 |
Description  |
|
|
The present invention relates to a universal motion and intrusion detection
system, and more particularly to a method and apparatus for utilizing a
video camera and associated circuitry to detect motion in a given field of
view and to sound an alarm when such motion is detected and/or to focus
attention on the motion.
While the present invention is described herein with reference to a
surveillance system, it should be clear that the invention is applicable
to any other type of video or television system wherein it is desired to
detect motion, changes in grey scale, changes of position or intrusion in
a given field of view of a video camera, and to sound an alarm and/or
focus on the motion and follow the motion. Moreover, while the invention
is described with respect to a conventional type of video camera, the
techniques are equally compatible with low-light level infrared as well as
the visible spectrum. There is no limitation as to the type of photosensor
used in the system.
Several video surveillance techniques are generally known, as exemplified,
for example, by U.S. Pat. Nos. 3,590,151 and 3,740,466, both of which are
assigned to Jackson and Church Electronics Co., Inc. The video detection
systems presently available, including the systems described in the above
two U.S. patents employ video averaging techniques which gives them a
certain inherent minimum threshold of sensitivity. This characteristic
does not allow the prior art systems to be used for detecting all forms of
motion. The video averaging motion detectors consider a certain area of an
alarm site and on an analog basis integrates and stores the video
waveform. The prior art systems then re-scan the area on the next field or
the next frame, re-compute the average and look for differences in this
average value. Due to circuit noise and the quality of the analog
processing and frame-to-frame drifts in the circuitry, the resulting
system has a minimum sensitivity which render them unsuitable for reliable
detection of certain types of motion.
The main object of the present invention is to provide a more sensitive and
accurate motion and intrusion detection system than as presently known in
the art. More particularly, an object of the present invention is to
provide a system which does not use the video averaging techniques,
generally used in the art, in determining motion and/or intrusion in a
given field of view.
Another object of the present invention is to provide such a system which
is capable of focusing on the motion in the given field of view, mapping
it and/or zooming in on the motion and tracking same.
A further object of the invention is to provide a system in which only a
portion of the field of view of the video camera can be easily selected
for detection of alarm conditions only in the selected portion of the
field of view. This can be accomplished by means of a "light pen" or the
like.
A still further object of the present invention is to provide a digital
system which is substantially drift-free, thereby eliminating need for
frequent adjustment, and further allowing repeatable digital thresholds to
be set with regard to intruder size, intruder motion and intruder gray
scale.
A still further object of the present invention is to provide a system
which accurately pin points the motion and which further describes the
nature of the motion.
A further object of the present invention is to provide a system which is
less prone to generating false alarms. More particularly, an object of the
present invention is to prevent false alarm generation from certain forms
of motion that are not intrusions, such as snow, rain, wind vibration and,
for example, tree and other wind-generated motion when the alarm site is
out of doors.
A still further object of the present invention is to provide a system
which can discriminate an alarm condition further based on the size and
speed of the intrusion.
SUMMARY OF THE INVENTION
According to the present invention, a motion and intrusion detection system
comprises a video camera for viewing a given field of view and for
generating video signals corresponding to the field of view. The video
signals are then converted into a plurality of coded digital signals which
correspond to the grey level of the video signal at a plurality of points
in the field of view of the camera, and the coded digital signals are
stored in a main storage device. The coded information corresponding to
the plurality of points from a first scan of the video camera is compared
with coded information corresponding to the same plurality of points
generated during a subsequent scan of the video camera on a point-by-point
basis. Means is provided for generating an alarm signal when the
difference between the compared signals for corresponding points exceeds a
predetermined value, and an alarm indication means indicate the detection
of an alarm condition.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1a-1d show a basic block diagram of an embodiment of the present
invention;
FIG. 2 illustrates the alarm analyzer of the present invention in greater
detail;
FIG. 3 illustrates the sync generator of the present invention in greater
detail;
FIG. 4 illustrates the alarm area selector of the present invention in
greater detail;
FIG. 5 illustrates the pulse generator 32 of the present invention in
greater detail;
FIG. 6 illustrates the selector 19 of the present invention in greater
detail;
FIG. 7 illustrates the alarm map clock generator in greater detail;
FIG. 8 illustrates the field selector unit of the present invention in
greater detail; and
FIGS. 9-14 illustrate various mathematical relationships used in explaining
a mathematical model of the system of the present invention;
FIGS. 15a-15g show a waveform diagram at indicated points in FIG. 1; and
FIG.. 16 illustrates a modified form of the invention.
DETAILED DESCRIPTION OF ILLUSTRATED EMBODIMENTS
Before discussing the invention in detail, a discussion of the general
principles of operation of the detection system of the present invention
is given below.
In accordance with the present invention, specific points in time are
monitored on a video waveform which represent a given video field. A
pattern of sample points is generated on the given video field. The video
field used in this specific embodiment comprises commercial television
rate frequencies modified to present 256.5 lines per field or a total of
513 lines per frame. This includes the lines generated during vertical
retrace. The sample points are taken on every line of the field, there
being 128 sample points per line. The numbers used herein are only by way
of example, and can be increased or decreased in accordance with the
particular application. The pattern of the sample points on a line-to-line
basis is continually shifting in a pre-programed format. The total number
of points being sampled in a given field in the illustrated embodiment is
16,384. Each one of the sample points consists of a period of time that is
approximately 20 nanoseconds duration. Since a horizontal sweep at T.V.
frequencies consists of 63.5 microseconds, the percentage sample time
represents less than 0.0315 percent of the sweep time if it is further
considered that the bandwidth of the video camera is on the order of 4
megacycles. Thus, the sampling duration is on the same order of magnitude
as the video resolution. It has been found that there is no advantage or
improvement in sensitivity by reducing the sample period further than 20
nanoseconds. This sample point is carefully ear-marked in time by the
coordinates of a synchronous vertical counter which counts horizontal
sweeps and its position horizontally is defined by a precision oscillator
synchronized on a submultiple of the horizontal frequency which counts the
position of the sample point on the horizontal sweep line. This coordinate
information is known because of the synchronous generation of vertical
sync, horizontal sync and sample points. Therefore, every point is
referenced to the vertical even sync which is the top left of the field of
view. The sample point level which represents gray scale information of
the sample point is quantized by an analog-to-digital converter (A/D
converter) which provides, for example, 16 levels of quantization or 4
bits. The levels of quantization may be varied depending upon system
requirements. The A/D converter necessarily has to be extremely fast and
therefore is a full parallel A/D converter (Grey code).
The gray scale information for the sample point and its coordinates are
stored and then on the subsequent field of the same type (odd or even),
the same scan position is located by monitoring the vertical position
counter and the horizontal position counter until the same coordinates
reoccur and again a 20 nanosecond sample of the video is taken. The
subsequent sample is processed through the same 4 bit parallel A/D
converter. This new sample is compared with the previously stored sample
on a numerical basis. The quantization process involves converting the
analog signal into a Grey coded binary signal, which allows for
differences to be observed while avoiding ambiguities, at the threshold
points between levels. The two stored values are then converted into a
binary number and subtracted from each other. If a difference exceeding,
for example, plus or minus 2 is observed, an alarm condition exists. The
alarm condition which does not necessarily institute an alarm condition
external to the equipment, contains three important characteristics. The
time of the alarm, the magnitude of the alarm, and the exact location of
the alarm (since it relates to a particular sample point). This
information is available as three digital words. With this stored
information, the scanning process and the monitoring of the 16,384 alarm
points continues, and if another sample point indicates a change in level,
several decisions can be made based on the following criteria.
1. The time difference between the alarms in conjunction with the distance
of the two alarm sample points, gives a measure of the intrusion velocity.
The process of measuring velocity is a continuous one and covers many
points. It must be recognized that this process is occurring very rapidly
since the time between fields is approximately 32 milliseconds. This means
that 30 alarms can occur within a period of one second, for a small high
speed intrusion (i.e., intruder size on the order of the size of a sample
point).
2. The number of alarms that occur within one field gives an indication of
the size of the intrusion when related to the camera's field of view.
As the alarms occur, a recording device such as a video tape recorder may
be instantly started to record the entire intrusion interval until it is
automatically or manually stopped. In addition, the digital words
providing X and Y coordinates of the sampled point in the monitored site,
and the intensity (i.e. Grey scale) are stored and identified in a fashion
that indicates the location and Grey scale within the frame in which the
alarm took place. It is this digital information that provides the basis
for an alarm map which may be displayed on a monitor.
The mapping can be done without XY coordinates because, in the illustrated
embodiment, when a point is alarmed it is entered into a point memory
(16,384 bits) which is synchronously clocked with the sample point
generator clock.
The digital XY information identifies the location of the intrusion
relative to the camera's field of view. In other words, the field of view
could be considered as being centered at the origin of a Cartesian
Coordinate System and the digital alarm location, which is, for example,
an eight bit binary number, in both X and Y, provides information as to
where the motion occurred. Therefore, by using these coordinates the
camera can be remotely directed by a servo mechanism to re-center the
origin of the camera's coordinate system at the point where motion
occurred. From the size estimates, gathered by the differences in X and Y
addresses occurring in one given field, information can be gathered as to
the magnification required to fill the camera's field of view with the
motion. For example, if alarms were such that in one field that the
greatest difference in X coordinates and Y coordinates was the binary
number 26, the camera could be instructed to increase its magnification by
a factor of 10. This digital difference signal would be provided to a
motor-controlled zoom lens via, for example, a radio link. The largest
difference in coordinates would determine the maximum allowable degree of
magnification.
The detector generally performs a point analysis of 16,384 points
periodically and provides numerical data to the subsequent digital
processing circuitry to evaluate the nature of the alarm. The circuitry
for detecting and analyzing and quantizing the video signal, may be
provided only in a central station, allowing all the central processing
hardware to be common to all sites. This allows a minimum of hardware at
the remote monitored sites.
Referring to FIG. 1, a video camera 1 is provided to "observe" a given
field of view, hereinafter referred to as the "monitored site". The video
camera is provided with a transceiver 2, which may suitably be a low power
transceiver. The transmitter portion of the transceiver 2 transmits the
video information from the camera to a central station 5. The receiver
portion of the transceiver 2 receives positioning information which is
used at the "monitored site" to vary the position of the video camera 1
and/or to operate the zoom lens system thereof. When a plurality of remote
cameras are used, it is necessary for the transceiver to have decoding
equipment so as to be able to differentiate the information intended for a
particular remote site from information pertaining to other remote sites.
Also, if desired, means can be provided at the remote sites so as to be
responsive to general information received from the central station to
verify "normal" operation of the electronics at the remote site.
The video transmitter 3 at transceiver 2 amplitude modulates the video
signal with approximately a 6 megacycle base band response and transmits
it at, for example, UHF frequencies back to the central station 5. If the
frequency is sufficiently high, the transmitted beam width can be very
narrow and directed at the central station antennas. For beam widths of
approximately 2.10 of a st radiant, 10 millowatts of RF will provide 2
mile range with sufficient signal-to-noise ratio. Higher powers may be
used as desired. The video transmitter 3 runs on a continuous basis
providing video information to the central station 5 for analysis. Video
transmitters 3 from different remote sites are frequency multiplexed,
thereby allowing the central station 5 to identify the source of
transmission.
The receiver 4 receives coordinate information for the purpose of
positioning to the camera in digital form as well as digital messages
indicating the degree of magnification that is required. In addition, the
transmission from the central station is coded to a specific monitoring
site. The nature of the signal is preferably FSK and transmissions are
broadcast on the same frequencies to all sites and the sites are
identified by a digital address that is recognized by the FSK receiver.
Since this information is critical to the security of the site, redundant
coding is preferably used to insure the reliability of the transmission.
Further, information requested by the central station from the remote
sites will be requested via coded digital messages. These messages will
allow verification that the remote site electronics are functioning
properly, and responses from the remote sites will be coded into the video
waveform (i.e., during blanking intervals).
The remote site video electronics will be discussed in further detail
hereinbelow. The TV camera is a generally commercially available TV
camera, as is the basic transmitter and receiver, 3 and 4.
Still referring to FIG. 1, the central station comprises a transceiver 6
comprised of a receiver 7 and transmitter 8. The receiver is a
multi-channel UHF receiver capable of receiving video information from all
sites simultaneously and for receiving such information on a frequency
multiplex basis. The bandwidth required for the receiver 7 is
approximately 6 megacycles times the number of remote sites being
monitored. A modem 9 is coupled to the transmitter 8 for coding a message
which is desired to be sent to a particular remote site. Since each remote
site will have its own individual binary address, messages will carry the
site address which must be recognized by the particular site. The output
power of the transmitter 8 is determined by the frequency being used and
by the maximum distance required for the central station to transmit. The
central station further includes means, described hereinbelow, to detect
when a remote site fails to respond to interrogations from the central
station so as to cause an alarm condition to exist. This provides a means
to detect electromagnetic jamming attempts, power failure, or equipment
failure at the remote site.
The output of the receiver 7 is fed to a video buffer amplifier which
receives a demodulated waveform from the receiver 7 and filters and shapes
it, and provides the demodulated video wave with sufficient gain so as to
be useful in the remaining sections of the central station during
processing. The output of the video buffer amplifier 10 is fed to a
parallel-view A/D converter 21 which continually converts the incoming
video signal into a digital word. The A/D converter 21 has 16 levels or
thresholds and can divide the video signal into any one of these levels
with a high degree of precision. Since the A/D converter 21 is fully
parallel it operates on a continuous basis. The information fed out of the
A/D converter 21 is provided to a four bit Grey-to-binary converter 22.
The code used for the conversion is a Grey code. This is to insure that if
the video signal is going through one of the transition points of level
during a strobing period, no more than one bit will be in error. This is a
similar ambiquity problem to that generally found in encoding
technologies.
The A/D converter 21 has adjustable thresholds. This is a very important
feature since the portion of the video scenario in which alarms are being
monitored may be of a different average light level than the overall light
level in the monitored area. Therefore, adjustments are provided that can
be of an automatic or manual nature such that once the alarm within the
field of view of the camera is established, the minimum and maximum video
levels for that area will provide the extremes for the 16 bit converter
21. This gives the A/D converter 21 considerably more power as far as
resolution goes than if these thresholds were fixed. In the event the
threshold levels are to be set manually, an upper threshold control 21a
and lower threshold control 21b will be provided on the central station
monitor that will allow the operator to visually and very rapidly set
these levels by observing a special display for this purpose on the
monitor. This will be described in greater detail hereinbelow.
The speed at which the parallel A/D converter 21 will work will be well
beyond the bandwidth of the video camera. This insures that there will not
be a tracking error due to a conversion delay. Analog comparators are
readily available that will make level decisions in approximately 10 to 20
nanoseconds. Thus, conventional commercial grade integrated circuit
comparators can be used.
A latch circuit 20 receives the output of decoder 22 to "hold" the
information fed thereto so that new information from decoder 22 can be
inhibited from passing through for alarm comparison. Latch circuit 20 may
comprise a plurality of D-type flip-flops, one flip-flop for each incoming
signal line, as illustrated in FIG. 1 and the detailed function thereof
will be described later hereinbelow.
A further decoder 23 is provided to convert the Grey binary word from latch
circuit 20 into a natural binary word for each comparison or sample point
so that a numerical comparison can be made between the quantized value of
the present sample point and the previous corresponding sample point. The
converter 23 performs the conversion on a continuous basis in less than 35
nanoseconds.
The dynamic storage unit 24 is coupled to the output of decoder 23 via a
selector switch means 19 and stores the sample points as binary numbers
for one field. The storage media is, for example, LSI dynamic storage
registers coupled together to provide 65,536 bits of storage. The storage
register 24 is shifted precisely by a clock signal generated by a
phase-locked loop circuit in generator 11 at 2.0MHz. The re-circulating
register 24 is under the control of the position controller 15 via the
selector 19. In a storage mode the position controller 15, via gate 40
thereof sets the selector switch 19 to connect the output of the
Grey-to-BCD converter 23 into the register 24 and the sample point data is
entered into register 24 serially for one field. During the next field the
position controller 15 connects the Grey-to-BCD converter 23 (via latch
20) and the dynamic storage register 24 into a comparator or subtractor 25
wherein the sample points are numerically compared or subtracted.
The comparator 25 performs the numerical comparison between the stored data
in the dynamic register 24 and the real time sample point data coming in
from the Grey-to-binary converter 23. Comparator 25 operates on a purely
numerical basis by taking the information from the storage register 24 and
the information from the converter 23 and subtracting one from the other.
Since there is the possibility of a quantizing error occurring in the A/D
converter 21, a difference equal to or less than 1 is not considered an
alarm condition. Any difference in magnitude greater than 1 is considered
an alarm condition.
The output of numerical comparator 25 is fed to a magnitude detector 26
which can be manually set in terms of resolution (or sensitivity of the
system) by increasing the difference required for an alarm to be greater
than 1, i.e., 2 or more. This is done by varying the reference number set
into magnitude detector 26. This has the effect of coarsening or reducing
the overall system resolution. The magnitude detector 26 may comprise a
comparator or subtractor, similar to comparator 25.
The alarm analyzer 13 takes all the present data on what constitutes an
alarm and analyzes it to see if the alarm conditions being received from
the numerical comparator 26 constitute an actual alarm. The alarm analyzer
13 may receive inputs that relate to alarm size, alarm sensitivity, and
alarm movement, and further may make decisions based on the number of
alarms occurring per field to determine whether an actual alarm has
occurred. The alarm analyzer 13 determines the degree of automatic alarm
analysis that is used in the system to reduce the false alarm rate on an
automatic basis. Depending on the expected false alarms that might be
encountered in the particular monitored site, different programming and
different sets of alarm conditions will be provided to the alarm analyzer.
A simplified alarm analyzer 13 comprises only a counter which will indicate
an alarm condition only after a preset number of alarm signals are gated
thereto. Additionally, a reset may be provided so that only if a preset
number of alarms is received in a given period of time an alarm condition
will be indicated. Such an alarm analyzer 13 is illustrated, for example,
in FIG. 2 wherein a first counter 30 receives the vertical sync signal and
counts the number of vertical sync signals received. Counter 30 will
provide an output only after a predetermined number of vertical sync
signals, which number has been fed into the counter 30, is received. A
main counter 31 is provided to receive the gated alarm signal (see FIG. 1)
and if a predetermined number of gated alarm signals, which number is fed
into the counter 31, is received before a reset signal is provided by
counter 30, an output is provided to the alarm indicators 14.
A line-locked sync generator 11 is provided for generating horizontal and
vertical synchronizing signals and sampling point signals which are
precisely timed and locked to the line. The leading edges of the vertical
and horizontal sync signals are an accurate representation of the
beginning of the video and horizontal scans. The sync generator is shown
in more detail in FIG. 3.
The outputs of the sync generator 11 are fed to a position controller 15.
The position controller 15 receives the various clocking signals from sync
generator 11 and tracks the video cameras scanning in the vertical plane
and counts the horizontal synchronizing pulses up to 256 pulses, thereby
providing a complete field. The position controller 15 includes a field
selector 16 which selects the number of fields to be skipped between
successive alarm detections which preferably comprises a counter 16a and a
selectable decoder 16b. The count in the counter 16a is used to determine
the fields in which the intrusion or motion detection is to be carried
out. For example, if the selector switch 16c is set to the position shown
in FIG. 1, the system will conduct motion or intrusion detection every
other field of the video signal. If the selector switch 16c is set to the
64 ms position, the system will conduct detection every fourth field of
the video signal.
The position controller 15 further includes a vertical position controller
17, or otherwise termed a line selector. The line selector 17 likewise
comprises a counter 17a and a decoder 17b coupled to the output thereof. A
selector switch 17c is coupled to the output of the decoder so as to
select which lines of the frame in which alarm detection is to be carried
out. For example, with the selector switch 17c set as shown in FIG. 1, the
system operates on every other line of a field. A frame includes 513
lines, but in the implementation of the present invention, only one half
of the lines are used per field. Therefore, a setting of 128 of selector
switch 17c provides alarm detection on every other line of a frame.
Likewise, setting switch 17c to the 64 position, alarm detection is
conducted every fourth line per frame.
The position controller 15 further includes a point selector 18 for
determining on which points of a given line alarm detection is conducted.
For example, a line contains 128 points in the present embodiment and
setting the switch 18c to the position shown in FIG. 1 provides alarm
detection on every point on a line. Setting switch 18c to position 64
provides alarm detection at every other point on a line. The point
selector 18 comprises a counter 18a coupled to a selectable decoder 18b,
similar to the line and field selectors.
The outputs of the selectors 16, 17 and 18 are coupled to respective gates
40-43 as shown in FIG. 1 so as to provide the appropriate gated output
control signals to the remainder of the system. Gate 43 is a multi-input
AND gate which gates the alarm signals as a function of the settings of
the field, line and point selectors. In the present embodiment, alarm
detection is constantly taking place, and alarm indication is inhibited as
a function of the settings of the field, line and point selectors. This is
more efficient from a logic point of view than gating and controlling the
individual inputs to the substractor 25, for example, to control when
alarm conditions are to be considered. Referring to FIG. 8, the field
selector 16 is shown in greater detail. The line selector 17 and the point
selector 18 are of substantially identical construction, except that the
counters thereof and decoders are scaled with the appropriate numbers
required. In FIG. 8, the vertical sync signal is applied to a counter 16a,
the outputs of which are coupled to a decoder 16b, the design of which is
conventional. Switch 16c, which, in the illustrated embodiment comprises
three switch sections, is set by the operator so that the decoder 16b will
provide an output at the appropriate timing. The three switch sections of
switch 16c are shown as a signal switch section in FIG. 1 for convenience,
as are the switches 17c and 18c. The output of the decoder is applied to a
flip-flop 54 so as to shape the output and to hold same for a
predetermined period of time. The flip-flop is reset through an AND gate
and the next vertical sync pulse after the selected vertical sync pulse.
The output of the flip-flop 54 is the field select signal which is then
coupled to AND gates 40, 42 and 43.
The position controller further includes an alarm area selector 27 which
determines which specific area of a given field is to be considered for
alarm detection. The alarm area selector is shown in more detail in FIG.
4. The two outputs of the alarm area selector 27 are coupled to AND gate
43 through an AND gate 44 and a switch 45. As clearly seen from FIG. 4,
the selector 27 is responsive to the vertical and horizontal sync signals
and generates an output which gates the alarm signals to an alarm counter.
The first one shot multivibrator 46 has an adjustment resistor 47, as is
conventional, and determines the vertical position at which alarm
detection is to begin. The setting of the variable resistance 47 varies
this vertical position. The output of one shot 46 is fed to a one shot
multivibrator 48 having an adjustment resistance 49. The setting of one
shot multivibrator 48 determines the length of time that alarm detection
is to be carried out and effectively determines the vertical length of the
portion of the field that is subject to alarm detection. The horizontal
sync is fed to a one shot multivibrator 50 having an adjustment resistance
51. The setting of one shot multivibrator determines position in a given
line that alarm detection is to begin. The output of one shot
multivibrator 50 is coupled to a further one shot multivibrator 52 having
an adjustment resistance 53 which determines the position in the
horizontal line that alarm detection is to cease. The output of one shot
multivibrator 52 is gated with that of one shot multivibrator 48 in AND
gate 44 and is fed to the gate 43 for effectively gating the alarm signal.
The result of the provision of the alarm area selector 27 is that given
rectangles on the screen can be selected for alarm indication. Anything
that occurs outside of the selected rectangle will have no effect on the
generation of an alarm indication by the system.
As discussed above, alarm detection by means of the subtraction unit 25 and
alarm magnitude detector 26 is constantly being carried out at every point
being sampled. The AND gate 43 responds to the various gating signals
supplied thereto by the field selector 16, line selector 17, point the
selector 18, alarm area selector 27 so as to gate alarm indications
through to the alarm counter only at selected points in the frame. The
alarm area selector need not have extremely accurate timing elements since
minor variations in the timing of the various one-shot multivibrators
which comprise same will not introduce false alarms into the system. For
example, if the rectangular area selected by the alarm area selector is
supposed to begin at a given point in a field, if, on the next field the
timing of the one-shot multivibrator is such as to cause the alarm area
selector to initiate the rectangular area one point to the right of the
original point, this will have no effect on false alarms since the output
of alarm area selector merely gates alarm signals through the gate 43 to
the counter 13. The alarm area selector 27 has no effect on the comparison
circuitry wherein alarm conditions are initially detected.
One of the outputs of the position controller 15 is fed to a pulse
generator 32 which generate a "latch" signal which is coupled to an
inhibit input of the latch circuit 20 and a "gate" signal which is coupled
to one of the inputs of AND gate 43. The pulse generator is illustrated in
detail in FIG. 5 and comprises three serially connected one-shot
multivibrators 33-35. The output of the first one-shot multivibrator 33
comprises the latch signal and is preferably, in the present embodiment,
approximately 250 nanoseconds in duration. The second multivibrator has a
delay of approximately 70-80 nanoseconds, which is sufficient for the
latch circuit to "settle down", and the output of the third one-shot
multivibrator 35 provides a signal of approximately 100 nanoseconds
duration, but the leading edge of which is delayed by about 70-80
nanoseconds from the leading edge of the latch signal. The latch and gate
signals are also shown in FIG. 5. Thus, gate 43 is not enabled until a
sufficient amount of time (approximately 70-80 nanoseconds in the present
embodiment) has elapsed for the remaining circuitry to "settle down",
thereby preventing transient signals from affecting alarm indication. The
input signal to the pulse generator 32 is the anded output of the point
selector 18 and line selector 17. Pulse generator 32 thereby provides the
latch and gate pulses for each point on each selected line.
Referring again to the memory 24 and selector switching means 19, it is
pointed out that the selector 19 is operable as a function of the output
of AND gate 40 of the position controller 15. The output of the AND gate
40 is the "anded" combination of the outputs of the field and line
selectors. Thus, during the selected field and lines, the selector 19 is
operable so as to feed new information from the decoder 23 into the memory
24. During the non-selected fields and lines, the selector 19 is operable
so as to re-circulate the previously stored information back into the
memory 24 in their proper position. As illustrated in FIG. 1, the signals
a, b, c, and d correspond to "new" information coming in from the decoder
23, and signals a', b', c' and d' correspond to re-circulated information
which had already been stored in memory 24. The selector 19 is illustrated
in more detail in FIG. 6, and comprises a group of AND gates 56
interconnected with the "new" information signal line a and the
"re-circulated" informa | | |
