as a near-range blind zone for the transmitting channel and a loss of other gate data (dead gates) in other channels for a multichannel system. In this report the trade-off between the benefits and costs of pulse compression (lengthening) for multifunction phased array radars (MPARs) are analyzed. Specific results are presented for a three-channel MPAR and a two-channel terminal-area MPAR (TMPAR) that have been proposed as replacement systems for current U.S. civil-sector aircraft anad weather surveillance radar systems. The recommended maximum compression ratio is 130 ofr the MPAR and 80 for the TMPAR. The results are independent of radar peak power and antenna gain, and represent upper bounds. Acutal pulse compression ratios that would be employed are likely to be somewhat less tha these values, based on fulfilling specific sensitivity and scan-time requirements with specific radar physical parameters.
Summary
An active phased array radar with distributed low-peak-power transmit modules requires pulse compression to provide high sensitivity and fine range resolution. A long transmitted pulse, however, has accompanying problems such as a near-range blind zone for the transmitting channel and a loss of other gate data (dead gates) in other...
READ MORE
Support vector machines for speaker and language recognition
April 1, 2006
Journal Article
Published in:
Comput. Speech Lang., Vol. 20, No. 2-3, April/July 2006, pp. 210-229.
Topic:
R&D area:
Summary
Support vector machines (SVMs) have proven to be a powerful technique for pattern classification. SVMs map inputs into a high-dimensional space and then separate classes with a hyperplane. A critical aspect of using SVMs successfully is the design of the inner product, the kernel, induced by the high dimensional mapping. We consider the application of SVMs to speaker and language recognition. A key part of our approach is the use of a kernel that compares sequences of feature vectors and produces a measure of similarity. Our sequence kernel is based upon generalized linear discriminants. We show that this strategy has several important properties. First, the kernel uses an explicit expansion into SVM feature space - this property makes it possible to collapse all support vectors into a single model vector and have low computational complexity. Second, the SVM builds upon a simpler mean-squared error classifier to produce a more accurate system. Finally, the system is competitive and complimentary to other approaches, such as Gaussian mixture models (GMMs). We give results for the 2003 NIST speaker and language evaluations of the system and also show fusion with the traditional GMM approach.
Summary
Support vector machines (SVMs) have proven to be a powerful technique for pattern classification. SVMs map inputs into a high-dimensional space and then separate classes with a hyperplane. A critical aspect of using SVMs successfully is the design of the inner product, the kernel, induced by the high dimensional mapping...
READ MORE