DSP signal processing for open loop fiber optic sensors

Details DSP signal processing for open loop fiber optic sensors DSP signal processing for open loop fiber optic sensors

2000-07-13

2002-08-06

Turner, Samuel A. (Department: 2877)

Optics: measuring and testing

By light interference

Rotation rate

C356S483000

Reexamination Certificate

active

06429939

ABSTRACT:

FIELD OF THE INVENTION
The invention is directed to fiber optic sensors, and more particularly to digitally processing signals derived from fiber optic sensors.
BACKGROUND OF THE INVENTION
An open loop fiber optic sensor, such as an open loop fiber optic gyro (FOG), has the advantage of simplicity and low cost compared to a closed loop configuration. On the other hand, closed loop FOGs have the advantage of excellent bias stability, linearity and scale factor stability, although some of these characteristics require thermal modeling on an individual unit basis. The open loop gyroscope has a first order sinusoidal response to rotation, and the scale factor is dependent on both the optical intensity at the detector and the modulation depth. These considerations have limited the performance of most open loop gyroscopes.
The theory of open loop gyroscopes is known in the art. The direction of rotation can be determined and the sensitivity optimized by applying a phase modulation at a frequency f
1
to the light propagating in the fiber optic sensing coil, for example, through a piezoelectric element. Alternatively, with some sensor configurations, such as a reflective current sensor, a birefringent modulator could be employed. The Sagnac interferometer converts this modulation into a detected output signal represented by a series of Bessel functions. The amplitudes of the odd harmonics at the modulation frequency and its harmonics are proportional to the sine of the rotation rate, while the amplitudes of the even harmonics are proportional to the cosine of the rotation rate. All of the information required to determine the rotation rate and to linearize and stabilize the scale factor can be extracted from the fundamental signal and the second and fourth harmonics.
Conventional open-loop FOG systems process the signal in analog form, which represents an approximation of an analytic approach. A FOG can be operated with a low Sagnac scale factor by using a short length coil and restricting the maximum input rate. The operating regime is thus selected to be near the zero of the sine function at the fundamental frequency. For a small modulation depth, the sine function can be approximated by a linear function. The amplitude of the second harmonic is then at the peak of a cosine function and consequently varies very little with the rate. The peak value of the second harmonic can be used as a measure of the detected signal intensity to control and maintain the light source power over time and operating temperature. The self-oscillating PZT can be operated at approximately the same modulation depth over temperature. With these design criteria, analog signal processing can provide a rate gyro with acceptable rate gyro performance. However, further improvements of the analog processing are likely to be more complex, will require additional alignment steps during manufacture, and reduce the reliability. Even at present, the large number of discrete components makes FOGs with analog signal processing marginal in high reliability applications. In addition, the design constraints imposed by the required linearity and sensitivity are often difficult to attain jointly.
Some of the problems described above have been addressed in the prior art by employing digital signal processing (DSP) concepts in at least a part of the electronic circuitry. For example, individual synchronous detectors were driven with the fundamental or a desired harmonic, followed by A/D conversion. This approach requires the generation of the fundamental by dividing a clock frequency such that all of the required harmonics are generated individually. In another approach, the photo detector signal was converted from analog to digital (A/D), wherein the A/D sample rate does not have an integral relationship with the modulation frequency, so that the signal processing approach relies on a FFT technique to isolate the required frequencies. Further, since the signal may not be centered in each frequency bin, the amplitude of the Fourier transform may be affected and/or distorted by the particular windowing function used. According to yet another prior art approach, a downsampling technique was used, wherein the A/D sample rate clock is derived from the same oscillator as the PZT modulation frequency, but with a different divide ratio. While this substantially reduces the throughput demands on the A/D converter and the signal processing electronics, undersampling folds the gyro broadband noise into the processing bandwidth and increases the angle random walk (ARW).
It would therefore be desirable to provide a digital signal processing system and method for a fiber optic sensor that simplifies the circuitry, improves the linearity, the scale factor and the bias stability, and provides more flexible design rules, without introducing artifacts from the digital sampling process.
SUMMARY OF THE INVENTION
The invention is directed to a system and method for digitally processing the fiber optic sensor output signal using a limited number of frequencies where the frequency used to drive the birefringent modulator or phase modulator is an integer multiple of the A/D sampling frequency.
According to one aspect of the invention, a digital signal processing system for a fiber optic sensor used to measure a physical quantity, such as a rotation rate or a magnetic field, includes a digital signal processor (DSP); an oscillator or frequency syntheziser which may be controlled by the DSP or other means known in the art, producing a sample rate clock frequency which is an integer multiple of a modulator drive frequency of the fiber optic sensor; and a memory in communication with the DSP and storing frequency coefficients which are precomputed at the modulator drive frequency and at the second and fourth harmonic of the modulator drive frequency. The system further includes a signal converter for sampling at the clock frequency an output signal produced by the fiber optic sensor. The DSP multiplies the sampled output signal with respective ones of the frequency coefficients and forms in-phase and, if necessary, also quadrature components at the respective frequencies to compute the physical quantity. Alternatively, the phase of the frequency coefficients can be adjusted so that the reference signal is in phase with the gyro signal. This arrangement has the advantage of obviating the need for detecting the quadrature components, thereby eliminating the noise contribution from the quadrature components and improving the overall sensitivity.
The fiber optic sensor includes a fiber sensing coil; a light source supplying optical radiation to the fiber sensing coil; a modulator disposed between the fiber sensing coil and the light source, wherein the modulator is driven at a modulator frequency and modulates the optical radiation supplied to the sensing coil. A light detector detects return optical radiation returned from the fiber sensing coil and provides an output signal corresponding to a physical quantity detected by the fiber sensing coil. A light source driver controlled by the DSP supplies a current to the light source so as to maintain at least one of a constant intensity of the optical radiation produced by the light source and an optical power of the return optical radiation detected by the light detector.
According to another aspect of the invention, in a method of digitally processing a signal of a fiber optic sensor, a clock frequency is produced which is an integer multiple of a modulator drive frequency. Sine and cosine coefficients are computed at the modulator drive frequency and at a plurality of integer multiples of the modulator drive frequency for a plurality of discrete signal sampling times. The signal of the fiber optic sensor is acquired at the discrete signal sampling times over an integration period and the acquired signal at the discrete signal sampling times are multiplied with the corresponding sine and cosine coefficients obtained at the same discrete signal sampling times to define a sensor state vector. Characteristic o

Affiliated with

Also associated with

Rating

DSP signal processing for open loop fiber optic sensors does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with DSP signal processing for open loop fiber optic sensors, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and DSP signal processing for open loop fiber optic sensors will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2889617