Noise Compensation Technique Research Articles

Testing dialogue systems by means of automatic generation of conversations

This paper presents a novel technique that allows testing spoken dialogue systems by means of an automatic generation of conversations. The technique permits to easily test spoken dialogue systems under a variety of lab-simulated conditions, as it is easy to vary or change the utterance corpus used to check the performance of the system. The technique is based on the use of a module called user simulator whose purpose is to behave as real users when they interact with dialogue systems. The behaviour of the simulator is decided by means of diverse scenarios that represent the goals of the users. The simulator aim is to achieve the goals set in the scenarios during the interaction with the dialogue system. We have applied the technique to test a dialogue system developed in our lab. The test has been carried out considering different levels of white and babble noise as well as a VTS noise compensation technique. The results prove that the dialogue system performance is worse under the babble noise conditions. The VTS technique has been effective when dealing with noisy utterances and has lead to better experimental results, particularly for the white noise. The technique has permitted to detect problems in the dialogue strategies employed to handle confirmation turns and recognition errors, suggesting that these strategies must be improved.

Active noise compensation for multichannel magnetocardiography in an unshielded environment

A multichannel high-T/sub c/-SQUID-based heart scanner for unshielded environments is under development, Outside a magnetically shielded room, sensitive SQUID measurements are possible using gradiometers. However, it is difficult to realize large-baseline gradiometers in high-T/sub c/ materials, Therefore, the authors developed two active noise compensation techniques. In the Total Field Compensation technique, a Helmholtz type coil set is placed around the sensors. One magnetometer is used as a zero detector controlling the compensation current through the coil set. For Individual Flux Compensation, the reference signal is sent to the separate SQUIDs (or their flux transformer circuits) to compensate the local environmental noise fluxes, The latter technique was tested on low-T/sub c/ rf-SQUID magnetometers, each sensor set to a field resolution SQUID magnetometers, i.e. 0.1 pT/sub RMS///spl radic/Hz. The authors were able to suppress the environmental disturbances to such an extent that magnetocardiograms could be recorded in an ordinary environment. Here the two suppression techniques are described and experimental results are presented. >

A simple approach to improve the S/N ratio for TEM data using multiple receivers

Since electromagnetic (EM) techniques are very sensitive to cultural noise, many cases exist where standard processing techniques do not yield satisfactory signal‐to‐noise ratios (S/N). For EM surveys in areas with coherent regional EM noise we have developed a local noise compensation (LNC) technique. This technique, for transient EM methods, uses several mobile receivers referenced to a nearby (local) base station. LNC is applied as a prestack processing technique, and it is intended for multichannel systems with close receiver spacings. The main premise for LNC is that the regional noise is highly correlated over a region around the base station and that any other, more localized, noise is small. These assumptions must be tested at the beginning of the survey. The base station remains fixed with measurements being recorded continuously. Meanwhile, measurements are made at various other mobile stations in the zone of coherent noise. During the subsequent signal processing and removal of the transient signal, the actual noise at the base station is calculated as a function of time. Then, this noise is subtracted from the time series measured at the mobile sites. The method results in improvements in data quality, in depth of investigation and in productivity. Field tests carried out in Germany demonstrate results in an environment where the initial test showed that the necessary assumptions were not satisfied, although improvements in the S/N of the single records were obtained. Field tests in the Peoples Republic of China show a successful application of the technique, whereby the length of the usable signal was increased by a factor of about ten.

Noise compensation for autoregressive spectral estimates

The autoregressive spectral estimator possesses excellent resolution properties for time series which satisfy the "all-pole" assumption. When noise is added to the time series under analysis, the resolution of the spectral estimator decreases rapidly as the signal-to-noise ratio decreases. The usual approach to this problem is to model the resulting time series by the more appropriate autoregressive-moving average process and to use standard time series analysis techniques to identify the autoregressive parameters. This standard technique, however, does not result in a positive-definite autocorrelation matrix. As a result, it is shown that the resulting spectral estimator may exhibit a large increase in variance. An alternative approach, termed the noise compensation technique, is proposed. It attempts to correct the estimated reflection coefficients for the effect of white noise, assuming the noise variance is known or can be estimated. Simulation results indicate that a significant decrease in the degrading effects of noise may be realized using the noise compensation technique.