Publications

Refine Results

(Filters Applied) Clear All

Spatial and temporal analysis of weather radar reflectivity images

April 9, 1987
  |
Conference Paper
Author:
Mark W. Merritt
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, 6-9 April 1987, pp. 606-609.
Topic:
weather
R&D area:
R&D group:

Summary

This paper illustrates the use of a primitive symbolic description of an image to obtain more robust identification of amorphous objects than would be possible with more conventional edge or gradient-based segmentation techniques. An algorithm is described which uses a simple multi-level thresholding operation to form a symbolic representation of weather radar reflectivity images. This representation allows the use of detailed rules for the detection and quantification of the image features. A method is described for using this information to identify significant intensity peaks in an image, and examples of its performance are shown.
READ LESS

Summary

This paper illustrates the use of a primitive symbolic description of an image to obtain more robust identification of amorphous objects than would be possible with more conventional edge or gradient-based segmentation techniques. An algorithm is described which uses a simple multi-level thresholding operation to form a symbolic representation of...

READ MORE
Spatial and temporal analysis of weather radar reflectivity images

A block diagram compiler for a digital signal processing MIMD computer

April 9, 1987
  |
Conference Paper
Author:
Marc A. Zissman
…
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. 4, 6-9 April 1987, pp. 1867-1870.
Topic:
signal processing
R&D area:
R&D group:

Summary

A Block Diagram Compiler (BOC) has been designed and implemented for converting graphic block diagram descriptions of signal processing tasks into source code to be executed on a Multiple Instruction Stream - Multiple Data Stream (MIMD) array computer. The compiler takes as input a block diagram of a real-time DSP application, entered on a graphics CAE workstation, and translates it into efficient real-time assembly language code for the target multiprocessor array. The current implementation produces code for a rectangular grid of Texas Instruments TMS32010 signal processors built at Lincoln Laboratory, but the concept could be extended to other processors or other geometries in the same way that a good assembly language programmer would write it. This report begins by examining the current implementation of the BOC including relevant aspects of the target hardware. Next, we describe the task-assignment module, which uses a simulated annealing algorithm to assign the processing tasks of the DSP application to individual processors in the array. Finally, our experiences with the current version of the BOC software and hardware are reported.
READ LESS

Summary

A Block Diagram Compiler (BOC) has been designed and implemented for converting graphic block diagram descriptions of signal processing tasks into source code to be executed on a Multiple Instruction Stream - Multiple Data Stream (MIMD) array computer. The compiler takes as input a block diagram of a real-time DSP...

READ MORE
A block diagram compiler for a digital signal processing MIMD computer

Mixed-phase deconvolution of speech based on a sine-wave model

April 9, 1987
  |
Conference Paper
Author:
Thomas F. Quatieri
…
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. 2, 6-9 April 1987, pp. 649-652.
Topic:
signal processing
R&D area:
R&D group:

Summary

This paper describes a new method of deconvolving the vocal cord excitation and vocal tract system response. The technique relies on a sine-wave representation of the speech waveform and forms the basis of an analysis-synthesis method which yields synthetic speech essentially indistinguishable from the original. Unlike an earlier sinusoidal analysis-synthesis technique that used a minimum-phase system estimate, the approach in this paper generates a "mixed-phase" system estimate and thus an improved decomposition of excitation and system components. Since a mixed-phase system estimate is removed from the speech waveform, the resulting excitation residual is less dispersed than the previous sinusoidal-based excitation estimate of the more commonly used linear prediction residual. A method of time-varying linear filtering is given as an alternative to sinusoidal reconstruction, similar to conventional time-domain synthesis used in certain vocoders, but without the requirement of pitch and voicing decisions. Finally, speech modification with a mixed-phase system estimate is shown to be capable of more closely preserving waveform shape in time-scale and pitch transformations than the earlier approach.
READ LESS

Summary

This paper describes a new method of deconvolving the vocal cord excitation and vocal tract system response. The technique relies on a sine-wave representation of the speech waveform and forms the basis of an analysis-synthesis method which yields synthetic speech essentially indistinguishable from the original. Unlike an earlier sinusoidal analysis-synthesis...

READ MORE
Mixed-phase deconvolution of speech based on a sine-wave model

Multi-style training for robust isolated-word speech recognition

April 9, 1987
  |
Conference Paper
Author:
Richard P. Lippmann
…
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. 2, 6-9 April 1987, pp. 705-708.
Topic:
speech recognition
R&D area:
R&D group:

Summary

A new training procedure called multi-style training has been developed to improve performance when a recognizer is used under stress or in high noise but cannot be trained in these conditions. Instead of speaking normally during training, talkers use different, easily produced, talking styles. This technique was tested using a speech data base that included stress speech produced during a workload task and when intense noise was presented through earphones. A continuous-distribution talker-dependent Hidden Markov Model (HMM) recognizer was trained both normally (5 normally spoken tones) and with multi-style training (one token each from normal, fast, clear, loud, and question-pitch talking styles). The average error rate under stress and normal conditions fell by more than a factor of two with multi-style training and the average error rate under conditions sampled during training fell by a factor of four.
READ LESS

Summary

A new training procedure called multi-style training has been developed to improve performance when a recognizer is used under stress or in high noise but cannot be trained in these conditions. Instead of speaking normally during training, talkers use different, easily produced, talking styles. This technique was tested using a...

READ MORE
Multi-style training for robust isolated-word speech recognition

Two-stage discriminant analysis for improved isolated-word recognition

April 9, 1987
  |
Conference Paper
Author:
Edward A. Martin
…
Published in:
Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing, ICASSP, Vol. 2, 6-9 April 1987, pp. 709-712.
Topic:
speech recognition
R&D area:
R&D group:

Summary

This paper describes a two-stage isolated word search recognition system that uses a Hidden Markov Model (HMM) recognizer in the first stage and a discriminant analysis system in the second stage. During recognition, when the first-stage recognizer is unable to clearly differentiate between acoustically similar words such as "go" and "no" the second-stage discriminator is used. The second-stage system focuses on those parts of the unknown token which are most effective at discriminating the confused words. The system was tested on a 35 word, 10,710 token stress speech isolated word data base created at Lincoln Laboratory. Adding the second-stage discriminating system produced the best results to date on this data base, reducing the overall error rate by more than a factor of two.
READ LESS

Summary

This paper describes a two-stage isolated word search recognition system that uses a Hidden Markov Model (HMM) recognizer in the first stage and a discriminant analysis system in the second stage. During recognition, when the first-stage recognizer is unable to clearly differentiate between acoustically similar words such as "go" and...

READ MORE
Two-stage discriminant analysis for improved isolated-word recognition

An introduction to computing with neural nets

April 1, 1987
  |
Journal Article
Author:
Richard P. Lippmann
Published in:
IEEE ASSP Mag., Vol. 4, No. 2, April 1987, pp. 4-22.
Topic:
machine learning
R&D area:
R&D group:

Summary

Artificial neural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech and image recognition. These models are composed of many nonlinear computational elements operating in parallel and arranged in patterns reminiscent of biological neural nets. Computational elements or nodes are connected via weights that are typically adapted during use to improve performance. There has been a recent resurgence in the field of artificial neural nets caused by new net topologies and algorithms, analog VLSI implementation techniques, and the belief that massive parallelism is essential for high performance speech and image recognition. This paper provides an introduction to the field of artificial neural nets by reviewing six important neural net models that can be used for pattern classification. These nets are highly parallel building blocks that illustrate neural net components and design principles and can be used to construct more complex systems. In addition to describing these nets, a major emphasis is placed on exploring how some existing classification and clustering algorithms can be performed using simple neuron-like components. Single-layer nets can implement algorithms required by Gaussian maximum-likelihood classifiers and optimum minimum-error classifiers for binary patterns corrupted by noise. More generally, the decision regions required by any classification algorithm can be generated in a straightforward manner by three-layer feed-forward nets.
READ LESS

Summary

Artificial neural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech and image recognition. These models are composed of many nonlinear computational elements operating in parallel and arranged in patterns reminiscent of biological neural nets. Computational elements or nodes...

READ MORE
An introduction to computing with neural nets

Recognizing low-altitude wind shear hazards from doppler weather radar: an artificial intelligence approach

March 1, 1987
  |
Journal Article
Author:
Steven D. Campbell
…
Published in:
J. Atmos. Oceanic Technol., Vol. 4, No. 1, March 1987, pp. 5-18.
Topic:
aviation weather
R&D area:
R&D group:

Summary

This paper describes an artificial intelligence-based approach for automated recognition of wind shear hazards. The design of a prototype system for recognizing low-altitude wind shear events from Doppler radar displays is presented. This system, called WXI, consists of a conventional expert system augmented by a specialized capability for processing radar images. The radar image processing component of the system employs numerical and computer vision techniques to extract features from radar data. The expert system carries out symbolic reasoning on these features using a set of heuristic rules expressing meteorological knowledge about wind shear recognition. Results are provided demonstrating the ability of the system to recognize microburst and gust front wind shear events.
READ LESS

Summary

This paper describes an artificial intelligence-based approach for automated recognition of wind shear hazards. The design of a prototype system for recognizing low-altitude wind shear events from Doppler radar displays is presented. This system, called WXI, consists of a conventional expert system augmented by a specialized capability for processing radar...

READ MORE
Recognizing low-altitude wind shear hazards from doppler weather radar: an artificial intelligence approach

Correcting wind speed measurements for site obstructions

January 16, 1987
  |
Conference Paper
Author:
Marilyn M. Wolfson
…
Published in:
Sixth Symp. Meteorological Observations and Instrumentation, 12-16 January 1987, pp. 358-363.
Topic:
aviation weather
R&D area:
R&D group:

Summary

The FLOWS (FAA-Lincoln Laboratory Operational Weather Studies) Project is developing methods for automatically detecting and warning against aviation weather hazards, such as low-altitude wind shear, in airport terminal areas using NEXRAD-like Doppler weather radars. Currently, the FAA uses the Low Level Wind Shear Alert System (LLWAS), an anemometer array situated within and around an airport terminal area, for real-time detection of wind shear events. Even with the installation of Terminal Doppler Weather Radars (TDWRs) at some airports, the LLWAS systems there could still play an important role in the accurate detection of wind shear events, and at airports without TDWRs, the LLWAS will remain the primary detection system. The slowing or obstruction of wind by local obstacles is a well know n problem to those wishing to make accurate wind speed measurements. Anemometers should always be located where there will be, as nearly as passible, an unobstructed wind flow free from turbulent eddies in all directions. Because of the fairly precise required sensor configuration of the anemometers in an LLWAS system, it can occasionally be difficult or impossible to find sites with good exposure in all directions. The FLOWS project is interested in the unobstructed wind speed measurements for two main reasons. First, when analyzing a snapshot of the wind field over a mesonet (or LLWAS) for horizontal wind shear and/or for comparison with Doppler radar data, use of the measured, uncorrected winds would reveal spurious patterns of divergence or vorticity that depend little on time but greatly on the prevailing wind direction and that would, in some cases, obscure the true wind shear pattern. Second, when using surface wind measurements to estimate winds aloft that might be encountered by an aircraft on take-off or landing, an· appropriate power law can be accurately used if the original surface wind speed measurements are representative of the unobstructed flow.
READ LESS

Summary

The FLOWS (FAA-Lincoln Laboratory Operational Weather Studies) Project is developing methods for automatically detecting and warning against aviation weather hazards, such as low-altitude wind shear, in airport terminal areas using NEXRAD-like Doppler weather radars. Currently, the FAA uses the Low Level Wind Shear Alert System (LLWAS), an anemometer array situated...

READ MORE
Correcting wind speed measurements for site obstructions

The FLOWS automatic weather station network

January 12, 1987
  |
Conference Paper
Author:
Marilyn M. Wolfson
Published in:
Proc. Sixth Symp. on Meterological Observations and Instrumentation, 12-16 January 1987, pp. 294-9.
Topic:
aviation weather
R&D area:
R&D group:

Summary

Lincoln Laboratory is operating a network of 30 automatic weather stations for the FAA as part of the ongoing FLOWS (FAA-Lincoln Laboratory Operational Weather Studies) Project, which focuses on developing techniques for automated hazardous weather detection in airport terminal areas using NEXRAD-like Doppler weather radars. The stations, designed to measure temperature, relative humidity, barometric pressure, wind speed, wind direction, and precipitation amounts, originally used one of the first commercially available data collection platforms (DCPs) to transmit 5-min averaged data to the GOES satellites. Under FAA sponsorship, Lincoln has procured modern DCPs and has refurbished amd modified the sensors to create a reliable 30 station network capable of transmitting one minute averages of the variables mentioned above, as well as the peak wind speed each minute and some internal diagnostic variables, on a single GOES satellite channel. The complete system is described and some performance results from the FLOWS 1984-1985 Memphis operation are presented.
READ LESS

Summary

Lincoln Laboratory is operating a network of 30 automatic weather stations for the FAA as part of the ongoing FLOWS (FAA-Lincoln Laboratory Operational Weather Studies) Project, which focuses on developing techniques for automated hazardous weather detection in airport terminal areas using NEXRAD-like Doppler weather radars. The stations, designed to measure...

READ MORE
The FLOWS automatic weather station network

Speech transformations based on a sinusoidal representation

December 1, 1986
  |
Journal Article
Author:
Thomas F. Quatieri
…
Published in:
IEEE Trans. Acoust. Speech Signal Process., Vol. ASSP-34, No. 6, December 1986, pp. 1449-1464.
Topic:
speech enhancement
R&D area:
R&D group:

Summary

In this paper a new speech analysis/synthesis technique is presented which provides the basis for a general class of speech transformations including time-scale modification, frequency scaling, and pitch modification. These modifications can be performed with a time-varying change, permitting continuous adjustment of a speaker's fundamental frequency rate of articulation. The method is based on a sinusoidal representation of the speech production mechanism which has been shown to produce synthetic speech that preserves the waveform shape and is perceptually indistinguishable from the original. Although the analysis/synthesis system was originally designed for single speaker signals, it is also capable ot recovering and modifying non-speech signals such as music, multiple speakers, marine biologic sounds, and speakers in the presence of interferences such as noise and musical backgrounds.
READ LESS

Summary

In this paper a new speech analysis/synthesis technique is presented which provides the basis for a general class of speech transformations including time-scale modification, frequency scaling, and pitch modification. These modifications can be performed with a time-varying change, permitting continuous adjustment of a speaker's fundamental frequency rate of articulation. The...

READ MORE
Speech transformations based on a sinusoidal representation
2001-2010 of 2265