Publications
Magnitude-only estimation of handset nonlinearity with application to speaker recognition
May 15, 1998
Conference Paper
Published in:
Proc. of the 1998 IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. II, Speech Processing II; Neural Networks for Signal Processing, 12-15 May 1998, pp. 745-748.
Topic:
R&D area:
Summary
A method is described for estimating telephone handset nonlinearity by matching the spectral magnitude of the distorted signal to the output of a nonlinear channel model, driven by an undistorted reference. The "magnitude-only" representation allows the model to directly match unwanted speech formants that arise over nonlinear channels and that are a potential source of degradation in speaker and speech recognition algorithms. As such, the method is particularly suited to algorithms that use only spectral magnitude information. The distortion model consists of a memoryless polynomial nonlinearity sandwiched between two finite-length linear filters. Minimization of a mean-squared spectral magnitude error, with respect to model parameters, relies on iterative estimation via a gradient descent technique, using a Jacobian in the iterative correction term with gradients calculated by finite-element approximation. Initial work has demonstrated the algorithm's usefulness in speaker recognition over telephone channels by reducing mismatch between high- and low-quality handset conditions.
Summary
A method is described for estimating telephone handset nonlinearity by matching the spectral magnitude of the distorted signal to the output of a nonlinear channel model, driven by an undistorted reference. The "magnitude-only" representation allows the model to directly match unwanted speech formants that arise over nonlinear channels and that...
READ MORE
Global validation of single-station Schumann resonance lightning location
May 1, 1998
Journal Article
Published in:
J. Atmos. Sol.-Terr. Phys., Vol. 60, No. 7-9., May-June 1998, pp. 701-712.
Topic:
R&D area:
R&D group:
Summary
Global measurements of large, optically bright lightning events from the Optical Transient Detector (OTD) satellite are used to validate estimates of lightning location from single-station Schumann resonance (SR) data. Bearing estimates are obtained through conventional magnetic direction-finding techniques, while source range is estimated from the range-dependent impedance spectrum of an individual SR transients. An analysis of 40 such transients suggests that single-station techniques can locate lightning globally with an accuracy of 1-2 Mm. This is confirmed by further validation at close ranges from flashes detected by the National Lightning Detection Network (NLDN). Observations with both OTD and SR systems may be useful for globally locating lightning with necessary, if not sufficient, characteristics to trigger mesospheric sprites.
Summary
Global measurements of large, optically bright lightning events from the Optical Transient Detector (OTD) satellite are used to validate estimates of lightning location from single-station Schumann resonance (SR) data. Bearing estimates are obtained through conventional magnetic direction-finding techniques, while source range is estimated from the range-dependent impedance spectrum of an...
READ MORE
Multilateration on Mode S and ATCRBS signals at Atlanta's Hartsfield Airport
January 8, 1998
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-260
Topic:
R&D area:
R&D group:
Summary
The ATC community is seeking a way to obtain aircraft ID and improved surveillance on the airport movement area. Surface radars provide good surveillance data, but do not provide ID, may not cover the whole movement area, and suffer from false reflection targets and performance degradations in rain. This report describes an evolutionary technique employing multilateration, TCAS technology, and existing ATCBI transponders to provide the desired surface surveillance information. Five multilateration receiver/transmitters (RTs) based on TCAS units, and a central multilateration computer processor were procured and installed on the highest available buildings on the perimeter of the north side of Atlanta's Hartsfield airport. The resulting coverage was such that there was a 93% probability that a multilateration position would be computed on a given Mode S short squitter emitted from a a target at a randomly selected position on the movement area. Multilateration was performed on ATCRBS targets using replies elicited by whisper shout methods originally developed for TCAS. Measurements showed that whisper shout was successful in degarbling targets that were in close proximity on the movement area. The probability of obtaining an ATCRBS multilateration position in a given one second interval depended on the number of whisper shout interrogations transmitted. The equipment required over 10 interrogations per target per second to obtain per second multilateration update rates on two typical targets of 58% and 83% respectively. This less than anticipated performance was primarily due to the inefficient whisper shout interrogation technique that was used in the test equipment. This can be corrected in next generation equipment. The multilateration accuracy was about 20 feet one sigma, as anticipated from theoretical considerations and previous experience with other equipment. By combining the multilateration data with ASDE data and tracking the results, it would be possible to obtain track reliabilities on the airport surface similar to that obtained elsewhere in the ATC system but update rates of 1Hz as required for surface surveillance and control purposes. The RTs were also capable of receiving Mode S long squitters containing GPS position information. The probability of at least one of the 5RTs receiving a given long squitter was essentially 100% on the movement area.
Summary
The ATC community is seeking a way to obtain aircraft ID and improved surveillance on the airport movement area. Surface radars provide good surveillance data, but do not provide ID, may not cover the whole movement area, and suffer from false reflection targets and performance degradations in rain. This report...
READ MORE
Evaluation of Boeing 747-400 performance during ATC-directed breakouts on final approach
January 7, 1998
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-263
Topic:
R&D area:
Summary
The effects of three different levels of pilot training on the breakout response of pilots and the Boeing 747-400 aircraft were studied. The study examined response during ATC-directed breakouts on final approach and was conducted in three phases. Phase 1 tested performance during manual and autopilot-coupled approaches given current procedures and pilot training. Phase 2 tested the effect of increased pilot situational awareness and proposed ATC breakout phraseology on breakouts during manual and autopilot-coupled approaches. Phase 3 tested the effect of two B747-400-specific breakout procedures on breakouts during autopilot-coupled approaches. Pilot preferences regarding procedures and the tested training materials were also solicited.
Summary
The effects of three different levels of pilot training on the breakout response of pilots and the Boeing 747-400 aircraft were studied. The study examined response during ATC-directed breakouts on final approach and was conducted in three phases. Phase 1 tested performance during manual and autopilot-coupled approaches given current procedures...
READ MORE
Audio signal processing based on sinusoidal analysis/synthesis
January 1, 1998
Book Chapter
Published in:
Chapter 9 in Applications of Digital Signal Processing to Audio and Acoustics, 1998, pp. 343-416.
Topic:
R&D area:
Summary
Based on a sinusoidal model, an analysis/synthesis technique is developed that characterizes audio signals, such as speech and music, in terms of the amplitudes, frequencies, and phases of the component sine waves. These parameters are estimated by applying a peak-picking algorithm to the short-time Fourier transform of the input waveform. Rapid changes in the highly resolved spectral components are tracked by using a frequency-matching algorithm and the concept of "birth" and "death" of the underlying sine waves. For a given frequency track, a cubic phase function is applied to the sine-wave generator, whose output is amplitude-modulated and added to sines for other frequency tracks. The resulting synthesized signal preserves the general wave form shape and is nearly perceptually indistinguishable from the original, thus providing the basis for a variety of applications including signal modification, sound splicing, morphing and extrapolation, and estimation of sound characteristics such as vibrato. Although this sine-wave analysis/synthesis is applicable to arbitrary signals, tailoring the system to a specific sound class can improve performance. A source/filter phase model is introduced within the sine-wave representation to improve signal modification, as in time-scale and pitch change and dynamic range compression, by attaining phase coherence where sinewave phase relations are preserved or controlled. A similar method of achieving phase coherence is also applied in revisiting the classical phase vocoder to improve modification of certain signal classes. A second refinement of the sine-wave analysis/synthesis invokes an additive deterministic/stochastic representation of sounds consisting of simultaneous harmonic and aharmonic contributions. A method of frequency tracking is given for the separation of these components, and is used in a number of applications. The sinewave model is also extended to two additively combined signals for the separation of simultaneous talkers or music duets. Finally, the use of sine-wave analysis/synthesis in providing insight for FM synthesis is described, and remaining challenges, such as an improved sine-wave representation of rapid attacks and other transient events, are presented.
Summary
Based on a sinusoidal model, an analysis/synthesis technique is developed that characterizes audio signals, such as speech and music, in terms of the amplitudes, frequencies, and phases of the component sine waves. These parameters are estimated by applying a peak-picking algorithm to the short-time Fourier transform of the input waveform...
READ MORE
The Lincoln Near-Earth Asteroid Research (LINEAR) Program
January 1, 1998
Journal Article
Published in:
Lincoln Laboratory Journal, Vol. 11, No. 1, 1998, pp. 27-40.
Topic:
R&D area:
R&D group:
Summary
Lincoln Laboratory has been developing electro-optical space-surveillance technology to detect, characterize, and catalog satellites for more than forty years. Recent advances in highly sensitive, large-format charge-coupled devices (CCDs) allow this technology to be applied to detecting and cataloging asteroids, including near-Earth objects (NEOs). When equipped with a new Lincoln Laboratory focal-plane camera and signal processing technology, the 1-m U.S. Air Force ground-based electro-optical deep-space surveillance (GEODSS) telescopes can conduct sensitive large-coverage searches for Earth-crossing and main-belt asteroids. Field measurements indicate that these enhanced telescopes can achieve a limiting magnitude of 22 over a 2-deg2 field of view with less than 100 sec of integration. This sensitivity rivals that of much larger telescopes equipped with commercial cameras. Working two years under U.S. Air Force sponsorship, we have developed technology for asteroid search operations at the Lincoln Laboratory Experimental Test Site near Socorro, New Mexico. By using a new large-format 2560 X 1960-pixel frame-transfer CCD camera, we have discovered over 10,000 asteroids, including 53 NEOs and 4 comets as designated by the Minor Planet Center (MPC). In March 1998, the Lincoln Near-Earth Asteroid Research (LINEAR) program provided over 150,000 observations of asteroids--nearly 90% of the world's asteroid observations that month--to the MPC, which resulted in the discovery of 13 NEOs and 1 comet. The MPC indicates that the LINEAR program outperforms all asteroid search programs operated to date.
Summary
Lincoln Laboratory has been developing electro-optical space-surveillance technology to detect, characterize, and catalog satellites for more than forty years. Recent advances in highly sensitive, large-format charge-coupled devices (CCDs) allow this technology to be applied to detecting and cataloging asteroids, including near-Earth objects (NEOs). When equipped with a new Lincoln Laboratory...
READ MORE
The effects of compression-induced distortion of graphical weather images on pilot perception, acceptance, and performance
November 21, 1997
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-243
Topic:
R&D area:
R&D group:
Summary
The Graphical Weather Service (GWS) is a data link application that will provide near-real-time graphical weather information to pilots in flight. To assess the effect GWS, as well as to aid in the proper design, implementation and certification of the use of GWS in aircraft, two human factors studies have been conducted. The second study conducted (Phase Two) is the topic of this report. Phase Two was conducted to determine the maximum level of compression-induced distortion that would be acceptable for transmission of weather images to the cockpit. To make this determination the following data were collected and analyzed: pilot subjective ratings of the perceived amount of distortion of a compressed image, pilot subjective ratings of the acceptability of a compressed image for use in the flight task, and pilot route selections as a function of the amount of compression presented in an image. Results indicated that images of low to moderate compression levels were generally acceptable for transmission to the cockpit, while images that were highly compressed were generally unacceptable. In addition, computed measures of image quality have been identified to enable the establishment of a criteria for transmitting images to aircraft.
Summary
The Graphical Weather Service (GWS) is a data link application that will provide near-real-time graphical weather information to pilots in flight. To assess the effect GWS, as well as to aid in the proper design, implementation and certification of the use of GWS in aircraft, two human factors studies have...
READ MORE
High-performance low-complexity wordspotting using neural networks
November 1, 1997
Journal Article
Published in:
IEEE Trans. Signal Process., Vol. 45, No. 11, November 1997, pp. 2864-2870.
Topic:
R&D area:
Summary
A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neural networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker independent Switchboard corpus. Figure of Merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric, and voice transformations generate additional training examples by warping the spectra of training data to mimic across-talker vocal tract variability.
Summary
A high-performance low-complexity neural network wordspotter was developed using radial basis function (RBF) neural networks in a hidden Markov model (HMM) framework. Two new complementary approaches substantially improve performance on the talker independent Switchboard corpus. Figure of Merit (FOM) training adapts wordspotter parameters to directly improve the FOM performance metric...
READ MORE
The Weather-Huffman method of data compression of weather images
October 31, 1997
Project Report
Published in:
MIT Lincoln Laboratory Report ATC-261
Topic:
R&D area:
R&D group:
Summary
Providing an accurate picture of the weather conditions in the pilot's area of interest is a highly useful application for ground-to-air datalinks. The problem with using data links to transmit weather graphics is the large number of bits required to exactly specify the weather image. To make transmission of weather images practical, a means must be found to compress the data to a size compatible with a limited datalink capacity. The Weather-Huffman (WH) Algorithm developed in this report incorporates several subalgorithms in order to encode as faithfully as possible an input weather image within a specified datalink bit limitation. The main algorithm component is the encoding of a version of the input image via the Weather Huffman runlength code, a variant of the standard Huffman code tailored to the peculiarities of weather images. If possible, the input map itself is encoded. Generally, however, a resolution-reduced version of the map must be created prior to the encoding to meet the bit limitation. In that case, the output map will contain blocky regions, and higher weather level areas will tend to bloom in size. Two routines are included in WH to overcome these problems. The first is a Smoother Process, which corrects the blocky edges of weather regions. The second, more powerful routine, is the Extra Bit Algorithm (EBA). EBA utilizes all bits remaining in the message after the Huffman encoding to correct pixels set at too high a weather level. Both size and shape of weather regions are adjusted by this algorithim. Pictorial examples of the operation of this algorithm on several severe weather images derived from NEXRAD are presented.
Summary
Providing an accurate picture of the weather conditions in the pilot's area of interest is a highly useful application for ground-to-air datalinks. The problem with using data links to transmit weather graphics is the large number of bits required to exactly specify the weather image. To make transmission of weather...
READ MORE
Noise reduction based on spectral change
October 22, 1997
Conference Paper
Published in:
Proc. of the 1997 IEEE ASSP Workshop on Applications of Signal Processing to Audio and Acoustics, Session 8: Noise Reduction, 19-22 October 1997, 4 pages.
Topic:
R&D area:
Summary
A noise reduction algorithm is designed for the aural enhancement of short-duration wideband signals. The signal of interest contains components possibly only a few milliseconds in duration and corrupted by nonstationary noise background. The essence of the enhancement technique is a Weiner filter that uses a desired signal spectrum whose estimation adapts to the "degree of stationarity" of the measured signal. The degree of stationarity is derived from a short-time spectral derivative measurement, motivated by sensitivity of biological systems to spectral change. Adaptive filter design tradeoffs are described, reflecting the accuracy of signal attack, background fidelity, and perceptual quality of the desired signal. Residual representations for binaural presentation are also considered.
Summary
A noise reduction algorithm is designed for the aural enhancement of short-duration wideband signals. The signal of interest contains components possibly only a few milliseconds in duration and corrupted by nonstationary noise background. The essence of the enhancement technique is a Weiner filter that uses a desired signal spectrum whose...
READ MORE