United States Patent (19)
`Kimura et al.
`
`US005199080A
`5,199,080
`Patent Number:
`11
`(45) Date of Patent: Mar. 30, 1993
`
`54 VOICE-OPERATED REMOTE CONTROL
`SYSTEM
`(75) Inventors: Toshiyuki Kimura; Kazuo Yabe, both
`of Kawagoe, Japan
`73 Assignee: E. onic Corporation,
`s
`21 Appl. No.: 578,706
`
`4,837,830 6/1989 Wrench, Jr. et al.................. 381/42
`4,922,538 5/1990 Tchorzewsm ........................ 38/43
`FOREIGN PATENT DOCUMENTS
`0178699 7/1990 Japan ..................................... 381/42
`. Primary Examiner-Forester W. Isen
`Attorney, Agent, or Firm-Sughrue, Mion, Zinn,
`Macpeak & Seas
`
`A voice-operated remote control system which trans
`Foreign Application Priority Data
`30
`mits a remote control signal in response to a voice com
`Dec. 29, 1989 JP
`Japan .................................. -341629
`mand has a speech recognition circuit for recognizing
`51 Int. Cl. .......................................... H03G 3/20
`s:
`S.C. C.361/issi/43 the voice command. The speech recognition circuit has
`58 Field of Search ..................... 30.4 standard pattern data storage unit for storing a plurality
`395/2
`of standard pattern data with respect to each of voice
`commands. The input voice command is compared with
`the plural standard pattern data for accurate speech
`recognition.
`
`56
`
`References Cited
`U.S. PATENT DOCUMENTS
`4,718,093 1/1988 Brown ................................... 381/42
`... asi/16
`4,725,956 2/1988 Jenkins ......
`
`5 Claims, 10 Drawing Sheets
`
`
`
`
`
`
`
`
`
`
`
`MODE
`SELECTOR
`SWITCH
`
`M:MICROPHONE
`
`
`
`
`
`
`
`
`
`
`
`10A: TRANSMITTER
`
`SC: OPERATION CONTROL SIGNAL
`
`SR:REMOTE CONTROL
`INSTRUCTION SIGNA
`
`16
`
`CONTROLLER
`LEARNING UNIT
`
`TRANSITTING CKT
`
`INFRARED
`LIGHT-EMITTING
`DIODE
`
`Tz -
`
`RC: REMOTE CONTROL
`SIGNA ,
`
`Sw: CONTROL SIGNAL
`
`5: STANDARD
`PATTERN
`STORAGE UNIT
`
`POWER SUPPLY
`CONTROL
`CKT
`
`PA1-PAn
`PM-PM
`
`STANDARD
`PATTERN DATA
`
`Page 1
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`U.S. Patent
`
`Mar. 30, 1993
`
`Sheet 1 of 10
`
`5,199,080
`
`FIG 1
`
`100 : REMOTE CONTROL SYSTEM
`
`101
`
`
`
`102
`
`TRANSMITTER
`
`RECEIVER
`
`103
`CONTROLLED
`DEVICE
`(AW DEVICE)
`
`RC: REMOTE CONTROL SIGNAL
`
`FIG 2
`
`CONTROL CODE
`
`LEADER
`CODE
`
`CUSTOM
`CODE
`
`YEED
`CODE
`
`DATA
`CODE
`
`INVERTED
`DATA CODE
`
`RC:REMOTE CONTROL SIGNAL
`
`Page 2
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`U.S. Patent
`
`Mar. 30, 1993
`
`Sheet 2 of 10
`
`5,199,080
`
`FIG 3
`
`101: TRANSMITTER
`f
`SR:REMOTE CONTROL
`INSTRUCTION SIGNAL
`17
`
`16
`
`CONTROLLER
`
`TRANSMITTING
`
`INFRARED
`LIGHT-EMITTING
`DIODE
`
`-
`
`
`
`
`
`RC: REMOTE CONTROL
`SIGNA
`
`18
`
`
`
`
`
`
`
`15
`
`SPEECH
`RECOGNITION
`CKT
`
`M:MICROPHONE
`
`
`
`Page 3
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`US. Patent
`
`Mar. 30, 1993
`
`Sheet 3 of 10
`
`5,199,080
`
`
`
` s!in«:323::525.“.
`
`om<az<hw
`5E\(En—
`
`can.
`
`
`
`>4aa=m4omhzoo
`
`
`
`mmzoa>4mm=mmuse;
`
`pxoon
`
`FHZDww<m0hw
`
`zmwpp<a
`
`om<oz<hmum
`
`85$qu
`
`4<onm
`
`
`
`32$QOLwoos 3528
`ozfippmzm-pch4
`:60225525:I
`
`5.75528
`
`:2:@5525
`
`:szSEZSSm
`
`my
`
`mzozmomognz
`
`Page 4
`
`AMAZON 1006
`Amazon V. SpeakWare
`IPR2019-00999
`
`P au
`
`
`
`
`
`4<szmponhozmpmzHorIQHHgm
`
`
`
`
`
`
`
`.8528Eozmmuammosfiww
`
`awthzw2<mhnlxoP
`
`mfita
`
`
`
`2,55SEZSzozémaoam
`
`:oHsz
`
`X4<H
`
`
`
`.Amac:
`
`Page 4
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`
`
`
`
`
`
`

`

`US. Patent
`
`mM
`
`mu
`
`01fo4w.hS
`
`0.809991a5
`
`%.qN.
`
`mm
`
`Pm
`
`
`
` pxoonthwoomm:ommamuAVWHw6E
`
`%A9E48528o.—ImoimmHzH
`
`Iomwam
`
`momwmoomm
`I«83095
`
`oo._<z<
`
`
`
` ”kHz:me<mokmmmL.25::uiN92225nPlllllllllllllE05:
`
`1lllllllllllll
`
`<554:83
`
`onmH>Ho-w=Hhuo~
`
`mzozaomoE”Z
`
`Page 5
`
`AMAZON 1006
`
`Amazon V. SpeakWare
`IPR2019-00999
`
`Page 5
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`
`
`
`
`
`

`

`US. Patent
`
`0,“llv”3ua”2:wa
`85me1!$558I.x.5::Mmua-“o\<3‘mmwzoioxoguz“.33$2528E55“mmumPm
`..xz<mm.Eu-EEEE
` mowmmoommonZzwoowm:owwmmuNNmomwmooma
`
`
`
`oo._<z<n_.N 50285208?
`
`zomwawn<mpRSE
`
`
`
`m.......o0n........................................_Zw_mmm:23nJ,$5.8m“In...H5_2555.u..n"omgzgw”uuW85:I.a:ismmmuamSm58E:nmt_w-3328‘we:
`¢~h.3om“vmg225;mw3"$882:388uw.3rIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIILP9n.................,..........................J.mu33"5:25”?
`8.:meh8$2:.._S._
`
`6mN
`
`Amazon V. SpeakWare
`IPR2019-00999
`
`Page 6
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`
`
`

`

`U.S. Patent
`
`5,199,080
`
`
`
`(4) 9 DLJ (3)9 DE (D) 9 DE (D) 9 DEJ(q/g'0/-/
`
`||||||||||||||||-AAA)
`
`
`
`
`
`
`
`Page 7
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`U.S. Patent
`
`Mar. 30, 1993
`
`Sheet 7 of 10
`
`5,199,080
`
`FIG 9
`
`REGI
`STRATION
`ER NUMBER - 1
`BUFF
`S2
`ALK SW PRESSED
`
`S1
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`SPEECH
`REGISTRATION
`MODE 2
`
`(RECOGNITION
`MODE)
`
`OUTPUT VOICE REGISTRA
`TION COMMAND INDICATE
`REGISTRATION BUFFER
`NUMBER
`
`REGISTRA
`TION FISHED
`
`REGISTRATION
`BUFFER NyBER)Nmax
`
`
`
`y
`CANCEL SPEECH
`REGISTRATION MODE
`
`S11
`
`OUTPUT VOICE
`RECOGNITION COMMAND
`
`S12
`
`EFFECT RECOGNITION
`PROCESS
`
`S13
`
`S14.
`RECOGNI
`TION FISHED
`
`IS THE
`NUMBER OF STANDARD
`PATTERN DATA WHICH DO NOT
`COINCpE ONE
`
`OUTPUT REMOTE CONTROL
`INSTRUCTION SIGNAL
`BASED ON RECOGNIZED
`
`
`
`Page 8
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`PoFF
`
`w:2on
`wkozwxumnw\3mm.8u2:23-563
`11Aomkzou
`CG023.2524}:I
`
`SEEE
`
`$305.28
`
`:2:@5255;
`
`5&5
`
`”mica
`
`.58
`
`ESE
`
`mmtgmzéfimop
`
`
`
`223m.55282325.8“um
`
`US. Patent
`
`M
`
`3
`
`01f08teeh”a
`
`080”,991,5
`
`mm?
`
`
`
`CaonEzooomm:ommaw
`
`IHzooomm
`
`xommmw
`
`23.—
`
`mommwoomm
`
`-588$
`
`28me
`
`2o:
`
`mowmwoomm
`
`-Hzooowm
`
`:owwam
`
`2o:
`
`momwmoomm
`
`N—
`
`553m
`
`3:
`
`«203mm
`
`:0»5m
`
`moo:
`
`P-N~
`
`mzozaomoghz
`
`I.
`
`Page 9
`
`~--
`
`cJNN
`
`AMAZON 1006
`_
`Amazon V. SpeakWare
`IPR2019-00999
`
`Page 9
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`
`
`
`
`

`

`US. Patent
`
`0,3
`
`9
`
`a
`
`f
`
`080.,w1
`
`9EI?In“5IlII;II1.I
`M:03:I
`mIIIl
`
`H38%I
`
`r.$8385m‘zoEchmm
`
`mom—mam
`
`
`
`
`(:325::3:2;23:858El$5meII!22:58:II$3805
`
`czazcé$38on
`
`$me8325was:3Eazzn.
`
`mmzoEomoE“z
`
`
`
`eEa583%mmw222>a¢==éN
`
`
`
` COZOEZOOHmrummamnmmw\E6E
`
`P
`
`-NN
`
`AMAZON 1006
`
`Amazon V. SpeakWare
`IPR2019-00999
`
`Page 10
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`
`
`

`

`U.S. Patent
`
`Mar. 30, 1993
`
`Sheet 10 of 10
`
`5,199,080
`
`S3
`
`REGISTRATION BUFFER
`NUMBER -
`
`S32
`PROCESSOR NUM
`ER - ENERGIZE
`ONLY PROCESSOR IN QUESTION
`
`
`
`S33
`TALK SW PRESSED
`
`Y
`
`SPEECH
`REGISTRATION
`MODE 2
`
`S34.
`N
`
`S35
`ATION
`ISTRATION
`
`S36
`
`S37
`
`REGISTRA
`TION FISHED
`
`GD
`
`. FIG 12
`
`ENERGIZE ALL PROCESSORS
`
`OUTPUT VOICE RECOGNITION
`COMMAND
`
`S45
`
`EFFECT RECOGNITION
`PROCESS
`
`S46
`
`S47
`
`N
`
`RECOGNI
`TION FINISHED
`
`GO
`
`
`
`COUNT PR
`R
`T
`
`SENIOR RESULTS”-S48
`S49
`
`SUL
`
`SORS WHOSE RECOGNI
`RESULTS DO NOT
`
`
`
`
`
`
`
`
`
`SS1
`
`RECOGNI
`TION RESULTS
`OF HALF OR NORE OF
`OF PROCESSORS
`CONCCIDE 2
`
`N
`
`S38
`
`UPDATE STANDARD
`PATTERN DATA OF
`PROCESSOR
`
`
`
`OUTPUT REMOTE CONTROL
`INSTRUCTION SIGNAL
`BASED ON RECOGNIZED
`DATA
`
`S54.
`
`EFFECT ERROR
`PROCESS
`
`REGISTRATION
`BUFFER NyBERNma
`
`PROCESSOR
`NUMBE)Pmax
`
`
`
`PROCESSO
`R NUBER + 1
`ENERGIZE
`PROCESSOR
`ION
`IN QUEST
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page 11
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`1.
`
`VOICE-OPERATED REMOTE CONTROL SYSTEM
`
`O
`
`40
`
`BACKGROUND OF THE INVENTION
`The present invention relates to a remote control
`system for remotely controlling various electronic de
`vices, and more particularly to a remote control system
`for remotely controlling devices such as AV (audio
`visual) devices by way of voice commands.
`In recent years, various AV devices such as stereo
`sets, television receivers, cassette tape decks, video tape
`decks, compact disk players, laser vision disk players, or
`the like are equipped with remote control systems.
`A remote control system has a transmitter which is
`usually positioned remotely from a controlled AV de
`15
`vice. The transmitter, when operated, transmits a re
`mote control signal, such as an infrared remote control
`signal, which is received by a receiver in the controlled
`AV device. The received remote control signal is de
`20
`coded to control the AV device as intended by the
`remote control signal.
`There has recently been developed a voice-operated
`remote control system which employs voice control
`commands instead of control commands entered
`through keys. The voice-operated remote control sys
`25
`tem has a microphone mounted on a transmitter for
`converting a voice command into an electric voice
`signal, and a speech recognition LSI (Large Scale Inte
`gration) circuit for generating a remote control signal
`which corresponds to a voice pattern represented by
`30
`the voice signal. The remote control signal thus gener
`ated is transmitted to a receiver in a controlled AV
`device.
`In conventional voice-operated remote control sys
`tem, standard pattern data corresponding to voice com
`35
`mands given by the operator are registered in advance.
`When a voice command is applied for remote control,
`the input voice command is recognized on the basis of
`the registered standard pattern data.
`The voice sound uttered by the operator may not
`necessarily be the same at all times. Features of the
`voice sound of the operator may sometimes vary from
`those at the time the operator's standard pattern data
`were registered. If the voice sound features vary, then
`pattern data of an input voice command do not coincide
`with the standard pattern data, and the speech recogni
`tion rate is lowered.
`In order to avoid the above drawback, it is necessary
`to register the standard pattern data again at a suitable
`time. However, the process of registering the standard
`pattern data again is tedious and time-consuming, and
`makes the remote control system inconvenient to use.
`SUMMARY OF THE INVENTION
`It is an object of the present invention to provide a
`voice-operated remote control system which has an
`increased speech recognition rate even if voice sound
`features of the operator vary with time.
`According to a first aspect of the present invention,
`there is provided a voice-operated remote control sys
`tem comprising a microphone for converting a voice
`command into an electric signal, speech recognition
`means for converting the electric signal into pattern
`data and comparing the pattern data with registered
`standard pattern data to produce command data corre
`65
`sponding to the voice command, and transmitting
`means for generating and transmitting a remote control
`signal based on the command data, the speech recogni
`
`5,199,080
`2
`tion means comprising a plurality of standard pattern
`storage areas for storing a plurality of different standard
`pattern data for each of command data, and means for
`recognizing the voice command based on the plurality
`of different standard pattern data.
`When a voice command is entered through the mi
`crophone, the speech recognition means converts the
`contents of the voice command into pattern data and
`compares the pattern data with a plurality of different
`standard pattern data which have been registered in the
`standard pattern data storage areas. Then, the speech
`recognition means applies a remote control instruction
`signal to the transmitting means to produce a remote
`control signal.
`According to a second aspect of the present inven
`tion, there is provided a voice-operated remote control
`system comprising a microphone for converting a voice
`command into an electric signal, speech recognition
`means for converting the electric signal into pattern
`data and comparing the pattern data with registered
`standard pattern data to produce command data corre
`sponding to the voice command, and transmitting
`means for generating and transmitting a remote control
`signal based on the command data, the speech recogni
`tion means comprising a plurality of parallel speech
`recognition processors for independently recognizing
`the voice command based on respective standard pat
`tern data and outputting respective recognition results.
`When a voice command is entered through the mi
`crophone, the voice command is recognized by the
`speech recognition processors independently based on
`their standard pattern data, and the speech recognition
`processors output their respective recognition results.
`Based on the recognition results of the speech recogni
`tion processors, the speech recognition means applies a
`remote control instruction signal to the transmitting
`means to produce a remote control signal.
`The above and other objects, features and advantages
`of the present invention will become more apparent
`from the following description when taken in conjunc
`tion with the accompanying drawings in which pre
`ferred embodiments of the present invention are shown
`by way of illustrative examples.
`BRIEF DESCRIPTION OF THE DRAWENGS
`FIG. 1 is a block diagram of a general remote control
`system;
`FIG. 2 is a diagram showing a remote control signal
`by way of example;
`FIG. 3 is a block diagram of the transmitter of a
`general voice-operated remote control system;
`FIG. 4 is a perspective view of the transmitter of a
`voice-operated remote control system according to a
`first embodiment of the present invention;
`FIG. 5 is a block diagram of the transmitter of the
`voice-operated remote control system according to the
`first embodiment;
`FIG. 6 is a block diagram of a speech recognition
`circuit according to the first embodiment;
`FIG. 7 is a detailed block diagram of the speech rec
`ognition circuit according to the first embodiment;
`FIG. 8(a) is a diagram showing an analog processor;
`FIGS. 8(b) through 8(f) are diagrams showing the
`waveforms of signals in the analog processor shown in
`FIG. 8(a), respectively;
`FIG. 9 is a flowchart of an operation sequence of the
`transmitter according to the first embodiment;
`
`45
`
`50
`
`55
`
`Page 12
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`10
`
`5,199,080
`4.
`3
`transmitting circuit 17. In response to the supplied re
`FIG. 10 is a block diagram of the transmitter of a
`voice-operated remote control system according to a
`mote control instruction signal SR, the transmitting
`circuit 17 drives the infrared light-emitting diode D1 to
`second embodiment of the present invention;
`FIG. 11 is a block diagram of a speech recognition
`transmit a remote control signal RC. The controlled
`circuit according to the second embodiment; and
`device 103 is therefore remotely controlled by the re
`FIG, 12 is a flowchart of an operation sequence of the
`mote control signal RC.
`transmitter according to the second embodiment.
`First Embodiment.
`DETALED DESCRIPTION OF THE
`A voice-operated remote control system according to
`INVENTION
`a first embodiment of the present invention will now be
`General Remote Control System
`described below with reference to FIGS. 4 through 10.
`For a better understanding of the present invention, a
`External Structure
`general remote control system and a voice remote con
`As shown in FIG. 4, a transmitter 10A of the voice
`trol signal will first be described below.
`operated remote control system has a unitary casing 11
`15
`As shown in FIG. 1, a remote control system 100
`comprises a transmitter 101 for transmitting a remote
`which allows the operator to carry the transmitter
`freely around. The casing 11 supports a microphone M
`control signal from a position remote from a controlled
`on an upper panel thereof. The microphone M converts
`device 103 such as an AV device, and a receiver 102 for
`receiving the transmitted remote control signal, decod
`a voice command given by the operator into an electric
`ing the remote control signal, and sending the decoded
`20
`signal. An infrared light-emitting diode D1, for exam
`information to the controlled device 103.
`ple, is mounted in one end of the casing 11. The infrared
`FIG. 2 shows a general remote control signal. The
`light-emitting diode D1 is used to transmit a remote
`remote control signal is composed of a leader code
`control signal to the receiver of a remiotely controlled
`which informs a receiver of the transmission of data, a
`device (not shown). On one side of the casing 11, there
`25
`custom code and an inverted custom code which indi
`is disposed a voice input switch (hereinafter referred to
`cate a controlled device, a data code and an inverted
`as a "talk switch') 12 which is closed when pressed and
`data code which indicate a control command for the
`can automatically be released or opened when released.
`controlled device. The inverted custom code and the
`The talk switch 12 may be an automatic-return pushbut
`inverted data code are used to detect any error in the
`ton switch or a slide-type switch. When a voice com
`custom code and the data code, respectively.
`mand is to be entered, the talk switch 12 is closed to
`FIG. 3 schematically shows the transmitter 101 of the
`operate the transmitter 10A. Otherwise, the talk switch
`voice-operated remote control system 100. The trans
`12 is open keeping the transmitter 10A out of operation.
`mitter 101 has a microphone M for converting a voice
`. The casing 11 also supports, on its side, a mode selector
`command into an electric signal. The converted electric
`switch 13 in the form of a slide-type switch, for exam
`signal is applied to a speech recognition circuit 15 in the
`35
`ple. The mode selector switch 13 serves to select one of
`form of a speech recognition LSI circuit or the like
`the following modes at a time. The modes include a
`which includes a microprocessor. The speech recogni
`speech registration mode in which a voice command is
`tion circuit 15 recognizes the contents of the applied
`registered in the transmitter 10A and a speech recogni
`electric signal, and produces command data corre
`tion mode in which a voice command is recognized, as
`sponding to the recognized contents. The transmitter
`described later on. The casing 11 accommodates therein
`101 also has a controller 16 comprising a microproces
`an electronic circuit of the voice-operated remote con
`sor. Based on the command data from the speech recog
`trol system according to the present invention.
`nition circuit 15, the controller 16 produces and applies
`a remote control instruction signal SR to a transmitting
`Electronic Circuit Structure
`circuit 17, which then energizes an infrared light-emit
`45
`FIG. 5 shows in block form the electronic circuit of
`ting diode D1 to transmit a remote control signal RC.
`the transmitter 10A of the voice-operated remote con
`The above components of the transmitter 101 are sup
`trol system according to the present invention. The
`plied with electric energy from a power supply circuit
`transmitter 10A has a speech recognition circuit 15A
`18.
`including a standard pattern data storage unit 5 which
`When a voice command is received through the mi
`50
`stores a plurality of different standard pattern data with
`crophone M, the speech recognition circuit 15 converts
`respect to each of the voice commands. For example,
`the voice command into pattern data. The speech rec
`the standard pattern data storage unit 5 stores standard
`ognition circuit 15 compares the voice command pat
`pattern data PA1 through PAn with respect to a voice
`tern data with a plurality of standard pattern data which
`command A and standard pattern data PB1 through
`are stored therein, determines the distance between the
`55
`PBn with respect to a voice command B. Further, thus,
`voice command data and the standard pattern data, and
`standard pattern data PM1 through PMn with respect
`outputs command data corresponding to the standard
`to a voice command M are stored therein. One voice
`pattern data whose distance which from the voice com
`command which is entered from the microphone M is
`mand pattern data is smallest. There may also be en
`recognized using a plurality of standard pattern data,
`ployed another speech recognition process in which the
`60
`and the recognized data are converted into a remote
`similarity of the compared pattern data is determined
`control signal RC. The transmitter 10A has a controller
`according to a known simple similarity method and
`command data corresponding to the standard pattern
`16 to which the talk switch 12 and the mode selector
`switch 13 are connected. The controller 16 applies a
`data which has the highest similarity are outputted. The
`remote control instruction signal SR to a transmitting
`command data thus produced are applied to the con
`65
`circuit 17 which energizes the infrared light-emitting
`troller 16.
`diode D1 to transmit a remote control signal RC to the
`The controller 16 sends a remote control signal SR
`receiver of a remotely controlled device. The speech
`corresponding to the applied command data to the
`
`30
`
`Page 13
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`O
`
`15
`
`5,199,080
`5
`6
`cessing in one frequency band (e.g., through the band
`recognition circuit 15A, the controller 16, and the trans
`mitting circuit 17 are supplied with electric energy from
`pass filter BPF3) will be described. However, similar
`a power supply circuit 18 through a power supply con
`signal processing is carried out in the other frequency
`trol circuit 14 and power supply wires.
`bands.
`As shown in FIG. 6, the speech recognition circuit
`When a voice command is applied to the microphone
`15A comprises an analog processor 21 for processing an
`M, the output electric signal from the microphone M is
`amplified to a suitable signal level by the amplifier 30,
`analog voice commarid signal which is received
`through the microphone M and for outputting the pro
`which outputs an amplified signal A (see FIG. 8(b)).
`The amplified signal A is applied to the bandpass filter
`cessed analog voice command signal as a time-division
`digital data 20, a speech recognition processor 22 for
`BPF3, which then passes only a signal B in its passband.
`recognizing the voice command based on the time-divi
`The signal B is then applied to the rectifier RCT3 (see
`sion digital data 20 from the analog processor 21, a
`FIG. 8(c)). The signal B is rectified by the rectifier
`memory 23A for storing standard pattern data for
`RCT3, and a rectified output signal C (FIG. 8(d)) from
`speech recognition, and an interface 24 for transmitting
`the rectifier RCT3 is transmitted to the low-pass filter
`signals to and receiving signals from the controller 16.
`LPF3. The low-pass filter LPF3 removes ripples which
`The memory 23A serves as the standard pattern data
`may be contained in the signal C, and produces a ripple
`storage unit 5 shown in FIG. 5, and stores a plurality of
`free output signal D (see FIG. 8(e)) which is then input
`different standard pattern data PA1 through PAn, PB1
`ted to the A/D converter ADC3. The A/D converter
`through PBn,..., PM1 through PMn with respect to
`ADC3 then converts the supplied input signal D into a
`respective voice commands.
`signal E composed of 4-bit time-division digital data
`20
`As shown in FIG. 7, the analog processor 21 gener
`(1010), (0111), (0101), (0111), (1101),..., as shown in
`ally comprises an amplifier 30 for amplifying a voice
`FIG. 8(f).
`As illustrated in FIG. 7, the speech recognition pro
`command signal transmitted from the microphone M to
`a suitable level, a filter bank 31 for dividing an amplifier
`cessor 22 comprises a system controller 40 for analyzing
`output signal into signals in different frequency bands
`and processing control commands from the controller
`25
`and rectifying and outputting the signals in these differ
`16 and also for controlling the entire operation of the
`ent frequency bands, an analog-to-digital converter
`speech recognition processor 22, and a digital processor
`41 for effecting distance calculations and controlling the
`assembly (hereinafter referred to as an "A/D converter
`assembly”) 32 for converting the output signals in the
`memory 23A.
`different frequency bands from the filter bank 31 into
`The system controller 40 comprises a CPU (Central
`30
`digital signals, and an interface 33 for transmitting sig
`Processing Unit) 42 for controlling the overall opera
`nals to and receiving signals from the speech recogni
`tion of the transmitter 1, a ROM (Read-Only Memory)
`tion processor 22.
`43 for storing a control program to be executed by the
`As shown in FIG. 8(a), the filter bank 31 comprises a
`CPU 42 for the overall operation of the transmitter 1, a
`bandpass filter assembly 35 for dividing the input voice
`RAM (Random-Access Memory) 44 for temporarily
`35
`signal into signals in a plurality of frequency bands (four
`storing data, and an interface 45 for transmitting data to
`frequency bands in FIG. 8(a)), a rectifier assembly 36
`and receiving data from the analog processor 21 and the
`for rectifying output signals from the bandpass filter
`digital processor 41.
`assembly 35, and a low-pass filter assembly 37 for re
`The digital processor 41 comprises an arithmetic unit
`moving ripples from output signals from the rectifier
`46 for effecting distance calculations and identifying
`assembly 36.
`input voice commands based on the results of the dis
`The bandpass filter assembly 35 comprises a plurality
`tance calculations, a data RAM 47 for storing data nec
`of (four in FIG. 8(a)) bandpass filters BPF0 through
`essary for distance calculations, a ROM 48 for storing a
`BPF3 which have respective central frequencies fo, f1,
`program for distance calculations, a working RAM 49
`f2, f3 (fo (fl. Cf2<f3) corresponding to the frequency
`for temporarily storing processed data, an interface 50
`for transmitting data to and receiving data from the
`bands.
`The rectifier assembly 36 comprises four rectifiers
`analog processor 21 and the system controllier 40, and
`an interface 51 for transmitting data to and receiving
`RCTO through RCT3 connected in series with the
`bandpass filters BPF0 through BPF3 of the bandpass
`data from the memory 23A.
`The speech recognition processor 22 operates as fol
`filter assembly 35, respectively. The rectifiers RCTO
`50
`through RCT3 rectify the output signals from the band
`lows: When a control command is applied from the
`pass filters BPF0 through BPF3 in the respective fre
`controller 16 through the interface 24 to the speech
`quency bands.
`recognition processor 22, the system controller 40 re
`The low-pass filter assembly comprises four low-pass
`ceives the control command through the interfaces 50,
`filters LPF0 through LPF3 connected in series with the
`45 and analyzes the received control command. If the
`55
`result of analysis indicates a speech recognition process,
`rectifiers RCTO through RCT3 of the rectifier 36, re
`spectively. The low-pass filters LPF0 through LPF3
`the system controller 40 sends an instruction for speech
`recognition to the digital processor 41 through the in
`remove ripples from the rectified signals in the respec
`tive frequency bands.
`terfaces 45, 50.
`The A/D converter assembly 32 comprises four A/D
`When instructed by the system controller 40, the
`digital processor 41 introduces time-division digital data
`converters ADCO through ADC3 connected in series
`(input voice command signal) 20 from the analog pro
`with the low-pass filters LPFO through LPF3 of the
`low-pass filter assembly 37, respectively. The A/D
`cessor 21 through the interface 50 into the data RAM
`converters ADCO through ADC3 convert the analog
`47. The arithmetic unit 46 reads the standard pattern
`output signals from the low-pass filters LPF0 through
`data from the first address in the memory 23A which
`65
`LPF3 into respective digital signals.
`stores the different standard pattern data PA1 through
`Operation of the analog processor 21 will be de
`PAn, PB1 through PBn, .
`. . PM1 through PMn,
`scribed below. For the sake of brevity, only signal pro
`through the interface 51. Then, the arithmetic unit 46
`
`45
`
`Page 14
`
`AMAZON 1006
`Amazon v. SpeakWare
`IPR2019-00999
`
`

`

`5,199,080
`7
`8
`determines the logarithm of the first time-division digi
`tween the input voice command and the data PA2 is too
`tal data of a plurality of time-division digital data which
`large, then the standard pattern data item PA2 is up
`constitute one of the read standard pattern data and also
`dated or replaced with the pattern data of the input
`the logarithm of the first time-division digital data of the
`voice command.
`input voice command signal, and then determines the
`differences between the two logarithms. The arithmetic
`unit 46 further squares the differences, and adds the
`squares to determine a distance D. Therefore, the dis
`tance D is given by:
`D = i. (log(f(0) - log(f()
`
`O
`
`Overall Operation
`The transmitter 10A operates depending on whether.
`the talk switch 12 is pressed or released (i.e., turned on
`or off). If the talk switch 12 is pressed, the transmitter
`10A is capable of transmitting remote control signals,
`and if the talk switch 12 is released, the transmitter 10A
`is kept in the low power consumption mode, waiting for
`voice commands to be applied. There are two input
`modes for entering voice commands. In one input
`mode, voice commands of the operator are registered,
`and in the other input mode, voice commands of the
`operator are recognized. In the voice registration mode,
`a command word such as for "reproduction' is re
`corded in the transmitter 10A.
`Now, operation of the transmitter 10A will be de
`scribed below with reference to the flowchart of FIG.
`9. It is assumed that the talk switch 12 is not pressed and
`the transmitter 10A is in a standby condition in the low
`power consumption mode.
`First, the controller 16 initializes a registration buffer
`number to 1 in a step S1.
`Then, the controller 16 detects whether the talk
`switch 12 is pressed or not in a step S2 by detecting
`whether there is produced an operation control signal
`Sc from the talk switch 12 or not. If an operation con
`trol switch Sc is produced at this time, then it means
`that the talk switch 12 is pressed, and the controller 16
`sends a control signal Sv to the power supply control
`circuit 14. The power supply control circuit 14 supplies
`electric energy in a normal mode, enabling the transmit
`ter 10A in a step S4.
`If the talk switch 12 is not pressed, the transmitter 1
`is left in the low power consumption mode, and the
`steps S2 and S3 are repeated.
`Thereafter, the controller 16 reads the condition of
`the mode selector switch 13 to determine whether or
`not it indicates the speech registration mode for voice
`commands to generate standard pattern data, in a step
`S5.
`If the speech registration mode is indicated, control
`then goes to a step S6 in which the controller 16 outputs
`a command to instruct the speech recognition circuit
`15A to carry out a speech registration process. At the
`same time, the controller 16 sends a registration buffer
`number to the speech recognition circuit 15A in the step
`S6.
`The speech recognition circuit 15A then stores
`speech recognition standard pattern data in a corre
`sponding registration buffer in the memory 23A, i.e., a
`registration buffer having the registration buffer num
`ber=1, in a step S7.
`The controller 16 reads a status register (not shown)
`in the speech recognition circuit 15A to determine
`whether or not the registration of a voice command is
`finished in a step S8. If the registration is not yet fin
`ished, then the steps S7 and S8 are repeated until the
`registration is finished. If the registration is finished, the
`registration buffer number is incremented by 1 in a step
`S9.
`Then, the controller 16 determines whether or not
`the current registration buffer number has exceeded a
`maximum number Nimax that can be registered in a step
`
`15
`
`where
`x: the number of time divisions;
`f(t): the input voice command data (time-division
`digital data); and
`fs(t): the standard pattern data (time-division digital
`data).
`Likewise, the distances D are calculated in the same
`manner for all the standard pattern data. The smaller
`the calculated distances, the higher the probability that
`the standard pattern data are similar to the voice com
`mand. The recognition results thus obtained are col
`lected for each of voice commands. Then, command
`25
`data corresponding to the voice command to which the
`standard pattern data are most similar as a whole are
`outputted as command data from the speech recogni
`tion circuit 15A through the interface 24 to the control
`er 16.
`30
`Referring back to FIG. 5, the controller 16 is in the
`form of a microprocessor, for example. The micro
`processor of the controller 16 comprises a CPU, a
`ROM, a RAM, and an interface. The CPU executed
`arithmetic operations while referring to data stored in
`35
`the RAM, which serves as a working memory, accord
`ing to the algorithm (see FIG. 9) of a control progra