Abstract
A method for measuring a response from an optical fiber providing distributed back reflections using a system comprising an optical source comprising a laser, an optical receiver and a processing unit is disclosed. The method comprises establishing initial parameters of a distributed back-reflection processing. The method also comprises generating an interrogation signal and an optical local oscillator using the optical source, the interrogation signal being represented by an interrogation phasor and the optical local oscillator being represented by a local oscillator phasor; transmitting the interrogation signal into the optical fiber; and mixing the optical local oscillator with reflected light from the optical fiber and detecting a mixing product with the optical receiver to achieve a receiver output signal. The method further comprises performing a measurement that characterizes the interrogation phasor; updating the parameters of the distributed back-reflection processing based on the measurement result such that an effect of fluctuations in the interrogation phasor on the measured response from the fiber is reduced; and applying distributed back-reflection processing to the receiver output signal. Finally, the method comprises extracting the response from the optical fiber from the distributed back-reflection processing output. A system for measuring a response from an optical fiber providing distributed back reflections is also disclosed.
Claims
1. A method for measuring a response from an optical fiber providing distributed back reflections using a system comprising an optical source comprising a laser, an optical receiver and a processing unit, the method comprising: establishing initial parameters of a distributed back-reflection processing; generating an interrogation signal and an optical local oscillator using the optical source, the interrogation signal being represented by an interrogation phasor and the optical local oscillator being represented by a local oscillator phasor; transmitting the interrogation signal into the optical fiber; mixing the optical local oscillator with reflected light from the optical fiber and detecting a mixing product with the optical receiver to achieve a receiver output signal; performing a measurement that characterizes the interrogation phasor; updating the parameters of the distributed back-reflection processing based on the measurement result such that an effect of fluctuations in the interrogation phasor on the measured response from the fiber is reduced; applying the distributed back-reflection processing to the receiver output signal; and extracting the response from the optical fiber from the distributed back-reflection processing output.
2. The method of claim 1, wherein performing the measurement provides an estimated interrogation phasor and updating the parameters of the distributed back-reflection processing reduces fluctuations in the output from the distributed back-reflection processing when applied to the estimated interrogation phasor.
3. The method of claim 1, wherein the frequency of the interrogation signal is swept.
4. The method of claim 1, wherein the distributed back-reflection processing comprises applying a pulse compression filter.
5. The method of claim 4, wherein updating the parameters of the distributed back-reflection processing comprises updating the coefficients of the pulse compression filter to an estimate for the conjugated and time-reversed estimated interrogation phasor.
6. The method of claim 1, wherein updating the parameters of the distributed back-reflection processing comprises: establishing a compressed reference phasor, wherein performing the measurement provides an estimated interrogation phasor; applying the pulse compression filter to the estimated interrogation phasor to provide a compressed interrogation phasor; computing a noise suppression filter that when convolved with the compressed interrogation phasor provides a result that resembles the compressed reference pulse; and incorporating convolution with the noise suppression filter into the distributed back-reflection processing.
7. The method of claim 1, further comprising adding a fraction of the interrogation signal to the reflected light from the sensor fiber with a delay that does not overlap with the delays of the distributed back reflections from the sensor fiber, wherein information about the interrogation signal is extracted from the pulse compression filter output around the delay of the added fraction of the interrogation signal.
8. The method of claim 1, wherein the output from the monitor detector is fed back to the laser to minimize the laser phasor fluctuations.
9. The method of claim 1, wherein performing a measurement that characterizes the interrogation phasor comprises measuring a laser frequency detuning.
10. The method of claim 9, wherein measuring a laser frequency detuning comprises measuring the frequency of the laser relative to the laser frequency in the previous interrogation period.
11. The method of claim 9, wherein the distributed back-reflection processing comprises: establishing a compressed reference phasor; applying the distributed back-reflection processing to the estimated interrogation phasor to provide a compressed interrogation phasor; computing a detuning compensation filter that when convolved with the compressed interrogation phasor provides a result that resembles a compressed reference pulse frequency shifted by the measured laser frequency detuning; and incorporating convolution with the noise suppression filter and frequency shifting of the filtered signal according to the measured laser frequency detuning into the distributed back-reflection processing.
12. A system for measuring a response from an optical fiber providing distributed back reflections, the system comprising: an optical source comprising a laser, the optical source being configured for generating an interrogation signal and an optical local oscillator, the interrogation signal being representable by an interrogation phasor, and the optical local oscillator being representable by a local oscillator phasor, the optical source further being configured to transmit the interrogation signal into optical fiber; an optical receiver and signal characterization unit configured to generate and receive a mixing product of the optical local oscillator and reflected light from the optical fiber to provide a receiver output signal, and to perform a measurement that characterizes the interrogation phasor; and a processing unit configured to: establish an initial parameter of distributed back-reflection processing; update the parameters of the distributed back-reflection processing based on the measurement that characterizes the interrogation phasor such that an effect of fluctuations in the interrogation phasor on the measured response from the fiber is reduced; apply the distributed back-reflection processing to the receiver output signal; and extract the response from the optical fiber from the distributed back-reflection processing output.
13. The system of claim 12, wherein the optical source further comprises a modulator.
14. The system of claim 12, further comprising a separate coherent receiver configured to measure the interrogation signal.
15. The system of claim 12, further configured to add a fraction of the interrogation signal to the reflected light from the sensor fiber with a delay that does not overlap with the delays of the distributed back reflections from the sensor fiber.
16. The system of claim 12, further comprising a monitor detector configured for measuring fluctuations in the laser phasor characterizing the laser output.
17. The system of claim 16, wherein the monitor detector comprises an intensity detector.
18. The system of claim 16, wherein the monitor detector comprises an interferometer.
Description
BRIEF DESCRIPTION OF THE FIGURES
(1) FIG. 1 illustrates a system according to the invention,
(2) FIG. 2A-E show alternative implementations of the measurement of M.sub.n(t).
(3) FIG. 3 shows a flow diagram of an embodiment of the method according to the invention.
(4) FIG. 4 illustrates aspects of the invention.
(5) FIG. 5 shows a flow diagram of another embodiment of the method.
(6) FIGS. 6 and 7 are flow diagrams illustrating details of embodiments of the method according to the invention.
DETAILED DESCRIPTION
(7) FIG. 1 illustrates a system 100 according to the invention, the system comprising an optical source 101 providing an interrogation signal 102 and an optical local oscillator 103. The optical local oscillator is directed to an optical receiver and signal characterization unit 106. The interrogation signal is being launched into the optical fiber 104 (called “sensor fiber” in the following) via a tap and backreflection unit 105. Note that the term “sensor fiber” here merely means the fiber to be interrogated, not necessarily that the fiber contains or forms sensors as such. For instance, in some embodiments, the sensor fiber may be a regular optical fiber, such as an SMF-28-type fiber. The interrogation signal 102 is launched into the sensor fiber 104 will be reflected from different sections of the sensor fiber and directed to the optical receiver and signal characterization unit 106 via the tap and backreflection unit 105. The optical receiver and signal characterization unit 106 mixes the reflected light with the optical local oscillator and performs measurements of fluctuations in the local oscillator signal and/or the interrogation signal. The optical receiver and signal characterization unit 106 provides input for a processing unit 107 for further signal processing.
(8) Alternative implementations of the system according to the invention, and measurement of M.sub.n(t) are illustrated in FIGS. 2A-D. Throughout the figures, like reference numerals refer to the same or corresponding parts. Therefore, only differences of one embodiment in comparison with previous embodiments will be discussed.
(9) As mentioned above, laser noise may contribute to errors in the interrogation phasor M.sub.n(t). Such noise may also contribute to local oscillator phase noise. In many cases the laser noise will be dominated by phase noise, while the intensity noise from the laser will be negligible.
(10) To suppress laser induced noise an auxiliary laser frequency monitoring interferometer may be included. The output from this interferometer may be sampled and used in the pulse compression filter processing to suppress the noise contribution from the laser. Such an interferometer may for instance be implemented by leading part of the laser light to a fiber Michelson interferometer with Faraday reflectors (to suppress polarization fading) that is locked into quadrature at low frequencies, typically below the inverse repetition period. The monitoring interferometer can provide a measure for laser frequency fluctuations, which may be integrated with time to obtain the laser phase fluctuations.
(11) Other interferometer topologies than the Michelson design may also be used, such as Mach-Zehnder or Fabry-Perot. As an alternative to quadrature locking, the interferometer may be designed to allow for direct I/Q demodulation of the interferometer phase, for instance by use of a 3×3 coupler, an optical hybrid or by generating a phase generated carrier by either modulating the light passed to the interferometer or placing a modulator inside the interferometer.
(12) The Pound-Drever-Hall (PDH) technique is a sensitive method to measure the phase noise of a laser which can be used to suppress phase noise through feedback to the laser. The PDH error signal can also be sampled and feedforward to the pulse compression filter processing for further suppression of laser phase noise.
(13) Optionally, laser intensity noise may also be monitored, for instance by leading a fraction of the laser light to a separate intensity monitoring detector.
(14) FIG. 2A illustrates the optical source 101 as a laser 200 and modulator 201. The optical receiver and signal characterization unit 106 comprises a coherent receiver 205 to detect the reflected signal from the sensor fiber and a laser monitor 214. The laser monitor 214 comprises a coupler 215 is used to guide a fraction, for instance 20%, of the laser output via coupler 216 to a Michelson interferometer comprised by coupler 217, for instance a 50% coupler, a delay fiber 218, for instance with 100 ns return delay, and two Faraday mirrors 219 and 220. The local oscillator is derived from the laser output such that the measurement of the local oscillator gives a measurement of laser noise. Coupler 216 may for instance be a 30% coupler configured such that 30% of the input power is guided to the Michelson interferometer. The return signals from the Michelson interferometer are guided from coupler 217 to the upper and (via coupler 216) the lower input of the local oscillator characterization receiver 221. By scaling the two resulting detector signals such that they provide approximately equal fringe amplitudes and taking the difference, one obtains a signal that is a sine function of the optical frequency of the local oscillator, or more precisely a sine function of the local oscillator phase change ϕ.sub.n(t) within the dual path delay of the delay fiber 218 given by A sin(ϕ.sub.n(t)−ϕ.sub.n(t−τ.sub.IF))≈A(ϕ.sub.n(t)−ϕ.sub.n(t−τ.sub.IF)), where A is scale factor and τ.sub.IF is the imbalance of the interferometer. The response is proportional to frequency noise of the laser at frequencies<<1/τ.sub.IF, but falls off to a zero at 1/τ.sub.IF. The local oscillator phase can be computed by scaling by 1/A and applying an integrator. This integrator should be bandlimited to cut-off at low frequencies and at a frequency below 1/τ.sub.IF. If this signal is fed back to a mechanism that modulates the source laser frequency, for instance a laser pump current, one may stabilize the laser frequency, while the interferometer is kept close to quadrature, i.e. close to the zero crossing of the mentioned sine function. Alternatively, the interferometer may be kept in quadrature by applying feedback to a mechanism that modulates the delay of the delay coil, for instance via a Peltier temperature modulator. Intensity fluctuations in the local oscillator signal may be monitored by detecting the middle input to the local oscillator characterization receiver via the lower left output of coupler 216. In this way, signals that characterize both the phase and the amplitude fluctuations of the local oscillator signal are made available. These signals could be sampled by some signal characterization ADCs to provide a measurement of fluctuations in M.sub.LO(t) which could be used to remove the effect of local oscillator induced noise from the signals received from the coherent receiver 205. If the interrogation signal is derived from the same laser output as the local oscillator, which is the case in the example illustrated in FIG. 2A, the output from the signal characterization ADCs may be used to estimate for both the noise in M.sub.n(t) and M.sub.LO(t), providing the possibility to suppress the effect of both noise sources. This is exemplified above in equation (7) and the following discussions.
(15) In FIG. 2B a tap coupler 202 is used to guide a fraction, for instance 1%, of the transmitted light to a separate coherent receiver 204 which will provide an updated monitor measurement of M.sub.n(t)
(16) FIG. 2C illustrates that two weak couplers called Tap 202 and Combiner 206 and a delay fiber 207 are arranged in front of the circulator 203, which may define the start of the sensor fiber, in such a way that a fraction of the transmitted light passes through a monitor arm 208 and reaches the receiver before the light reflected from the sensor arm 209. In one preferred implementation of the invention M.sub.n(t) is nonzero in only a part of each repetition period T.sub.rep, and the first sensor reflections are delayed by T.sub.d, which is longer than the nonzero duration of M.sub.n(t). In this way, the full duration of M.sub.n(t) can be measured before the sensor reflections enter the receiver 106, thus avoiding interference with the sensor reflections.
(17) In an alternative implementation the delay T.sub.d can be shorter, and the nonzero duration of M.sub.n(t) may even be close or equal to T.sub.rep. Such implementations may for instance be useful in cases where M.sub.n(t) is a frequency sweep with sweep rate SWR (Hz/s). If the instantaneous frequency of M.sub.n(t) (the phase change rate) deviates from R(t) with less than SWR.Math.T.sub.d, the compressed interrogation phasor will only fluctuate significantly in the delay range [−T.sub.d,T.sub.d]. In that case, the deviations in M.sub.n(t) can be computed from the pulse compression filter output in the delay range±T.sub.d centered around the monitor delay. Such monitoring can be useful in cases where suppression of noise arising from low frequency deviations in M.sub.n(t) is important.
(18) FIG. 2D illustrates the use of a discrete monitor reflector R.sub.mon 210 in the sensor fiber with a reflectivity that significantly exceeds the combined Rayleigh backscattering strength of the sensor fiber within a delay range T.sub.d. may provide a monitor signal at the receiver 106 that dominates clearly above the Rayleigh scattering, and which can be used to monitor fluctuations in a swept signal M.sub.n(t) at frequencies<SWR.Math.T.sub.d. The reflector 210 may for instance be formed by a fiber Bragg grating (FBG), a reflective splice or a connector. Due to the very high amplitude difference between the monitor signal and Rayleigh scattered signals, the receiver will need a very high dynamic range to allow the Rayleigh scattering signal, and thus the sensor phase information, to be extracted. However, if the sensor part of the fiber comprises sensor reflectors R.sub.s that are comparable in strength to the monitor reflector R.sub.mon, such as FBGs, the dynamic range requirement on the receiver will be relieved. In this case, by delaying the first sensor reflector with T.sub.d relative to the monitor reflector, e.g. with a delay fiber 211, it will be possible to monitor fluctuations in a swept signal M.sub.n(t) at frequencies<SWR.Math.T.sub.d. Note that T.sub.d in this case is the dual path delay of the fiber between the two reflectors, as opposed to FIG. 2C where T.sub.d is the combined single path delay of the fiber from the Tap 202 to the circulator 203 and from the circulator 203 to the combiner 206, minus the delay of the direct path from the Tap 202 to the Combiner 206.
(19) FIG. 2E illustrates a technique useful in connection with the present invention, in which an optical switch 212 selectively allows the signal that otherwise is transmitted to the sensor fiber 104 to be routed directly to the receiver 106, to measure the shape of the interrogation phasor M.sub.n(t) for a short period before the interrogation starts. The measurements may be averaged over several repetition periods, to get an estimate M.sub.n(t) for the typical shape of M.sub.n(t). This measurement may provide a basis for the reference interrogation signal R(τ). The deviation between the estimated interrogation phasor and a goal interrogation phasor may be used to compute a premphasis on the drive signal to the modulator to minimize this deviation.
(20) FIG. 3 illustrates an embodiment of the method 300 according to the invention, for compensating for fluctuations in the interrogation signal. In a first step 302, the interrogation signal is generated and measured, before the interrogation signal is launched into the sensor fiber. Based on the measurement of the interrogation signal, processing parameters of the signal processing are adjusted 304, to minimize an influence of the interrogation signal fluctuations on the signal detected from the sensor fiber. The distributed backreflection processing is applied with the updated processing parameters in step 306
(21) FIG. 4 exemplifies the real part of the compressed reference phasor 402, A(τ), and the compressed interrogation phasor 404, X.sub.n(τ), obtained from a measurement according to one of embodiments 2A-2D. If the deviations in M.sub.n(t) can be described as a multiplication with a signal of moderate bandwidth, the two compressed phasors will deviate mostly close to zero delay. In the case of the embodiments shown in FIGS. 2C and 2D, the delay range of the deviation in the compressed interrogation phasor that can be measured may be limited by the delay difference between the monitor signal path and the first sensor reflector (the start of the sensor fiber or the first discrete reflector).
(22) In one embodiment of the invention, the pulse compression filter is updated to C(τ)={circumflex over (M)}*.sub.n(−τ), where {circumflex over (M)}.sub.n(τ) is the estimate of the interrogation phasor. In another embodiment of the invention a noise suppression filter is calculated and dynamically updated such that the convolution between compressed interrogation phasor and the noise suppression filter approximates the compressed reference phasor in a least square sense. The convolution of the original pulse compression filter and the noise suppression filter combined into an improved pulse compression filter, or the noise suppression filter can be applied to the sampled signal either before or after the primary pulse compression filter.
(23) FIG. 5 shows one embodiment of the method 500 that can be used to reduce the sensitivity to laser noise on the interrogation signal and the local oscillator signal. In a first step 502, the fluctuations in the local oscillator signal is measured according to the embodiment in FIG. 2A. The embodiment in FIG. 2A provides a measured laser noise that is used to correct for local oscillator induced noise on the receiver output signal, as described by the term involving ϵ*.sub.LO in Eq. (7) is applied in step 504. Correction for local oscillator phase noise can be implemented in signal processing by rotating the complex receiver output signals with the measured laser phase errors. The required sign of the rotation will depend on the hardware configuration and sign convention used. One efficient implementation of such a rotation is a coordinate rotation processor (CORDIC). Additional intensity correction can be implemented through multiplication with a signal that is proportional to the inverse square root of the detected laser intensity signal. The phase and intensity measurements can also be combined into the variable ϵ.sub.LO(t), and the noise on the local oscillator can for instance be suppressed by multiplying the receiver output signal with 1−ϵ.sub.LO(t). In step 508, ΔM.sub.n(τ) is computed as ϵ.sub.n(τ)R(τ), where ϵ.sub.n(τ) comprises the laser phase and/or amplitude noise. This deviation is used to update the pulse compression filter such that C(τ)={circumflex over (M)}*.sub.n(−τ)=R*(−τ)+ΔM*(−τ), and/or to calculate a noise suppression filter T(τ) based on the deviation of A(τ)−{circumflex over (M)}.sub.n(τ)*C(τ) as described in eq. (13). This approach may be useful in cases where laser noise is a dominating contributor to fluctuations in the interrogation phasor. In step 510, the noise suppression filter and the pulse compression filter are combined into a modified pulse compression filter, which is applied to receiver output signal in step 506. However, they may also be applied individually on the receiver output signal.
(24) FIG. 6 shows an embodiment of the method 600 which may further extend the embodiments of FIG. 3 or FIG. 5. The method extracts and applies a noise suppression filter based on a measurement of the interrogation phasor according to the system of FIG. 2C or 2D. In step 602 a delay range of the output from a primary pulse compression filter 306 is extracted which is limited to not overlap with responses from strong sensor reflectors. The extracted signal is compared to the target A(τ) to obtain the noise suppression filter in step 604 with the methods described above. The noise suppression filter is finally applied in step 606 to the output from the pulse compression filter. This process is repeated for every repetition period. When a swept interrogation pulse is used, the noise suppression filter will correct for multiplicative error contributions to M.sub.n(t) with frequencies up to SWR.Math.T.sub.d, where T.sub.d is the delay range that is included in the computation of the compensation filter.
(25) FIG. 7 describes yet another embodiment of the method 700 according to the invention. A preliminary estimate of the detuning can be calculated as the spatial derivative of the phase of the output from the pulse compression filter, with or without involving the noise suppression techniques described in the previous embodiments. The non-linearity of the detuning estimate can be suppressed by applying Eq. (14) to the pulse compression filter output. This is implemented in this embodiment by first computing a detuning compensation filter 706 based on the laser frequency detuning 704 and the compressed reference phasor A(τ). The output from the detuning compensation filter 708 is frequency shifted by −Δν. The phase of the frequency shifted signal 710 should be independent of the laser frequency detuning. In some embodiments the first detuning correction is computed based on detuning estimates calculated during previous repetition periods. In some embodiments it may be beneficial to repeat this processing where the detuning output of the first stage is used as the preliminary input to the second stage.
