Voiceband signal classification using statistically optimal combinations of low-complexity discriminant variables

Sewall, J.S.; Cockburn, B.F.

doi:10.1109/26.803495

“…Past research into automatic identification of modulation type has used a combination of short-term spectral features, i.e. [2]. A weakness of features of this type is that they are sensitive to frequency translations which may be induced during the initial demodulation of the signal into the audible range by the human operator.…”

Section: Signal Interceptionmentioning

confidence: 99%

“…For the spec P vector, four scalar features which were also insensitive to frequency shift were extracted including entropy and bandwidth of spec P , and the mean and standard deviation of the demodulated spectrum estimate. The normalized central second moment of the analytic signal [2] was also used for our features. We refer to these five features collectively as "spectral estimate" features.…”

Section: Data Collectionmentioning

confidence: 99%

Non-Stationary Signal Classification Using Joint Frequency Analysis

Sukittanon¹,

Atlas²,

Pitton³

et al. 2003

View full text Add to dashboard Cite

Time-varying short-term spectral estimates have been successfully applied in many classification tasks. However, they are still insufficient for many non-stationary signals where time-varying information is useful. In this paper, we propose to improve the deficiencies of current short-term feature analysis by adding information to describe the time-varying behavior of the signals. Our proposed method which is motivated by the human auditory system can be applied to several non-stationary signal types. Real world communication signals were used for experimental verification. These experimental results, assessed with a conventional probabilistic classifier, showed significant improvement when the new features were added to short-term spectral estimates.

show abstract

“…The normalized central second moment of the analytic signal [2] was also used for our features. We refer to these five features collectively as "spectral estimate" features.…”

Section: Data Collectionmentioning

confidence: 99%

“…Short-term estimates, for example Fourier or autocorrelation coefficients, have been widely used for many signal types (e.g. [1,2]). Basically, the signal is blocked into successive frames using a small data analysis window.…”

Section: Introductionmentioning

confidence: 99%

Non-stationary signal classification using joint frequency analysis

Sukittanon

¹

,

Atlas

²

,

Pitton³

et al.

2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03).

View full text Add to dashboard Cite

Time-varying short-term spectral estimates have been successfully applied in many classification tasks. However, they are still insufficient for many non-stationary signals where time-varying information is useful. In this paper, we propose to improve the deficiencies of current short-term feature analysis by adding information to describe the time-varying behavior of the signals. Our proposed method which is motivated by the human auditory system can be applied to several non-stationary signal types. Real world communication signals were used for experimental verification. These experimental results, assessed with a conventional probabilistic classifier, showed significant improvement when the new features were added to short-term spectral estimates.

show abstract