 |
Description  |
|
|
FIELD OF THE INVENTION
The field of this invention relates generally to providing apparatus and
method for human voice control of certain types of welding systems.
BACKGROUND OF THE INVENTION
Human voice controlled machines are known in the art as are remote control
systems for welding equipment. U.S. Pat. No. 4,216,367 to Riseberg, issued
Aug. 5, 1980, entitled "Wireless Remote Control for Electric Welder",
discloses a welding machine which can be remotely controlled. The welding
machine has a controllable power supply which is initially set by the
human operator to provide a desired power level using a main rheostat. An
operator, working at a location remote from the welding machine, carries a
pen-type adjustable rheostat which carries calibrations indicative of
various percentages of the preset welding current. If the human operator
wishes to change the magnitude of the welding current, he need not return
to the main rheostat at the welding machine, but instead sets the remote
rheostat between the electrode holder and the workpiece. A current
transformer in the welding machine senses the output (calibrating) current
flowing through the welding cable. A signal proportional to the desired
welding current reference is compared with the stored value of the present
welding current reference, to produce the control signal, which is fed to
an up/down counter whose output is the stored welding current reference.
The welding current reference is thus either increased or decreased until
it reaches the value called for by the adjustable remote rheostat. After
the power level has been adjusted as desired, the remote rheostat is
removed from the electrode holder and placed like a pen back into the
pocket of the operator.
U.S. Pat. No. 4,266,114 to Hansen issued May 5, 1981, entitled "Apparatus
for the Remote Control of Mains Operated DC and AC Welding Machines",
discloses a portable regulating device intended to be connected into the
welding current circuit between the electrode or electrode holder and the
workpiece or clamp therefor. A regulating device selectively generates
different control signals in the welding cables when receiving a current
supply through the welding cables. A regulating circuit in the welding
machine is adapted to receive these control signals through the welding
cables from the regulating device, and to cause regulation of functions in
the welding machine.
The regulating circuit consists of a decoding circuit which, in response to
the pulse-shape control signals in the welding cables, produces regulating
signals for the welding machine, depending upon the number of pulses or
the code of pulses being provided by the regulating device. A blocking
circuit prevents the formation of regulating signals in the decoding
circit when currents of the same order of magnitude as the welding current
flow in the welding cables.
U.S. Pat. No. 4,275,266 to Lasar issued June 23, 1981, entitled "Device to
Control Machines by Voice", discloses a machine which responds to a
plurality of predetermined musical tones which are delivered in a sequence
in order to generate a digital control output signal. The audio signal is
converted to a sequence of digital number signals which are temporarily
stored in a memory. Then a sequence of ratio signals are generated by
performing arithmetic division in a systematic order. The resulting
sequence of ratio signals is then converted to digital numbers comprising
the digital control output signal. A microprocessor, with its associated
storage or memory, processes the digital data and controls the conversion
of the musical tones into corresponding digital numbers. The device is
then made electrically connectable to an apparatus whose operation is to
be voice controlled.
U.S. Pat. No. 4,340,797 to Takano and Ueda issued July 20, 1982, entitled
"Voice Actuated Heating Apparatus", discloses a heating apparatus such as
an electric oven which includes a voice recognition part capable of
recognizing voice commands of the user. It produces a recognition code by
receiving the voice command and then performing the operation commanded.
Heating sequences are preliminarily stored in a memory of a control part,
one of them is selected by a recognition code produced by the voice
recognition part, and preset in the memory. Heating members are controlled
in a manner so as to heat the object according to the preset heating
sequence.
U.S. Pat. No. 4,340,799 to Ueda, Takano and Suzuki issued July 20, 1982,
entitled "Heating Apparatus with Voice Actuated Door Opening Mechanism",
discloses a heating apparatus such as an electric oven having an enclosure
with an openable door having a locking means. A voice recognition circuit
for recognizing voice commands of the user of the apparatus is connected
to a releasing mechanism which permits opening the door in response to a
human voice command.
The above-cited U.S. Patents are incorporated by reference into this
Specification. Generally, speech recognition machines fall into two broad
categories, machines which recognize either (1) isolated words, or (2)
continuous speech. Because of the complexity of human speech, most
research has concentrated on solving specific tasks, such as recognizing
either isolated words or continuous speech with a small vocabulary.
Fortunately, many practical applications exist for limited recognition.
Isolated word recognition, for example, is adequate for logging freight
destinations in warehouses, or identifying and counting items for
inventory control.
Even though word recognition machines vary greatly in detail, they all use
the same basic recognition process. First, a spoken word is converted into
an electrical signal by a microphone. Second, the signal is processed to
extract a set of identifying features. Third, the features are then
compared to a library of templates representing the machine's vocabulary.
Fourth, a word is recognized if it matches one of the templates stored in
the machine's memory.
The stored reference templates are created either in a laboratory
(speaker-independent), or by using the recognition machine itself in a
special training mode (speaker-dependent). When training the machine with
his voice, the human speaker repeats each word several times to enable the
machine to compute an average template for (1) that word, and (2) that
speaker (or class of speakers). The analysis stage of speech recognition
consists of the voice recognition unit's extracting identifying
characteristics from the electrical analog of the speech signal. Speech
recognition machines are described in "Unraveling the Mysteries of Speech
Recognition", Michael Elphick, High Technology, Vol. 2, No. 2, March/April
1982, pages 71-78; which is hereby incorporated by reference into this
Specification.
None of the above-identified patents or articles provides a means for
verbally adjusting the power level of an overall welding system, or
suggests a combination which would provide for a speech-controlled power
adjustment as part of a welding system.
The application of voice recognition technology to welding systems presents
many problems. The higher levels of audible noise and electromagnetic
interference found in the welding environment are examples of problems
which adversely affect the operations of the electronic equipment required
for voice recognition.
Making power adjustments manually by the welder-operator is often time
consuming and costly, and is sometimes dangerous. For example, a welder on
a construction crew, often several floors above the location of a welding
power supply, precariously perched on scaffolding or steel I-beams and
holding a welding torch, often cannot make the necessary adjustments
readily. He might have to return to the ground level where the power
supply is located to adjust the power output manually, he might employ an
assistant to make the changes for him, or he might carry a foot pedal,
connected to the welding power supply by cables, with which he can make
the necessary adjustments. Alternatively, if the welder has to manually
adjust a device such as that disclosed in U.S. Pat. No. 4,216,367 cited
above, there is a time lag in the power adjustment as well as
inconvenience to the operator in having to use his hands to make the power
adjustment.
Therefore, there is a need for an apparatus and method which would permit a
remotely located welder to quickly and safely adjust the power output of a
welding power supply located at some distance from the welder, while
welding is actually taking place.
OBJECTS AND SUMMARY OF THE INVENTION
Accordingly, in order to resolve the above and other problems of the
present technology it is a general object of this invention to provide
apparatus and method for human voice controlled welding systems.
More specific objects of this invention are to provide welding apparatus
and method which permits a human operator who is physically removed from a
welding power unit to give verbal commands which will:
(a) adjust the power level (either voltage and/or current) delivered to the
welding torch being used by the human operator; and/or
(b) start, stop, and control an internal combustion engine (gasoline or
diesel) used to drive the welding power supply, when such an internal
combustion engine is used; and/or
(c) adjust the speed at which a consumable wire electrode is fed to the
welding torch, when such a consumable wire electrode is used.
Additional objects, advantages and novel features of the invention will be
set forth in part in the description which follows, and in part will
become apparent to those skilled in the art upon examination of the
following, or may be learned by practice of the invention. The objects and
advantages of the invention may be realized and attained by means of any
instrumentalities and combinations particularly pointed out in the
appended claims.
In one embodiment a human operator is provided with an audio transmitter
fitted with a microphone which is capable of receiving an acoustic command
signal spoken by the operator. The transmitter converts the acoustic
command signal into a transmitted command signal such as electromagnetic
waves. A receiver is provided, and receives and converts the transmitted
electromagnetic waves into an audio electrical signal in the form of an
electric current which contains audio information corresponding to that of
the acoustic command signal. An audio interface receives and conditions
the audio electrical signal into a series of audio interface output
signals. A Voice Recognition Unit (VRU) having a VRU memory contains
previously stored voice recognition information (i.e., templates) in the
VRU memory, and is capable of receiving and comparing the audio interface
output signals with the voice templates stored in the VRU memory, such
that only those audio interface output signals which match the voice
templates stored in the VRU memory are forwarded as VRU output signals.
A computer with computer memory receives and compares the VRU output signal
with welding information stored in the computer memory, and computes a
computer digital command signal, which is transmitted as a series of
computer output signals. A computer interface is connected to the computer
to receive and convert the computer output signals into computer interface
output signals. The VRU, the computer, and the computer interface form a
Voice Recognition System (VRS). A power control interface receives the
computer interface output signals arriving from the computer interface,
and supplies an output of power control signals. A welding power supply is
connected to a power source by electric power lines, for supplying
electric power. The welding power supply is provided with power control
means and a power output line connected to a welding torch. The power
control means may be preset to the desired power level and range by the
human operator prior to use of the welding torch. The welding power supply
is responsive to the power control signals produced from the acoustic
command signal, to thereby adjust the power control means to supply a new
power level to the welding torch.
The power control means on the welding power supply may be adjusted, using
a human voice, by the steps of: (a) manually pre-setting the desired power
range of the welding power unit; (b) positioning a transmitter within
speaking distance of a human operator; (c) transmitting an acoustic
command signal to adjust the power level; (d) receiving the command
signal; (e) routing the command signal to a Voice Recognition Unit (VRU)
having VRU memory for recognition and conversion into a computer
recognizable VRU output signal; (f) sending the VRU output signal to a
computer which translates the VRU output signal into a computer output
signal; and (g) converting the computer output signal to an analog signal
and selectively applying that signal to the welding power source to
control the operation thereof.
In a second embodiment, the Voice Recognition Unit (VRU) and the digital
interface are combined into a Voice Recognition System (VRS) without a
separate computer; the additional functions of the computer are performed
by the VRU. The VRU memory contains previously stored voice recognition
information. The VRU is capable of receiving and comparing the audio
interface output signals with the voice recognition information (i.e.,
templates) contained in the VRU memory. Only those audio interface output
signals which match the voice recognition information in the VRU memory
are forwarded as VRU output signals. These VRU output signals are
transmitted to a digital interface which converts VRU output signals into
digital interface output signals.
In a third embodiment, the welding torch is adapted to receive and consume
during welding a consumable wire electrode. The power output line is
adapted to receive within itself the wire electrode, which is wound onto a
rotatable wire spool mounted within a wire electrode supply unit. In
response to verbal commands received and forwarded by the VRS and the
other components with which it cooperates, the wire electrode is advanced
by motor driven first and second rollers, through the power output line,
toward and into the tip of the welding torch for consumption during
welding.
In a fourth embodiment, an engine means is mechanically coupled to the
welding power supply. The welding power supply is provided: (a) internally
with an electric power producing means responsive to the engine means to
produce electric power, and (b) with a power output line which is
connected to a remote welding torch. The welding power supply may be
preset to a desired power range by a human operator prior to use of the
welding torch. The welding power supply is responsive to the power control
signals generated in response to the acoustic command signal, to thereby
adjust the power control means to supply a new power level to the welding
torch in response to the acoustic command signal. Additionally, the engine
means is responsive to the engine control signals generated in response to
the acoustic command signal, thereby allowing control of the engine means.
The benefits and advantages of the present invention over the prior art
will become particularly apparent from the Detailed Description below. The
most significant improvement offered by this invention is that it provides
apparatus and method permitting a human welder to verbally adjust the
power delivered to the welder's hand-held welding torch, conveniently and
immediately, and without interrupting the welding process, simply by
issuing verbal commands into a transmitter which communicates through a
VRS with the welding power unit. When a consumable wire electrode is used
in the welding process, this invention provides method and apparatus for
verbally controlling the speed at which the consumable wire is fed through
the welding torch. When used at remote locations where there are no
electric power lines and the welding power supply is driven by an internal
combustion engine, this invention provides the apparatus and means for
verbally controlling the internal combustion engine.
BRIEF DESCRIPTION OF THE DRAWINGS
The accompanying drawings, which are incorporated into and form a part of
the Specification, illustrate various embodiments of the invention and,
together with the Detailed Description, serve to explain the principles of
the invention. In the drawings:
FIG. 1 is a simplified block diagram showing the present invention and how
it differs from the prior art.
FIG. 2 is a schematic view of the assembly and operation of a first
embodiment of the voice controlled welding system of this invention.
FIG. 3 is a block diagram of the power control interface shown in FIG. 2.
FIG. 4 is a schematic view of additional embodiments of the assembly and
operation of the voice controlled welding system of this invention.
FIG. 5 is a block diagram of the power control interface shown in FIG. 4.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
The present invention is particularly useful in electrical welding systems
utilizing gas tungsten arc welding (GTAW) and gas metal arc welding
(GMAW). Both the GTAW and GMAW systems are continuous processes wherein
the welding operation is performed manually by a skilled and well-trained
welder. An example of the GTAW system is shown in FIG. 2. A GMAW system is
shown in FIG. 4. The present invention can be used with a welding system
driven by a source of electrical power (see FIG. 2) or by an internal
combustion (I.C.) engine (see FIG. 4).
Referring now to FIG. 1 wherein a simplified block diagram shows the
present invention and how it differs from the prior art. The prior art
system 100, contained within the dashed lines, is a system having two
elements, human operator 104 and welding equipment 106, which interact
with the welding process 102. To adjust the power output of welding
equipment 106, operator 104 must perform a manual operation, indicated by
dashed line 108, such as depress a foot pedal (not shown) or turn a knob
(not shown), both connected to welding equipment 106. This manual
operation 108 may be performed remotely. Welding equipment 106 then
responds, as indicated with response line 110, by delivering a new power
output to the welding process 102. The operator 104, who is closely
observing the welding process 102, receives sensory information 112
(primarily visual) from the welding process 102 and moves the welding
torch (not shown) along the weld by operator action 114.
In contrast, the present invention 116 incorporates a Voice Recognition
System (VRS) 118. Now, according to the invention, operator 104 issues a
voice command indicated by line 120, which is received by VRS 118,
interpreted, and then sent as an electrical command signal indicated by
line 122. Welding equipment 106 receives and responds to signal 122, to
generate the response indicated with line 110, which is a new power output
for use in welding process 102. Therefore, the manual adjustment 108 of
welding equipment 106 is eliminated and replaced by verbal operation using
the means of voice command 120, VRS 118, and signal 122. FIG. 2 shows a
GTAW welding system 10 according to an embodiment of the invention in
which a human operator 11 carries a transmitter 12 having a microphone 14.
The microphone 14 and transmitter 12 are attached to welding mask 13 worn
by welders to protect their vision and face during welding, or
alternatively, to the operator's clothing or person.
The welding system includes a welding power supply 44. The power output of
supply 44 may be adjusted by regulating the voltage and/or current.
Current controls, however, are much more common. Thus, the present
invention will be described with respect to a power supply 44 that is
regulated by adjusting current.
Operator 11 must perform two steps prior to commencing welding: (1)
manually pre-set the power range of the welding power supply 44 with power
control means 54 to obtain the desired range of current, which current
flows through power output line 46 to the hand-held welding torch 48 used
by operator 11 to weld on workpiece 56; and (2) if the Voice Recognition
Unit (VRU) 28 provided with voice recognition unit memory (i.e., VRU
memory), discussed below, is not preprogrammed, then operator 11 must
place VRU 28 in a special training mode to "teach" the VRU memory of unit
28 to recognize the specific commands or words, as spoken by that speaker
or operator.
If operator 11 is using welding torch 48 on workpiece 56 but desires to
adjust the power output of welding power supply 44, operator 11 speaks a
command into microphone 14, which converts this acoustic command signal 16
into an electrical signal which can be converted by transmitter 12 into a
transmitted command signal 18. Alternatively, wire 17 can be connected to
and span the distance between transmitter 12 and receiver 20, through
which wire the electrical signal representing acoustic command signal 16
can be sent. Receiver 20 receives the electrical signal 18 and converts it
into an analog audio electrical signal 22, which contains essentially the
same information (but in different form) as acoustic command signal 16,
the command issued by operator 11. Audio electrical signal 22 is sent to
audio interface 24 for conditioning to produce an audio interface output
signal 26 which can be accepted by the voice recognition unit 28.
A voice recognition unit (VRU) 28 with VRU memory receives signal 26 and
compares it to information (i.e., voice templates) stored in its memory.
If signal 26 matches one of these stored voice templates, then VRU 28
transmits this recognized command as a series of digital VRU output
signals 30 to computer 32 which is provided with its own computer memory.
VRU 28 performs the functions of: (1) comparing the audio interface output
signal 26 to the voice templates which operator 11 stored in the memory of
VRU 28 in the training mode (speaker-dependent), or which information was
preprogrammed into the VRU memory by, for example, the manufacturer of the
VRU 28 (speaker-independent), and (2) generating the appropriate output
signal 30 when an audio interface output signal 26 is found to match one
of the voice templates.
As alluded to, there are two classes of speech-recognition systems:
speaker-independent and speaker-dependent. These systems may recognize
either isolated words or connected words.
A speaker-dependent system requires users to "train" it to their voices,
using a predetermined vocabularly. To recognize a speaker's words, the
voice input system analyzes samples of that person's speech. These samples
capture the normal variations of that speaker in the form of voice
templates.
Most commercially available voice-input systems today are
speaker-dependent, because of their lower cost and relatively higher
performance.
Speaker-independent recognition allows any person's voice to be recognized
without that individual training the system. For each word in the system's
vocabulary, speech samples from hundreds of persons must be collected,
properly processed, and clustered. Only then can an unknown person's
speech be reliably recognized.
Isolated-word recognition requires a brief silence between each word to
allow the system to separate the spoken words from one another and
eliminate the effects of coarticulation. A shortcoming of isolated-word
recognition is the time it takes to complete the data entry.
Connected-word recognition eliminates the short pauses between words.
Hence, users can string a series of words together for faster data entry.
Connected-word recognition is becoming less costly, more efficient, and
more widely accepted through the use of low-cost powerful digital signal
processing chips.
Speech recognition systems may utilize a digital inverse filter approach.
This approach involves four steps. An autocorrelation step extracts
relevant parameters from the spoken word. These parameters are passed
through an inverse filter formed by the reference templates. When a spoken
word matches a word whose template is stored in memory, the energy in a
residual or output signal falls below a threshold. A dynamic programming
algorithm allows a match to occur even if the spoken word is uttered at a
different speed from the stored word. An event detector senses possible
matches and passes them to a high-level decision function. There the right
word is selected from the pre-stored reference set.
Preprogrammed voice templates (speaker-independent) for each word in the
VRU's vocabularly are made up from samples of speech from many persons.
Alternately, in the case of a "known" welder or speaker
(speaker-dependent), the VRU is placed in a teach or training mode and
that speaker repeats the words in the vocabulary several times so that the
VRU can build templates of that speaker's speech pattern. Speech
recognition systems are described in greater detail in the
above-referenced article by Michael Elphick, "Unraveling the Mysteries of
Speech Recognition", incorporated by reference.
An example of commercially available electronic voice recognition unit 28
is the Voice Recognition Chip Model VRC 008, made by Interstate Voice
Products, a Figgie International Company, of Orange, Calif., described in
the four-page brochure dated September 1984 entitled "Voice Recognition
Chip Model VRC 008", which is incorporated by reference into this
Specification. The VRC 008 system employs a unique method for processing
of analog speech data and recognition of spoken utterances. Designed for a
wide variety of voice automation products, the system is
speaker-independent, and can recognize with high accuracy eight spoken
words or phrases, translating verbal commands (e.g., stop, faster, slower,
etc.) into action by associated circuitry.
The manufacturer can customize the VRC 008 for a specific user vocabulary
by programming the templates or recognition parameters for the selected
vocabulary in the VRC 008. The VRC 008 recognizes speech by detecting
significant parameters in the incoming word phrase, and comparing them
with the stored voice templates of the prespecified vocabulary. With
recognition accomplished, the system then generates an output signal
specific to the recognized word.
In the context of the present invention, the VRC 008 could be programmed to
respond to the following commands for the applications and actions listed
below:
______________________________________
COMMAND
WORD APPLICATION ACTION
______________________________________
ON IC engine driven welder
Start IC engine
OFF IC engine driven welder
Stop IC engine
UP GTAW or GMAW Increase weld current
DOWN GTAW or GMAW Decrease weld current
FAST GMAW Increase wire speed
feeder
SLOW GMAW Decrease wire speed
feeder
START GTAW or GMAW Start welding power
supply
STOP GTAW or GMAW Stop welding power
supply
______________________________________
As discussed previously, if control over both voltage and current is
desired, the command vocabulary may be modified accordingly. For instance,
as shown below:
______________________________________
COMMAND
WORD APPLICATION ACTION
______________________________________
VOLTS UP GTAW or GMAW Increase voltage
VOLTS DOWN GTAW or GMAW Decrease voltage
AMPS UP GTAW or GMAW Increase current
AMPS DOWN GTAW or GMAW Decrease current
______________________________________
Alternatively, for the GTAW system shown in FIG. 2, where the VRU 28
comprises the standard version of the VRC 008 which has been
pre-programmed to respond to the verbal commands GO AHEAD and STOP, the
welding power supply 44 may be turned on and off by utilizing these
respective commands. The commands FASTER and SLOWER could be utilized to
increase or decrease the power output of the welding power supply.
Computer 32, in response to VRU output signal 30, sends computer output
signals 34 to a computer or digital interface 36, which operates on
computer output signals 34 to produce computer interface output signals
38. These sig | | |
