 |
Claims  |
|
|
We claim:
1. An ultrasonic Doppler system for measuring velocity of fluids moving
within internal patient structures, comprising:
means for transmitting an acoustic beam of addressing pulses of ultrasonic
frequency acoustic waves into said patient structure, and receiving the
reflected acoustic echo energy from both stationary and moving acoustic
scatterers within the path of said beam and demodulating same into
corresponding electrical signals containing both Doppler and stationary
echo information;
means for sampling said echo signals into a plurality of channels, each
corresponding to a portion of the time between said addressing pulses;
means for determining from the digital form of said Doppler echo
information, the velocity of said fluids, said means including a zero
cross detector and a flow velocity sign detector; and
filter canceler means receiving signals from said means for sampling and
supplying an output to said velocity determining means, said filter in
turn processing each of said channels, said filter employing a recursive
loop including a digital memory for comparing the echo signals from each
addressing pulse with its immediate echo signal predecessor, a subtractor
to subtract any unchanged non-Doppler components for each channel, and an
analog to digital converter for rapidly digitizing the remaining changing
components of said echo signal, said means including comparator means in
parallel with said analog to digital converter for identifying the highest
active digital bit between the digital dynamic range of bits in the data
path of the incoming echo signals, and the dynamic range of the analog to
digital converter, said comparator means providing said highest active bit
to be added along with the output of said analog to digital converter into
said memory, so that upon comparison of the output of said memory with the
incoming echo signal for each given channel and subtraction therefrom, and
subsequent iterations thereof, the number of iterations is decreased, and
the response time is correspondingly improved for large stationary
signals.
2. An ultrasonic Doppler system as in claim 1, in which said comparator
means includes two sets of a like plurality of comparison circuits, one
set being for the positive signals, the other set being for the negative
signals, and logic means monitoring the output of said comparison circuits
and identifying the one of said comparison circuits having the largest
absolute value threshold, the corresponding bit being said highest active
bit.
3. An ultrasonic Doppler system as in claim 2, which said digital dynamic
range is of N bits capacity, said analog to digital converter is of n bits
capacity, and in which said plurality of comparators in each of said sets
is given by the expression:
N-(n+1)
4. An ultrasonic Doppler system as in claim 3 which further includes a
digital to analog converter between the output of said digital memory and
an input of said subtractor, said digital to analog converter being of N
bits capacity.
5. An ultrasonic Doppler system as in claim 3, in which said comparator
means is in parallel with said analog to digital converter, and in which
said comparator means has a nonzero output only when the dynamic range of
said analog to digital converter is exceeded.
6. An ultrasonic Doppler system as in claim 5, in which said memory for
each channel stores data due both to the output of said analog to digital
converter, and the output of said comparator, and in which said subtractor
is an analog subtractor having an input receiving an incoming signal due
to a new addressing pulse, and another input receiving said stored data
resulting from the previous addressing pulse, said subtractor output up
with supplying both said comparator and said analog to digital converter,
said stored data comprising both the output from said comparator and from
said analog to digital converter being thereby subtracted from said
incoming signal, whereby further iterations over said loop are
substantially reduced.
7. A sampled ultrasonic Doppler system for measuring the velocity of fluids
moving within internal body structures comprising:
means for transmitting an acoustic beam of addressing pulses of ultrasonic
frequency acoustic waves having a regular phase repetition rate into said
body structure and for receiving the acoustic echo energy from both
stationary and moving acoustic scatterers within the path of said beam,
and converting same into corresponding electrical signals;
means for demodulating said echo signals into quadrature echo signals, said
quadrature signals containing both Doppler and stationary echo
information;
means for sampling said echo signals into a plurality of equal channels,
each corresponding to a portion of the time between said addressing
pulses;
filter means accepting and digitizing said echo signals for each channel
and cancelling therefrom any stationary non-Doppler information;
means for determining, from the remaining signal comprising Doppler echo
information in digital form, the zero crossings in one direction of said
Doppler information-bearing signal to obtain the Doppler frequency for
each channel;
means for determining from said Doppler frequency information, and
information as to whether the sign of the Doppler shift is positive or
negative, the Doppler velocity for each channel, said means including
means accumulating digital counts representative of said zero crossings;
and
means for determining the sign of said Doppler shift, and supplying said
information to said means for determining velocity, including
means for comparing the quadrature components of said quadrature echo
signals to determine the instantaneous sign of said Doppler frequency
shift for each channel, and
means for comparing said accumulating counts against a predetermined
frequency, and upon said counts exceeding the value corresponding to said
predetermined frequency, substituting the sign which was determined just
prior to exceeding said value, in preference to said instantaneous sign,
whereby the accuracy of said Doppler velocity-determining means is
improved.
8. A sample ultrasonic Doppler system as in claim 7, in which said
predetermined frequency is the Nyquist criterion of one-quarter the
interrogating pulse repetition rate for quadrature multiplexed sampled
systems.
9. A sample ultrasonic Doppler system as in claim 7, in which said means
for determining the Doppler shift sign includes
a first memory storing a representation of the current sign of the Doppler
velocity,
a second memory,
logic means to compare the absolute value of said accumulating counts with
said predetermined frequency, and causing said second memory to store an
indication if said frequency is exceeded in any channel,
switch means receiving both signal representing said instantaneous sign,
and a signal from said second memory indication to substitute said stored
sign for said instantaneous sign as an input into said means for
determining the Doppler velocity.
10. An ultrasonic Doppler system for measuring velocity of fluids moving
within internal patient structures comprising:
means for transmitting an acoustic beam of addressing pulses of ultrasonic
frequency acoustic waves having a regular pulse repetition rate into said
patient structure, and receiving the reflected acoustic echo energy from
both stationary and moving acoustic scatterers within the path of said
beam, and demodulating same into quadrature electrical signals containing
both Doppler and stationary echo information;
means for sampling said echo signals into a plurality of equal channels,
each corresponding to a portion of the time between said addressing
pulses;
means for determining from the digital form of said Doppler echo
information, the velocity of said fluids;
filter canceller means receiving signals from said means for sampling, and
supplying an output to said velocity determining means, said filter in
turn processing each of said channels, said filter employing a recursive
loop including a digital memory for comparing the echo signals from each
addressing pulse with its intermediate echo signal precedessor, a
subtractor to subtract any unchanged non-Doppler components for each
channel, and an analog-to-digital converter for rapidly digitizing the
remaining changing components of said echo signal, said means including
comparator means in parallel with said analog-to-digital converter for
identifying the highest active digital bit between the digital dynamic
range of bids in the data path of the incoming echo signals, and the
dynamic range of the analog-to-digital converter, said comparator means
providing said highest active bit to be added along with the output of
said analog-to-digital converter into said memory, so that upon comparison
of the output of said memory with the incoming echo signal for each given
channel and subtraction therefrom, and subsequent iterations thereof, the
number of iterations is decreased;
means included in said velocity determining means for determining from said
remaining changing components of said signal in digital form, the zero
crossing in one direction to obtain the Doppler frequency for each
channel;
means included in said velocity determining means for determining from said
Doppler frequency information, and information as to whether the sign of
the Doppler shift is positive or negative, the Doppler frequency for each
channel, said means including means accumulating digital counts
representative of said zero crossing; and
means for determining the sign of said Doppler shift and supplying said
information to said means for determining velocity, including
means for comparing the quadrature components of said quadrature signals to
determine the instantaneous sign of said Doppler frequency shift for each
channel, and means for comparing said accumulating counts against a
predetermined frequency, and upon said counts exceeding the value
corresponding to said predetermined frequency, substituting the sign which
was determined just prior to exceeding said value, in preference to said
instantaneous sign. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
DESCRIPTION
This invention relates generally to ultrasound imaging of internal
structures and fluids, and more specifically, to improvements in the
measurement, imaging and mapping, by use of the Doppler principle, of the
movement and flow of internal body fluids, especially blood, and the
vessels containing such flow, for example, blood vessels and heart walls.
BACKGROUND OF THE INVENTION
Ulrasonic technology has in recent years become ever more important in
medical diagnosis. Such technology finds application where it is desired
to examine internal body organs and fluids with the objective of locating
features or aspects which may be indicative of disease or absnormalities.
Typical instruments detect the amplitude of the echo signals returning
from the structure being examined, and usually display the information in
a two-dimensional "B Scan" image. Less common, and less straightforward,
is the detection of velocity along the axis of the interrogating sound
beam, rather than amplitude. Such detection can provide an image of the
blood flow pattern, or vessel network, information of high diagnostic
significance. The detection of velocity is based upon the Doppler
principle, whereby a change in observed frequency of the reflected echo
pulse is indicative of a corresponding change in the velocity in the
region from which the echo emanates.
Fortunately, much basic work on such Doppler based flowmeter systems has
already been done. A basic system is described in M. Anliker, titled
"Current and Future Aspects of Biomedical Engineering", Triangle, Volume
16, No. 3/4, 1977, 129, 130-132. Another later system of the type is
described in M. Brandestini, "Topoflow-A Digital Full Range Doppler
Velocity Meter", IEEE Transactions in Sonics and Ultrasonics, Volume
SU-25, No. 5, September 1978, Pages 288-291. Other similar papers are
"Blood Flow Imaging Using a Discrete Time Frequency Meter", Brandestini
and Forrester, 1978 Ultrasonic Symposium Proceedings, IEEE Catalog
78CH1344-ISU and F. D. McLeod, M. Anliker, "A Multiple Gate Pulsed
Directional Doppler Flowmeter, Proceedings IEEE Ultrasonic Symposium,"
Miami, December 1971; and F. E. Barber, D. W. Baker, D. E. Strandness Jr.
and G. D. Mahler, "Duplex Scanner II", Ultrasonic Symposium Proceeding,
IEEE Catalog 74, CHO08961SU, 1974.
Most of the prior Doppler systems involve an RF ultrasonic pulse
transmitting and receiving section, and some form of quadrature phase
detection, transmitting by means of a transducer an interrogating pulse
train into the structure under examination and receiving and resulting
echo information for processing. The RF frequency is of the order of
megacycles, while the pulse repetition frequencies are typically in the
kilohertz range. Thus, a pulse repetition interval of 100 to 200
microseconds between pulses may be expected, and a useful range of no more
than 10 to 20 centimeters into the patient's body. Along with the received
actual echo signal, the quadrature echo signal, accomplished by mixing
with a local oscillator signal differing in phase by 90.degree. from the
transmitter frequency, preserves phase and the ability to later detect the
sign, i.e., flow direction, of the fluid movement under examination.
The next section typically found in such Doppler systems usually involves
the sampling of the Doppler information-carrying envelope with both the
original and orthogonal detected echo signal into a number of channels
similar to the number of microseconds interval between the pulses,
typically into 128 channels. Of course, each such channel or portion of
the time interval between pulses also corresponds to a portion or interval
of a range within the patient's body under interrogation. If structures or
fluids in any such interval within the body have a velocity component in
the direction of the axis of ultrasonic radiation, a Doppler frequency
change is impressed upon the echo emanting from such interval or channel.
However, complicating the detection of such Doppler frequencies is the
existence of quasi-specular stationary reflecting tissue interfaces which
yield echo signals of large amplitudes, thus masking the much lower
amplitude signals scattered by the moving blood cells, and which actually
contain Doppler information of interest. The difference in amplitude may
be as much as two orders of magnitude. Such large echo "clutter" signals,
being from stationary interfaces, have no Doppler information, and change
little, if at all, from one interrogating pulse to the next, while the
echo signals from moving scatterers such as blood will change rapidly. It
was realized that decomposing the received echo signals into the large,
but relatively fixed, clutter components, and a small, rapidly-changing
signal component could provide the key to resolving this masking problem.
Accordingly, as the next stage of some Doppler systems, certain
investigators, (especially Anliker and Brandestini), utilized
analog/digital recursive stationary canceler-filters based on principles
first utilized in radar in order to digitize the incoming signals and
remove therefrom those components from fixed tissue interfaces, and
low-pass filter the non-blood flow low-frequency Doppler signals from such
reflectors as moving lumen walls. Such filters attempted to split the
analog digital version of the sampled echo signals into a "fast" and a
"slow" section, and utilized a tracking type conversion relying on the
Doppler difference between subsequent pulses to track and eliminate
clutter, since it is nearly constant from pulse to pulse, while attempting
to digitize with maximum speed the small-amplitude, but fast-changing
Doppler portion of the system remaining after subtraction of the clutter.
Such digital canceller-filters, while a substantial improvement over prior
expedients, nevertheless have not performed well enough to enable Doppler
flowmeter systems to function at practical levels, with performance levels
sufficient to provide truly acceptable commercial instruments. Rather
severe demands on such filters result from the fact that not only are the
amplitude changes in the echoes as much as two orders of magnitude
different, but also such changes are very sudden, and may cause transfer
of low frequency energy content into neighboring channels. The handling of
such large-amplitude changes obviously requires a high degree of
resolution and dynamic range, and the suddenness of such changes further
requires short response times, if substantial amounts of Doppler
information-bearing echo signals are not to be lost under the influence of
clutter signal amplitudes, and because of the time required for the filter
to respond and to eliminate same. In these respects, the stationary
filter-cancellers of the prior art have been less than satisfactory, and
have been found lacking, especially in the dynamic range and response time
necessary to perform at a sufficient resolution level.
The filter-canceller stage has then typically supplied the input for a zero
crossing detector and a companion flow velocity sign detector circuit. The
function of the crossing detector is to detect the zero crossings which
the Doppler signal undergoes over a period of time in one direction for
each channel, which then gives the measure of the Doppler frequency. The
sign detector is important in determining whether a positive or negative
Doppler shift is occurring within each channel. This is critical for the
operation of the conventional Doppler frequency to velocity converter with
which these systems are finally equipped, and which then yield a velocity
for each channel. Such converters accumulate over some predetermined time
period counts corresponding to the occurrence of zero crossings for each
channel, and must be instructed as to whether the zero crossing is in one
sense or the opposite sense, that is, whether the count should be added or
subtracted. The sign detector, by comparing quadrature components through
which phase information has been preserved, and obtaining the
instantaneous direction of the Doppler frequency, supplies such
instruction to the velocity converter.
Again, while such zero crossing detectors and sign detector means have been
the best expedients heretofore available, they, too, have had serious
shortcomings. These have primarily to do with the inherent limitations of
sample systems, in particular as imposed by the well-known Nyquist
criterion. In other words, in quadrature multiplexed systems, it is
well-known that when the detected frequency exeeds one-quarter the
repetition rate of the interrogating acoustic pulse, the phase information
can no longer be preserved. Thus, while the fact of a zero crossing may be
reliably detected in such prior art expedients, its direction will not be
reliably detected, under the foregoing conditions. Therefore the operation
of any Doppler frequency to velocity converter in the prior art under
these circumstances is correspondingly also unreliable and unsatisfactory.
Accordingly, it may be regarded as an object of the present invention to
provide a Doppler flowmeter system with improved resolution, dynamic range
and response time to enable a practical level of performance in measuring
the flow of body fluids.
It is a further object of the invention to provide a Doppler system with an
improved digital stationary canceler-filter for a Doppler flowmeter system
having improved dynamic range and response time at high resolution level.
It is a still further object of the invention to provide a Doppler system
with improved zero crossing and sign detector circuits having an extended
Doppler bandwidth for improved handling of sampled echo information to
assure satisfying the Nyquist criterion.
SUMMARY OF THE INVENTION
The foregoing objects of the invention are met by providing an improved
sampled ultrasonic system utilizing the Doppler principle which measures
the velocity of fluids moving within internal body structures. The system
includes means for transmitting an acoustic beam comprised of addressing
pulses of ultrasonic-frequency acoustic waves having a regular pulse
repetition rate into the body structure under consideration, and for
receiving the acoustic echo energy from both stationary and moving
acoustic scatterers within the path of the beam, and finally converting
same into corresponding electrical echo signals. The echo signals, which
contain both Doppler and stationary echo information, are also demodulated
into quadrature echo signals in order to preserve phase change information
imparted by the moving scatterers. Means are provided for sampling the
quadrature echo signals into a plurality of equal channels, each
corresponding to a portion of the time between addressing pulses. Each
channel then also corresponds to an equal portion of the overall range of
the addressing pulse within the body. This system also includes filter
means accepting the quadrature echo signals and canceling therefrom any
stationary non-Doppler information. Also provided are means for
determining from the digitized form of the remaining Doppler echo
information the velocity of the fluids under interrogation.
In accordance with one particular aspect of the invention, the filter,
inserted between the means for sampling and the velocity determining
means, and servicing each of the channels, employs a recursive loop
including a digital memory for comparing the echo signals from each
addressing pulse with the immediately proceeding echo signal, a subtractor
to subtract any unchanged non-Doppler components for a given channel, and
an analog digital converter rapidly digitizing the remaining changing
components of the echo signal. Such a recursive filter requires a definite
number of iterations in order to handle large amplitude stationary echo
signals which are reflected, for example, from vessel walls, and which
mask the much lower amplitude signals containing Doppler information from
moving body fluids within such vessels. The time required for the filter
to respond to such situations detracts from the filter's ability to "see"
Doppler information bearing signals, which may then be lost. Accordingly,
response time is greatly improved by a comparator means in parallel with
said analog/digital converter, for identifying the highest active digital
bit between the digital dynamic range of the incoming echo signal, and the
dynamic range of the analog/digital converter, said comparator providing
said highest active bit to be added along with the output of said
analog/digital converter into said memory, so that upon subsequent
comparison with incoming echo signals for said given channel, subtraction
therefrom, and subsequent further iterations thereof, the number of
iterations is decreased, and the response time is correspondingly improved
for large stationary echo signals. In this manner, the dynamic range,
response time, and resolution of the system is greatly improved as
compared to earlier expedients.
In accordance with a further aspect of the invention, means are provided in
the system for determining, from the signal which remains after processing
by the filter and which comprises Doppler echo information in digital
form, the zero crossings in one direction of said signal, in order to
obtain the Doppler frequency for each channel. This information is
directed to the means for determining the fluid velocity, which also
requires instruction as to whether the sign of the Doppler frequency shift
is positive or negative, and which includes means accumulating digital
counts representative of the zero crossings. Supplying such instruction as
to the direction of the Doppler shift is a means for determining the sign
of the Doppler shift. This means includes means for comparing the
quadrature component of the quadrature echo signals to determine the
instantaneous sign of the Doppler frequency shift for each channel. Also
included is means for comparing the accumulated counts within the
frequency to velocity converter against a predetermined frequency (the
Nyquist criterion), and upon said counts exceeding the value corresponding
thereto, substituting the sign of the velocity determined just prior to
said value being exceeded, in preference to said instantaneous sign. In
this manner, the limitations of the sampled system in accurately detecting
the direction of the Doppler frequency shift, and consequently accurately
determining the velocity, are obviated within the ranges which concern
human diagnostic needs. It is well known that quadrature sampled systems
such as the foregoing are normally limited because of the Nyquist
criterion. Thus, if the accumulated count of a channel indicates a Doppler
frequency above this criterion of one-quarter the addressing pulse
repetition frequency, it is well-known that phase can no longer preserved,
and thus, the usual sign detector instruction to the frequency to velocity
converter as to whether to add or subtract pulses representative of zero
crossings is no longer reliable. The foregoing arrangement overcomes these
limitations to a substantial extent, and in effect extends the bandwidth
of the system to accommodate Doppler bandwidths normally beyond the
Nyquist criterion.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is an electrical block diagram schematic illustrating an improved
ultrasonic Doppler flowmeter system in accordance with the present
invention;
FIG. 2 is a graphical depiction of the various electrical signals
transmitted, received and produced by the system of FIG. 1, plotted on a
comparable time scale or spatial range scale, to illustrate the
functioning of the system of FIG. 1;
FIG. 3 is an electrical circuit block diagram of the improved
filter-canceler of the Doppler system of FIG. 1, in accordance with the
further aspects of the present invention;
FIG. 4 is a graphical depiction of response time of the filter of FIG. 3,
as compared to prior art filters; and
FIG. 5 is an electrical circuit block diagram of the improved zero detector
and the improved sign detector of the Doppler system of FIG. 1, in
accordance with another aspect of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
A complete ultrasonic Doppler flowmeter system in accordance with the
invention is illustrated in FIG. 1. Initially, means are provided for
transmitting an RF ultrasonic pulse via a transducer 10 into the body of a
patient whose internal fluid flow or blood vessel structure is desired to
be examined. Associated therewith are means for receiving echos resulting
from the reflections of the interrogating pulse set up by the internal
tissue interfaces and scatterers, such as the blood flowing within the
vessels under examination. FIG. 2 may be referred to to show the
orientation of the transducer 10 with respect to the patient's skin 11,
vein 12, and artery 13 below the skin. Transducer 10 is excited with
radio-frequency energy by transmitter 14 operating at a typical frequency,
of the order of four megahertz, under the influence of oscillator 16 to
emit pulses of corresponding the high frequency ultrasound. The
transmitter is pulsed at a pulse repetition frequency of the order of 6.4
kilohertz, thus allowing an interval of some 160 microseconds between two
consecutive interrogating pulses. The length of time travelled by sound
within such 160 microsecond intervals, corresponds to roughly 10
centimeters, which defines the range of the instrument into the body from
the transducer-skin interface within which echos can be picked up from
structures of interest.
Line B of FIG. 2 plots this interrogation time, or the corresponding range,
along the horizontal axis, out to 10 centimeters. Line B of FIG. 2
illustrates the initial interrogating burst as it starts out along the
range into the body. As the interrogating pulse travels inwardly, it will
excite return echo information tion, of amplitude and phase which depends
on the degree of discontinuities encountered, and their motion. This is
graphically illustrated by Line C of FIG. 2. It will be seen, for example,
the echos emanating from the tissue interfaces at 12a at the outer wall of
vein 12, as well as the similar interface at 12b at the inner vein wall,
are very large in amplitude as compared to the echos set up by scattering
from the blood flow within the vein. In actuality, the difference in
amplitude may be as much as two orders of magnitude.
Such echo information, received back from the transducer 10 and transduced
into corresponding electrical signals as illustrated in Line C, then is
passed to the mixer-demodulator 20, where it is mixed with local
oscillator signals, one differing in phase by 90.degree. from the
transmitter signal, by means of phase delay 21 and local oscillator 16. In
this manner, two signals are in effect provided, the detecting original
received signal carrying the echo information, and the orthogonal signal.
This quadrature arrangement enables phase to be preserved, and enables the
detection of the sign, whether negative or positive, of the Doppler
velcity component by the sign detector circuit to be described below.
The mixer-demodulator 20 separates the original Doppler information
carrying envelope of both the original and orthogonal RF echo signal, and
then directs each of the quadrature demodulated signals to sample and hold
units 23 and 24, respectively. Multiplexer 25 alternately switches between
the two sample and holds to alternately present the demodulated echo
signals of each channel to stationary canceler-filter 26, (which actually
is duplicated in parallel to accommodate both of the orthogonal
demodulated echo signals of each channel at the high sampling rate of the
order of 1 sample per microsecond). Thus, with the aid of sample and hold
units 23 and 24 and multiplexer 25, the demodulated echo signals are
scanned into a plurality (here 128) of uniform time intervals or channels,
as illustrated in line A of FIG. 2, with total time equal to the interval
between successive interrogating pulses. Of course, each such channel or
portion of the time interval between pulses also then corresponds to an
interval of the 10 centimeter range within the patient's body which is
under interrogation. The overall problem of detecting the position and
magnitude of velocity changes within a patient's structure then is
resolved into one of detecting the Doppler frequency change at the level
associated with each of the channels of the range. The particular number
of channels is not critical, and has been chosen for convenience, in view
of the capabilities of the digital components, the sampling times, and the
pulse repetition and ultrasound frequencies involved.
The large amplitude "clutter" signals as mentioned above mask the lower
amplitude signals from the moving fluid scatterers which contain the
Doppler information sought to be detected. However, the influence of such
clutter signals may be substantially mitigated by the use of stationary
canceler-filter 26. This filter utilizes the fact that the clutter
signals, being from stationary or quasi-stationary tissue interfaces,
change little if at all from pulse to pulse, in contrast to the
rapidly-changing smaller-amplitude signals from the moving fluid
scatterers. The stationary canceler-filter of the present invention is
illustrated in FIG. 3.
The filter includes an analog subtractor 30 which receives inputs from
multiplexer 25 and which includes an output connected to both 8-bit
analog/digital converter 31, and a Highest Active-Bit Detector 32,
hereinafter the HABD 32. These two units both have outputs supplying a
digital adder 34, while the A/D converter 31 also supplies its output to
the zero crossing detector of FIG. 1. Adder 34 in turn inputs a storage
unit 36, which may be a RAM. Storage 36 is a memory of 16 by 128-bit
capacity, since 128 channels will have to be serviced, and for each, a
16-bit word capacity is provided. Storage 36 is connected both to 16-bit
digital-to-analog converter 37 which in turn is connected to the second
input of subtractor 30, and also supplies a second input of adder 34 with
its output. The basic recursive nature of the operation can be readily
appreciated from FIG. 3, but the finer points of the iterative process
under which the filter operates, requires more detailed explanation of the
operation of the filter.
In operation, the analog values within the sample and hold units 23 and 24
are directed by multiplexer 25 in turn for each of the channels into one
input of subtractor 30. A representative one of such values for
convenience may be termed M.sub.p, that is, the echo values associated
with a transmit or interrogating pulse p in a representative channel R.
The immediately preceding value in that channel R, from the immediately
preceding interrogating pulse p-1, is then M.sub.p-1. Let us assume that
the latter has been earlier processed by the filter, digitized, and is now
stored in storage 36 of the filter. Note that in this particular case, the
data path is 16 bits wide, which is approximately 100 decibels of dynamic
range. It has been found that such a dynamic range is needed to resolve
the Doppler information from background clutter. Thus, the digital/analog
converter 37 is 16 bits in capacity. Because the operation of the filter
will eliminate the need for the higher order bits by filtering the
clutter, A/D converter 31 is of 8-bit capacity. Thus, the faster
components of the echo signals, which have a spectral component above the
stationary clutter, should be adequately serviced, in theory.
Now let us examine the basic and normal mode of operation for the recursive
filter in channel R as echo information from a new interrogating pulse p
is received. Upon presentation of the value M.sub.p to the positive input
of analog subtractor 30, the previously stored value M.sub.p-1 from
storage 36 and previously associated with channel R is taken from the
storage unit output, and presented to digital/analog converter 37. The
latter then directs the information, now in analog form, to the negative
input of analog subtractor 30, where it is subtracted from the already
queueing value M.sub.p belonging to pulse p. Thus, the difference value
M.sub.p -M.sub.p-1 is formed at the output of subtractor 30 and directed
to analog to digital converter 31 and to the input of comparator 32. As
will be explained below, the comparator's output is zero and it
effectively is out of the circuit as long as the dynamic range of A/D
converter 31 is not exceeded. Under this assumption, A/D converter 31
digitizes the difference value (M.sub.p -M.sub.p-1) whereupon the value is
passed to the zero crossing detector of the FIG. 1 system as will be
described, and also to one of the inputs of digital adder 34. Digital
adder 34 through another of its inputs also receives the stored value
M.sub.p-1 from the output of storage 36. Also receiving the stored value
M.sub.p-1 from storage 36 is digital/analog converter 37, which thereafter
passes it to the minus input of analog subtractor 30. Digital adder 34
then performs the addition M.sub.p-1 +M.sub.p -M.sub.p-1 of its input,
resulting in the digital value M.sub.p. The latter is then stored in
storage unit 36 and thereupon becomes available to average against the
next subsequent value M.sub.p+1 for the same channel R, which value will
result upon the occurrence of the next following interrogating pulse p+1.
Thus, a recursive loop is described, and a means for cancelling the
unwanted low frequency components by in effect tracking the portions of
the signals which do not change between subsequent pulses and removing
same, while digitizing the residual remaining small amplitude signal,
which should be the Doppler component, at high speed.
It has been found that in any such filter system as thus far described
without the benefit of HABD 32, and where the maximum number of bits in
the digital data stream is N (any number such as 10, 14, 16 etc.), and the
digitizing bit rate for the Doppler component at the A/D converter 31 is n
(any number such as 4, 6, or 8 etc.), then the longest period in which the
input value is not the same as the digital output is given by the
expression
2.sup.(N-n) /f.sub.PRF
where f.sub.PRF is the pulse repetition frequency, and 1/f.sub.PRF is the
corresponding time period of the repetition.
In this case, of course, this input value is the value of the echo signal
for the channel being studied, and the digital output is the corresponding
digital form of the echo signal, out of which the higher order stationary
components have been filtered. The foregoing expression is then a measure
of the "dead time" of the filter, during which the filter is not seeing
useful small-amplitude Doppler-information-bearing signals from moving
scatterers because of masking by a large clutter signal.
Another way of expressing the same type of concept is that if the filter is
saturated by a large clutter signal of high order, equal to or surpassing
its 16-bit capacity, the quantity
2.sup.(N-n)
gives the number of iterations which the recursive loop of the filter must
perform in order to respond to the clutter signal and to eliminate same.
Of course, the number of iterations is a measure of response time
necessary to take care of such clutter signals. A system with as small a
number of iterations as possible is highly desirable in terms of response
time, and to eliminate dead time during which the filter is not reading
useful Doppler information. Filters along the above-described lines,
without the benefit of the action of HABD 32, have been found to have
unacceptably long response times when faced with processing high-order or
saturation-level clutter signals, as too often occurs in the real world,
and thus, to be highly susceptible to problems in handling signals
exceeding the dynamic range of the A/D converter.
It is instructive to appreciate the numbers of iterations necessary in such
a system to handle a saturation signal. Let us assume, to simplify the
analysis, that digital/analog converter 37 is of 7-bit capacity, while
analog digital/converter 31 is only of 4-bit capacity. Thus, in the above
"dead time" formula, N=7, and n=4 for this example. An incoming clutter
signal saturating the system would then represent 128 counts, i.e.,
2.sup.7, the highest number of counts handleable by the data lines. The
A/D converter 31 would then be saturated with its maximum number of
counts, or 2.sup.4 =16. In accordance with the above mode of operation,
A/D converter 31 would pass a count of 15 through adder 34 and into
storage 36, whereupon it would return via D/A converter 38 to be
subtracted from the 127 input count, to result in 112 counts residue.
Another recursion would then take place, with D/A converter again
inputting its maximum 15 count value into the loop and the now 30 counts
in the storage would then be subtracted from the inputted 127, leaving 97
counts after the second iteration. Upon the third iteration, the memory
would contain 45 counts which would then be subtracted from the
still-saturation 127 input counts, still leaving 82 residue. It may be
seen that the iterations would continue for 8 cycles, until the residual
counts finally dropped below the 15 count dynamic capacity of A/D
converter 31, in accordance with the above expression.
By contrast, it has been found that a unique and very marked improvement in
dynamic range and response time, for a given number of bits resolution,
can be obtained by the implementation of a system as above with HABD 32
used and positioned as described. In fact, the filter with HABD 32
functioning will only have a maximum number of iterations represented by
the quantity:
(N-n)
rather than 2.sup.(N-n) as is the case with the non-HABD-based system as
described above. Similarly, the longest period over which Doppler
information contained in the output of the fast A/D converter 31 may be
lost, is now given by the quantity:
(N-n)/f.sub.PRF
where f.sub.PRF is, as before, the pulse repetition frequency, N is the
maximum number of bits in the data path, and n is the digitizing bit rate
at A/D converter 31. The present system thus obviates the slow response
heretofore typically seen with such filter/cancelers, and thus cuts the
number of iterations, and consequently the dead time during which the
filter does not respond to Doppler information, very drastically,
especially for systems accommodating the higher dynamic ranges,
resolutions and greater number of bits in the data path.
HABD 32 itself is structured to identify the highest active bit between the
n-bit and the (N-1)-bit whenever the dynamic range of the fast n-bit A/D
converter is exceeded. In the actual described embodiment, since n=8 and
N=16, the comparator identifies the highest active bit between the 8th-bit
and the 15th-bit whenever the 8-bit dynamic range of the A/D converter is
exceeded by a saturation signal. This highest active bit is then fed into
the digital adder 34, together with the saturated value output from the
A/D converter 31, and the iteration by the recursive loop continues as
above described until the dynamic range of A/D converter 31 is no longer
exceeded.
In order to effect this performance, HABD 32 includes two sets of a like
plurality of comparators, the plurality being given by the quantity
[N-(n+1)]; in this case, seven comparators per set. One set of seven is
for the positive signals, and one set is for the negative. Thus, HABD 32,
like the analog/digital converter 31, is bipolar, that is the summation
point in the recursive loop is consistent for both positive and negative
signs, and sign magnitude consistency is maintained. In addition to these
comparators, a logic circuit is included which monitors the output of the
comparators and identifies the one with the largest positive or negative
threshold, and thus the corresponding highest active bit.
It is instructive to extend the above simplified analysis of the iterations
necessary to handle a saturation signal to the present system, as a rough
measure of the improvement obtained as compared to earlier filters. We
again assume the same conditions as previously, i.e., N=7 and n=4, an
incoming saturation-level clutter signal of the maximum 127 count value,
and thus saturation for the A/D converter again at a value of 15 counts.
HABD 32 will detect and add the highest active bit between N-1 and n. In
this case, N-1-n=2. Thus, the HABD is to detect the two upper bits (below
the MSB), then present them to the adder along with the low order bits
from A/D converter 31, which then go into storage 36 and are subtracted
from the input signal in the recursive fashion as previously. The two
upper bits in this case are represented by 2.sup.6, 2.sup.5, whose value
is 63. Thus, the HABD and A/D converter deposit the value 63 counts into
adder 34 and thus into storage 36. This value is then subtracted from the
initial saturation input value of 128 counts, leaving a residue of 65
which is still greater than the dynamic range of 16 counts of A/D
converter 31; thus the iteration continues. After another input of 63
counts to the memory, which is then subtracted from the residue of 65
counts, we are down to 2 counts, which is within the range of the A to D
converter 31. Thus, within three iterations, as predicted from the
aforementioned expression, (N-n), the large-amplitude saturating signal
has been eliminated and the filter is again operating within the dynamic
range of the fast A/D converter 31. A significant reduction in the
response time of the filter to large input signals is then manifested.
FIGS. 4A and 4B show the results of comparison between the response time of
the instant system, and the response times of corresponding earlier
filters without the improvements herein. In FIG. 4A, intervals a and c
represent the times during which a filter without the present improvements
recursively executes iterations upon reception of a large stationary echo
signal component. These of course are also time periods during which any
received Doppler information is being masked out. Intervals b and d then
represent times after reception of large stationary components during
which Doppler information is actually being processed. These intervals a
and c represent "dead times" for the filter. The much improved situation
for a filter in accordance with the invention is illustrated on a
comparable scale in FIG. 4B. The dead times a and c after reception of a
large stationary component are drastically reduced, leaving
correspondingly much increased intervals b and d during which the filter
sees Doppler information.
At this point in our description of the overall FIG. 1 system, we have
arrived at the output of the canceler-filter 26, which output is in
digital quadrature multiplexed binary offset code, and contains the
Doppler information, with the stationary components eliminated, for each
channel in turn and for both quadrature components. This signal is then
passed to a zero crossing detector 48, which also includes a threshold
noise eliminator in the form of a Schmitt trigger. The object is to detect
the number of zero crossings which the Doppler signal undergoes over a
period of time, which then gives a measure of the Doppler frequency, and
hence, the velocity of the structures or fluids under observation. Working
with the zero crossing detector is a flow velocity sign detector 50, which
in turn is inputted by the zero detector and which then enables
determination of the direction of the flow of the fluid under observation,
that is, whether the flow is toward or away from the transducer.
FIG. 5 shows both zero crossing detector 48 and sign detector 50 in more
detail. With the incoming signal in binary offset form as described, the
digital data bus for each channel may be truncated to a single synchronous
pulse string with the aid of the Schmitt trigger 51, functioning as a
threshold noise eliminator. As shown, the most significant bit (MSB) of
the incoming signal, which is the sign bit, goes directly to zero crossing
detector 48, while the remainder of the bits d | | |
