August 2014
Spotlight Summary by Nathan Newbury
Range-resolved vibrometry using a frequency comb in the OSCAT configuration
One of the intriguing uses for laser Doppler vibrometry is to eavesdrop on a conversation by picking up the vibrations from surrounding surfaces. The principle is simple. A person’s voice will cause a nearby surface to vibrate. If a laser beam is shined on the surface, then the back-reflected laser beam will have a Doppler shift that is proportional to the vibration-induced surface velocity. The Doppler shift can be measured in a coherent heterodyne receiver to detect the surface vibrations and thereby eavesdrop on the conversation. Of course, in practice, this is a very challenging measurement problem since the signal-to-noise is low and there can be significant background clutter. Here, Boudreau and Genest have demonstrated a range-resolved vibrometer that is based on frequency-comb heterodyne detection in the so-called OSCAT (optical sampling by cavity tuning) configuration. In its essence, the technique allows one to detect the vibration from a particular, range-resolved surface while ignoring the vibrations at other ranges. This technique could find other applications as well, but the paper illustrates its effectiveness by detecting surface vibrations from someone’s voice. In the final demonstration, they successfully detect the voice from wall vibrations while looking through a glass pane that has much stronger applied vibrations (and that would have completely masked the desired voice signal from the wall vibrations in any conventional cw laser vibrometer).
In the OSCAT configuration, the basic layout is a conventional interferometer but the source is a frequency comb. In time, the output of a frequency comb is a series of pulses. This output is split in the interferometer into a signal and local oscillator path. The signal path is amplified, transmitted to the surface, and a portion of the back-reflected signal is then heterodyned with the local oscillator pulse train. The pulsed-nature of the output provides the range resolution. In fact, since the local oscillator is pulsed, the system only “detects” the returning signal pulses that arrive at the detector at the same time as the local oscillator pulses. Therefore by tuning the delay of the local oscillator branch, one can “listen in” to the vibrations from surfaces at different ranges. To avoid baseband noise, it is critical to shift this vibrometry signal to higher rf frequencies. The authors accomplish that task here either through insertion of an acousto-optic modulator in the LO branch or through a triangle sweep of the comb repetition rate that, when combined with the delay line, effectively gives a frequency shift of the heterodyne signal away from baseband.
The basic technique allows for range-resolved detection of the vibration signal. In addition, the authors implement significant amount of signal processing to remove the intrinsic frequency noise of the free-running frequency comb and of the LO path that would otherwise mask the vibration signal. They also implement notch filters to remove 60 Hz effects and finally use a spectral noise gating technique that has been previously shown to work well with audio signals. The final demonstration is given in their “media 8” file where the voice signal (a count from one to ten in French) can be heard despite the much stronger signal from an intervening glass pane. With higher power-aperture products and with continuous processing, it is intriguing to consider the future possibilities of long-range, continuous range-resolved vibrometry.
You must log in to add comments.
In the OSCAT configuration, the basic layout is a conventional interferometer but the source is a frequency comb. In time, the output of a frequency comb is a series of pulses. This output is split in the interferometer into a signal and local oscillator path. The signal path is amplified, transmitted to the surface, and a portion of the back-reflected signal is then heterodyned with the local oscillator pulse train. The pulsed-nature of the output provides the range resolution. In fact, since the local oscillator is pulsed, the system only “detects” the returning signal pulses that arrive at the detector at the same time as the local oscillator pulses. Therefore by tuning the delay of the local oscillator branch, one can “listen in” to the vibrations from surfaces at different ranges. To avoid baseband noise, it is critical to shift this vibrometry signal to higher rf frequencies. The authors accomplish that task here either through insertion of an acousto-optic modulator in the LO branch or through a triangle sweep of the comb repetition rate that, when combined with the delay line, effectively gives a frequency shift of the heterodyne signal away from baseband.
The basic technique allows for range-resolved detection of the vibration signal. In addition, the authors implement significant amount of signal processing to remove the intrinsic frequency noise of the free-running frequency comb and of the LO path that would otherwise mask the vibration signal. They also implement notch filters to remove 60 Hz effects and finally use a spectral noise gating technique that has been previously shown to work well with audio signals. The final demonstration is given in their “media 8” file where the voice signal (a count from one to ten in French) can be heard despite the much stronger signal from an intervening glass pane. With higher power-aperture products and with continuous processing, it is intriguing to consider the future possibilities of long-range, continuous range-resolved vibrometry.
Add Comment
You must log in to add comments.
Article Information
Range-resolved vibrometry using a frequency comb in the OSCAT configuration
Sylvain Boudreau and Jérôme Genest
Opt. Express 22(7) 8101-8113 (2014) View: Abstract | HTML | PDF