 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates generally to distance measuring devices and
techniques, and, more particularly, relates to a two-color synthetic
Michelson interferometer for optically measuring distance over relatively
long paths with resolution to a small fraction of an optical wavelength
without the range ambiguity inherent in conventional interferometry.
The origins of employing an optical wavelength, presumed to be a constant,
to measure absolute distance stems from the early classical experiments by
Michelson and Benoit to measure the international meter in terms of the
wavelength of the red line of cadmium. Laser interferometry enhances the
precision with which interferometric measurement can be made, but because
of its superior coherence length makes possible extended range
interferometry, and through precision frequency stabilization provides the
basis of an absolute wavelength standard.
The ambiguity of conventional interferometers may be greatly reduced
through the use of multiple wavelengths. Measuring absolute distance
interferometrically requires that the fringe order number in the
interferometer be identified. One would like to employ a multiwavelength
source with an ambiguity length longer than the greatest distance to be
measured; however, for most practical applications this is unnecessary.
Interferometer ambiguity distances large enough to be resolved by some
form of a priori measurement are considered acceptable. One technique to
extend the interferometer ambiguity distances employs a number of
well-characterized, suitably-spaced wavelengths produced by a CO.sub.2
laser source operating in the 10.4 .mu.m wavelength band. The differences
in a selected set of these wavelengths, and the differences in the
differences . . . ad infinitum . . . were used to generate a hierarchy of
wavelengths whereby, using fractional fringe measurement techniques, and a
simple algorithm, distance employing any wavelength in the hierarchy could
be established with sufficient accuracy to identify the next lower
wavelength order number. By working downward through the wavelength
hierarchy (from the longest wavelengths to the optical fringes), distance
is ultimately established in terms of a well known optical wavelength, the
unit of measure.
Analysis shows that the ideal wavelength hierarchy consists of a
geometrical progression of wavelengths of sufficient density so that the
fractional fringe measurement resolution of any wavelength in the
hierarchy could reliably measure distance to a small fraction of the next
lower wavelength. Practically, however, the availability of appropriately
spaced wavelengths occurs as an act of nature. The use of isotopes can
modify the available wavelengths somewhat, but this would have only a
small effect on the desired progression of wavelengths.
One technique of using multiple wavelengths is disclosed in U.S. Pat. No.
4,355,899, entitled "Interferometric Distance Measurement Method" and a
device for using this technique is further disclosed in U.S. Pat. No.
4,355,900, entitled "Self-Calibrating Interferometer". This particular
device uses two Michelson interferometers to measure an unknown distance.
Another approach uses heterodyne photodetection to provide a basis for
quantum-noise-limited operation that gives a high signal-to-noise ratios
with a large signal with a small target return power. It further provides
for a large variation in signal level and permits the phases (fractional
fringes) of the optical signals to be measured electronically at a
convenient frequency. Optical offsets necessary for heterodyne operation
are provided by several Bragg cells that also serve to spectrally isolate
the laser from the target returns which upset the laser's stability.
One optical setup employing heterodyning uses a CO.sub.2 laser, emitting
power from both ends, where one beam is used to control and stabilize the
laser and the other beam split into two beams, one of which is employed in
a synthetic interferometer and the other as a local oscillator (LO) beam.
Because heterodyne photodetection is used, the local oscillator beam and
interferometer beams are frequency offset with Bragg cells and spatially
separated. Ordinarily a single frequency translation would suffice but a
single Bragg cell generates, in addition to the desired
frequency-translated component, a second contaminating component due to a
small backward (reflected) acoustical wave. This component has in the past
been large enough to impair the phase measurement process. Employing a
second Bragg cell at a second frequency provides a basis for isolating
this component but increases the complexity. Two detectors are used to
detect phases in the two beams which are required because of spatial
separation and no frequency offsets.
One element of the prior Absolute Distance Sensor is the "two-color laser"
(T-C laser), which is capable of stabilizing and operating simultaneously
on any of four sets of two-color pairs (for a total of five different
rotational-vibrational lines in the CO.sub.2 -10.4 .mu.m band), and of
rapidly switching through the various color pairs by means of a
piezoelectric mirror drive and control subsystem. All of the basic
features of the T-C laser (its stability, switching capability, line
pairing sequence, states-of-operation, switching speed, and derivable
wavelength hierarchy) are fundamental to the operation of the Absolute
Distance Sensor system.
Progressing through the optical train of the prior Absolute Distance
Sensor, separate target/reference sensing beams are established at a
beamsplitter and pass in the vicinity of an optical switch (chopper)
employed to alternately range to the target and reference legs of the
interferometer. Since it is the difference between these measurements that
is of interest, phase noise in the electronics and all optics upstream
from the beamsplitter is common to both measurements and may be largely
cancelled. By making rapid phase measurements and by switching between
target and reference at a high rate (240 Hz is used here) significant
common mode noise cancellation benefits may be achieved. The
target/reference beams leaving the interferometer are made colinear at the
beamsplitter, and made congruent with the LO beam at a second
beamsplitter, and progress to a grating where the R- and P-lines are
separated and directed to separate detectors. The heterodyne photodetected
signals are then digitally processed for phase. A 36 MHz clock is started
and stopped by axis crossing detectors using target and reference returns
from the detected R- and P-carriers as well as an electronic reference
signal derived from the Bragg cell drivers, all of which are heterodyned
to a convenient working frequency (in this case, 10 kHz). Statistically,
the digital error of this scheme is approximately 1:2.8.times.10.sup.3 for
each phase measurement. Statistical improvement is achieved by averaging
over 10 cycles per phase measurement. Relative times (reference signal
period and target return delay times) are directly available in a form
convenient for computer input without further processing.
Although the prior Absolute Distance Sensor proved the concept of accurate
distance measurement that sensor required an elaborate signal processing
scheme and a complex optical system suitable only in the laboratory.
SUMMARY OF THE INVENTION
The instant invention overcomes the problems encountered in the past by
providing an improved Absolute Distance Sensor utilizing spectral
isolation of the R and P beams, one photo detector and one Bragg cell to
result in a more practical Absolute Distance Sensor configuration.
A multi-state two-color CO.sub.2 laser in response to an electronic laser
controller outputs a carrier beam having a selected pair of lines from the
CO.sub.2 laser spectrum. This carrier beam is input to a multi-frequency
Bragg cell driven by several selected frequencies. The Bragg cell outputs
several different R and P line beams that are frequency offset from the
carrier. One pair of R and P lines form a local oscillator beam and
another pair of R and P lines form an interferometer beam.
The local oscillator beam is input into a detector for measuring phase. The
interferometer beam is input into a synthetic Michelson interferometer
that has a beam switch that causes the interferometer beam to be returned
from a target and reference reflector alternatively. The return target and
reference beams are made congruent with the local oscillator beam and also
input to the phase detector.
In the improved Absolute Distance Sensor, the synthetic Michelson
interferometer is illuminated by each R and P line pair in rapid
succession and in a given order. The phase at a convenient heterodyne
frequency is measured alternately for each of two beams directed at
retroreflectors, one designated "target", the other "reference". The
phases are then subtracted, yielding a phase differential or "fractional
fringe" indicative of the optical path difference between target and
reference lengths. This is done simultaneously for the R and P lines of
each pair. Four basic wavelength pairs generated by the laser may be
combined to produce the long synthetic wavelengths. These, in turn,
provide a means to determine the exact number of 10 .mu.m wavelengths in
the optical path difference. This is accomplished by measuring the optical
path difference for each synthetic wavelength (obtain number of
wavelengths plus the fractional fringe, beginning with the longest) with
sufficient accuracy that a handover to the next shorter wavelength can be
accurately done. The process is repeated until the 10 .mu.m wavelength
number and fractional fringe are obtained.
It is therefore an object of this invention to provide an improved absolute
distance sensor using spectral isolation to reduce cross talk
contamination;
It is another object of this invention to significantly reduce the number
of optical components required to make distance measurement;
It is a further object of this invention to provide an improved Absolute
Distance Sensor that is capable of practical use.
These and many other objects and advantages of the present invention will
be readily apparent to one skilled in the pertinent art from the following
detailed description of a preferred embodiment of the invention and
related drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a diagrammatic representation of the improved absolute distance
sensor of this invention;
FIG. 2 is a block diagram of the electronics used in the improved absolute
distance sensor of this invention;
FIG. 3 illustrates oscillators for outputting multi-drive frequencies to
the Bragg cell;
FIG. 4 illustrates the laser CO.sub.2 gain curve in the 10.8 .mu.m band;
FIG. 5 illustrates wavelength heirarchy used in the two color laser;
FIG. 6 illustrates the four-state two-color laser frequency/wavelength
hierarchy pyramid; and
FIG. 7 illustrates switching sequence used in the four-state two-color
laser.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to FIG. 1 which illustrates the optical arrangement of an
improved absolute distance sensor 10, a four-state two-color CO.sub.2
laser 12 outputs a R/P carrier beam 14. Four-state two-color CO.sub.2
laser 10 is more fully described in U.S. Pat. No. 4,513,422 entitled
"CO.sub.2 Laser Stabilization and Switching" issued to the same assignee
and is hereby incorporated by reference. A reflector 16 causes beam 14 to
enter a multi-frequency Bragg cell 22. Bragg cell 22 outputs multiple
beams, not shown, to a beam corrector 24. Beam corrector 24 outputs, in
particular, an interferometer beam 26 having offset R and P line beams of
about 58 MHz which are collinear and congruent, and local oscillator R and
P beams 30 and 32, respectively, having a frequency of 17 MHz and 48 MHz,
respectively. These frequency offsets are with respect to the frequency of
R and P carrier beam 14.
R beam 30 is incident on a first beam splitter 34 that functions to align
beam 30 and to reduce the intensity of LO R beam 30. A first beam stop 38
absorbs the undesired portion of R beam 30. Similarly, P beam 32 is
incident on a second beam splitter 40 that functions to align P beam 32
with R beam 30. The undesired part of P beam 32 is absorbed by a second
beam stop 42. Then a congruent and collinear R and P beams 30 and 32 are
incident on a third beamsplitter 46. The undesired parts of R and P beams
30 and 32 are absorbed by a third beam stop 48. Finally reduced, collinear
and congruent R and P beams 30 and 32, shown as one line, are incident on
a single detector 54.
Returning to beam corrector 24, interferometer beam 26 enters a synthetic
Michelson interferometer 56. Interferometer beam 26 reflects off a mirror
58 and is incident on a fourth beam splitter 57. Interferometer beam 26 as
a result of fourth beam splitter 57 divides into a reference beam 62 and a
target beam 60. Target beam 60 reflects off a third mirror 64. Then both
target beam 60 and reference beam 62 pass through a conventional beam
chopper 66 that allows only a return from a target retroreflector 68 or a
reference retroreflector 70 to be analyzed. Target retroreflector 68 is
mounted on a slide 72. A reference interferometer 74 is used for checkout,
but is not required to obtain the distance of concern.
T-C laser 12 is capable of stabilizing and operating simultaneously on any
of four sets of two-color pairs (for a total of five different
rotational-vibrational lines in the CO.sub.2 10.4 .mu.m band), and of
rapidly switching through the various color pairs by means of a
piezoelectric mirror drive and control subsystem, not shown.
The basic concept behind T-C laser 12 stabilization technique is related to
the fact that a linear power exchange exists between the R and P lines as
a function of laser frequency in the region of two-color operation.
Equalizing the R and P line powers thus provides a laser frequency
discriminant which is based on the CO.sub.2 molecule and which, in
principle, should be absolute. Laser stabilization test data obtained by
beating two stabilized nearly identical T-C lasers 12 together indicates
that the residual noise of one T-C laser is approximately .+-.35 kHz, or 1
part in 0.8.times.10.sup.9 per 15 minutes. Short term stability is 1 part
in 7.times.10.sup.9 per msec.
The signature of T-C laser 12 (dominant resonant frequency as a function of
laser cavity length) is chosen to give the desired sequence of R and P
lines in the 10.4 .mu.m vibrational band. A piezoelectric mirror mount in
conjunction with the stabilization circuit adjusts the cavity length to
correspond to a chosen segment of the signature. Adjusting the cavity
length over a portion of a specific 5 .mu.m range allows one to obtain the
signature segment shown in FIG. 7. Also shown here are the four
color-pairs, the off-line-center stabilized operation points and the
four-state switching sequence.
TABLE I
__________________________________________________________________________
##STR1##
##STR2##
##STR3##
##STR4##
##STR5##
__________________________________________________________________________
Table 1 displays the hierarchy of differential and synthetic frequency
lines obtained from the eight optical frequencies shown on the first
(bottom most) level. The second level consists of differential (beat)
frequencies of the various two-color line pairs (simultaneous data). The
third, fourth, and fifth levels are "synthetic" frequencies since they are
derived using more than one color pair (non-simultaneous data) and hence
are not physically observable frequencies. The sixth and seventh levels
are also "synthetic" since they rely upon data obtained from either side
of line center (two different color pairs). The wavelength values are
calculated from the frequency values using the speed of light.
c=2.99792(00).times.10.sup.8 m sec, and assuming a refractive index, n=1.
Signal processing of the R and P lines is accomplished with the electronics
shown in FIGS. 2 and 3. As shown in FIG. 2, the output from T-C laser 12
is directed to Bragg cell 22 used to generate local oscillator beams 30
and 32, R and P beams respectively, and interferometer beam 26.
In order to drive Bragg cell 22, predetermined frequencies generated by a
LO battery 82 and a Target battery 84, shown in FIG. 3, are applied as
determined by a computer 80. Two drive frequencies, F.sub.PLO and
F.sub.RLO separate the R and P beams 30 and 32. The frequency offsets are
relative to T-C laser 12 R and P carrier beam 14. Computer 80 is
programmed to activate the appropriate oscillators in consonance with a
given T-C laser 12 state. Compensation for the effects of the frequency
variations on phase measurements and heterodyning take place in the signal
conditioning circuitry, not shown, in detector 54, prior to R and P phase
meters 76 and 78.
R and P interferometer beam 26 is generated in a similar manner with
oscillators. The large electronic spectral separation employed with the LO
R and P beams 30 and 32 is not necessary with interferometer beam 26.
Interferometer beam 26 having closing spaced R and P lines exiting cell 22
are made collinear by appropriate choice of drive frequencies F.sub.PT and
F.sub.RT, and are kept colinear by appropriate adjustment to cell 22 drive
frequencies derived from a second battery of oscillators 84. The resulting
spectrum for both the LO and interferometer beams 30, 32, and 26 are shown
in FIG. 7. Beam corrector 24 makes necessary adjusts in beam direction as
necessary and is of conventional design.
Referring to FIG. 7, R and P carriers are superimposed and the relative
offset frequencies of the subcarriers, either R or P, are shown to the
same scale. The large separation between R and P LO beams 30 and 32
provides the spectral isolation needed to simultaneously process the R and
P target and reference return signals (i.e., measure their phases) without
electronic crosstalk contamination. A small frequency separation (.about.2
MH.sub.z) between the R and P interferometer beam 26 is present and is
needed to compensate cell 22 for R and P line dispersion (i.e., provide
the beam overlap shown in FIG. 6 and also ensure that photo-mixing between
the main subcarriers and the contaminating components due to a backward
acoustical wave in Bragg cell 22 (-F.sub.RLO, -F.sub.PLO, F.sub.PT and
F.sub.RT) as well as the forward scattered components, R' and P', in the
LO and target directions will be outside the main 106 MHz and 75 MHz data
channels. The residual in-band photomixing that does occur in the data
channels, however, will be second order and, hence, of no consequence.
Measuring absolute distance interferometrically requires that the fringe
order number in the interferometer be identified. One would like to employ
a multiwavelength source with an ambiguity length longer than the greatest
distance to be measured; however, for most practical applications this is
unnecessary. Interferometer ambiguity distances large enough to be
resolved by some form of a priori measurement are considerd acceptable.
The basic approach used is to extend the interferometer ambiguity
distances employed to a number of well-characterized, suitably-spaced
wavelengths produced by a CO.sub.2 laser source operating in the 10.4
.mu.m wavelength band, FIG. 5. The differences in a selected set of these
wavelengths, and the differences in the differences . . . ad infinitum . .
. are used to generate a hierarchy of wavelengths, Table 1, whereby, using
fractional fringe measurement techniques, and a simple algorithm, distance
employing any wavelength in the hierarchy can be established with
sufficient accuracy to identify the next lower wavelength order number. By
working downward through the wavelength hierarchy (from the longest
wavelengths to the optical fringes), distance is ultimately established in
terms of a well known optical wavelength, the unit of measure.
Analysis shows that the ideal wavelength hierarchy would consist of a
geometrical progression of wavelengths of sufficient density so that the
fractional fringe measurement resolution of any wavelength in the
hierarchy could reliably measure distance to a small fraction of the next
lower wavelength. Practically, however, the availability of appropriately
spaced wavelengths occurs as an act of nature. The use of isotopes can
modify the available wavelengths somewhat, but this would have only a
small effect on the desired progression of wavelengths.
FIG. 5 exhibits a wavelength display obtained by selecting a specific path
through the values shown in Table 1. A logarithmic scale was employed so
that an approximation to a geometric wavelength progression (hierarchy
wavelength versus order number) would appear with the wavelength values
distributed along a straight line. The dashed line, a visual best fit,
shows that the wavelength hierarchy available from the CO.sub.2 laser does
indeed come remarkably close to a geometrical progression. Also
illustrated is the measurement accuracy (shown as the right-hand limit in
the overlap between the various wavelengths) required of each wavelength
to reliably capture the next lower wavelength in the hierarchy. Fractional
fringe measurement techniques have been developed that permit the required
measurement accuracy to be realized.
A total of sixteen different paths can be traced through the hierarchy
pyramid in Table 1. Four of these paths can be used in improved Absolute
Distance Sensor 10, both as a diagnostic for the system and as a means of
obtaining information on all eight laser line frequencies. Table II
displays these four paths. Eventually, only Path 2 will be used, thus
reducing the computer program complexity and increasing computational
speed.
TABLE II
______________________________________
Wavelength Hierarchy.
Level Path #1 Path #2 Path #3 Path #4
______________________________________
7 25 m 25 m 19 m 19 m
6 5.7 m 5.7 m 5.7 m 5.7 m
5 397 mm 397 mm 397 mm 397 mm
4 21.1 mm 21.1 mm 22.3 mm 22.3 mm
3 5.53 mm 7.49 mm 7.40 mm 5.60 mm
2 310 .mu.m 329 .mu.m 344 .mu.m 366 .mu.m
1 R.sub.2 (18)
R.sub.1 (18)
R.sub.2 (16)
R.sub.1 (16)
P (22) P.sub.2 (20)
P.sub.1 (20)
P (18)
______________________________________
Referring to FIG. 2, digital phase meters 76 and 78 determine phase
information from the heterodyne photodetected signals of the R and P lines
of LO R and P line beams 30 and 32, respectively, and R and P reduced
target and reference beams 60 and 62, respectively. This phase information
is then processed to yield range measurements using a dedicated Cromemco
2D microcomputer operating at 4 MHz. The microcomputer controls laser
switching through a laser phase data collection, display of results, and
system calibration. However, the rapid conversion of raw phase data into
range with values good to 2 parts in 10.sup.8 per meter is the primary
function of the microcomputer.
Phase information from each of the four color pairs (for both the target
and reference ranges) is determined by phase meters 76 and 78 and stored
in computer 80 as a fractional fringe number. From this phase information,
fractional fringe values are then calculated for each successively higher
level of the hierarchy pyramid. Having obtained a complete hierarchy of
fractional fringes, range determination may be viewed as a process of
successive approximation. The range is first determined at the longest
wavelength (hierarchy Level 7) by determining the integral number of the
whole fringes using fractional fringe data and an initial range estimate.
This new range value is then used, along with the fractional fringe data
from the next lower level, to determine a second range value. This process
is repeated down through the hierachy until a final range value is
determined at the 10 .mu.m wavelenghts.
The final range value is good to 0.025 .mu.m as long as the measurement
accuracy (phase resolution) and the laser frequency stability are
sufficient to correctly determine the integral number of whole fringes for
each wavelength down through the hierarchy. Each wavelength transition in
the hierarchy places different numerical criteria for measurement accuracy
and laser stability on the system. These criteria were derived by assuming
a total acceptable error of .lambda./10 for each wavelength,
.lambda..sub.n, in the hierarchy. The phase meters provide a measurement
accuracy of .+-.1.4.times.10.sup.-3 fringe (.+-.0.05 degrees) which is
double the most stringent measurement accuracy requirement at the
.lambda..sub.2 level in the hierarchy transition (.+-.3.1.times.10.sup.3
fringe). With range L set at 5 meters, an RMS laser stability of .+-.140
kHz is required. This, for contrast, is to be compared with the measured
stability of the laboratory T-C laser 12, which is .+-.35 kHz.
The analysis up to this point has neglected any changes in optical length
which occur during the collection of phase data. Small vibratory
excursions do occur which result in inconsistent phase measurements. Phase
measurements with a spread of greater than 3% may prevent accurate
determination of the range. The role of simultaneous R- and P-line phase
measurements in the reduction of system sensitivity to path length change
has been analyzed. Results indicate a significant increase in synthetic
wavelength phase measurement accuracy using simultaneous measurements.
This increase in accuracy is proportional to the ratio of wavelengths
between Level 1 and Level 2 of the hierarchy (approximately 35:1). The
simultaneous measurement of color pairs provides improved Absolute
Distance Sensor 10 with a crucial insensitivity to path length changes due
to target vibration and atmospheric turbulence.
To experimentally obtain a measure of accuracy as well as determine its
stability and general performance properties, comparison measurements were
made using an HP Interferometer (Model 5525) as reference 74. In an effort
to establish coincident measurement geometries the ADS and HP beams were
accurately colocated, a common retroreflector target 70 was employed, and
the target-to-reference distances were set nearly equal for the two
systems.
Clearly, many modifications and variations of the present invention are
possible in light of the above teachings and it is therefore understood,
that within the inventive scope of the inventive concept, the invention
may be practiced otherwise than specifically claimed.
* * * * *
|
|
 |
|
 |
Description |
|
