JP2003513341A - System and method for increasing recognition rate of voice input command in telecommunications terminal - Google Patents

Abstract

(57) [Summary] A method for improving the accuracy of voice dialing of a telecommunications terminal and a terminal using the method are disclosed. The analog voice input indicating the desired telephone number is converted to a digital signal. The automatic speech recognition module recognizes the number and generates an output signal indicative of the number. The determining module performs a test to determine whether one or more digits of the telephone number were not recognized by the conversion module. If the telephone number contains unrecognized digits, the search module searches the associated memory module for a telephone number having a number that matches the recognized number of the telephone number entered by the user. The user may be notified of the phone number from the matched memory either visually or by voice. If desired, the remote terminal may connect to a telephone number selected from the memory module.

Description

Detailed Description of the Invention

[0001]   background

The present invention relates to recognition of voice input in a communication device, and more particularly to a system and method for increasing the accuracy of a voice dialing system in a telecommunications terminal.

Telecommunication terminals, such as mobile phones, are ubiquitous in many modern industrial countries. Most telecommunications terminals use a keypad as an input device. However, keypads have drawbacks. First, in order to use the keypad, the user must pay attention to the communication device, even for a short time. In certain cases, such as while driving, this is not desirable. Market forces are also continually driving manufacturers to produce smaller remote telephone terminals, also called handsets. When the terminal device is downsized, a keypad error is likely to occur and the accuracy of the keypad as an input device is lowered.

Manufacturers have implemented voice input devices that accept voice input, recognize the input, and act on the input. For example, Kuniyoshi U.S. Pat. No. 4,959,850 discloses a wireless telephone device having voice recognition capability for voice dialing of a telephone. Similarly, Sakanishi U.S. Pat. No. 5,042,063 and Gerson et al. U.S. Pat. No. 4,870,686 disclose telephone devices that utilize voice recognition capabilities to enable voice dialing. Speech recognition features are described in Will U.S. Pat.No. 5,917,891, Maekawa et al. U.S. Pat.No. 5,884,257, Eting et al. U.S. Pat.No. 5,651,056, Mead
or US Pat. No. 5,638,425, Peterson US Pat. No. 5,509,049, Jakatdar US Pat. No. 5,495,553, and Hunt et al. US Pat. No. 5,303,299.

However, voice recognition is difficult, especially when noise from the surrounding environment such as automobile sounds and crowds is mixed in the voice signal. The device may not be able to recognize the user's voice if the pronunciation is insufficient or if ambient noise interferes. If applied to voice dialing, the telephone device would dial the wrong number. Alternatively, the user may be prompted to repeat all numbers or sequences of numbers that the telephone device does not recognize. Depending on the accuracy of the voice recognition system, the user must repeat the number with a high probability, making voice dialing less convenient for the user.

Therefore, there is a need for techniques to improve voice dialing systems and methods.

[0007]   Outline

The present invention solves these and other problems by providing an apparatus and method for facilitating voice dialing of telecommunications terminals, including mobile telephones. According to the invention, the remote terminal uses the information stored in the memory to enhance the accuracy of the speech recognition routine. Preferably the information is a priori information about a phone number previously placed from the remote terminal and can be matched with the phone number entered by the voice dialing method to increase the accuracy of the voice recognition system.

According to one aspect, the present invention provides a system for facilitating voice dialing of a communication device. The system includes a conversion module that receives a voice input indicating an input character string and generates a signal indicating each character of the input character string, and a determination module that determines whether the input character string includes an unrecognized character. The memory module includes a plurality of character strings corresponding to the network address, and a search module that searches the memory module for a character string including a character corresponding to the recognized character in the input character string. In use, if the conversion module is unable to convert one or more characters in the input string, the retrieval module retrieves from the memory module one or more strings having characters that match the recognized characters in the input string. You can search.

According to another aspect, the present invention provides a method for facilitating voice dialing of a communication device. The method receives a voice input indicating a desired character string, generates a signal indicating each character of the character string, determines whether or not the character string includes an unrecognized character,
If included, the method includes searching the memory module for a matching character string having a character corresponding to the recognized character in the input character string and generating a signal indicating the matching character string.

[0011]   Detailed description

Many digital wireless systems in use today use time slot access systems. User information (eg voice) is split, compressed,
It is packetized and transmitted in pre-assigned time slots. Time slots can be assigned to different users and this approach is commonly referred to as Time Division Multiple Access (TDMA). GSM (Global System for Europe)
Mobile communications) system and North American D-AMPS (Digial-Advanc
ed Mobile Phone System) system, Japanese PDC (Personal Digital Cellular)
In a time division multiple access (TDMA) communication system, such as a system, multiple remote devices can share a single radio frequency channel, thus increasing the capacity of the communication system.

Hereinafter, embodiments will be described with respect to a time division multiple access (TDMA) wireless communication system. However, the TDMA scheme is described for illustrative purposes only, and the present invention is not limited to frequency division multiple access (FDMA), TDMA, code division multiple access (CDMA).
), And / or any combination thereof, as will be appreciated by those skilled in the art.

The operation of a cellular communication system compatible with the GSM standard is described in the European Telecommunications Standards Institute (ETSI) document ETS 300 573, ETS 300 574, ETS 30.
0 578, which is here cited as a reference. Therefore, GSM
The operation of the example system is limited to a brief description here. Although the present invention is described as an example in a GSM system, one of ordinary skill in the art will appreciate that the present invention can be used in other communication systems.

FIG. 1 shows a communication system 10 in which the present invention can be implemented. System 1
0 is a hierarchical network with multiple levels for managing calls. Using a set of uplink and downlink radio frequencies, remote wireless terminals 12 operating in system 10 make calls using their assigned time slots at these frequencies. At higher hierarchical levels, mobile switching centers (
A group of MSCs) 14 routes the call from the caller to the destination. In particular,
The MSC 14 is responsible for call setup, control and disconnection. One of the MSCs 14, commonly referred to as a gateway MSC, handles communication with the public switched telephone network (PSTN) 18 or with other public and private networks.

Each MSC 14 is connected to one or more base station controllers (BSC) 16. According to the GSM standard, the BSC 16 has a CCITT No. It communicates with the MSC 14 through a reference interface known as an A-interface, which is based on the 7 signaling mobile communication application.

Each BSC 16 controls one or more wireless base station devices (BTS) 20. Each BT
S20 uses uplink and downlink radio frequencies (RF channels) to provide service in a particular area, such as one or more communication cells 21, 1
The above transceiver (TRX) (not shown) is included. BTS 20 primarily provides an RF link for transmitting and receiving data bursts to remote station 12 within each cell. In one embodiment, multiple BTSs 20 are radio base stations (RBSs).
22 included. The RBS 22 may be configured according to the RBS-2000 product series, for example. Those products are provided by Telefon Acty Boraget LM Ericsson, the assignee of the present invention. An example remote station 12 and R
For more information regarding the implementation of BS22, see Frondigh et al., US Pat. No. 5,909,46.
See issue 9.

FIG. 2 schematically shows a remote terminal 200 used in the present invention. Remote terminal 20
0 is preferably a mobile telephone used in a digital TDMA cellular communication system, eg GSM system, PDC system or D-AMPS system. However, as described above, the present invention is applicable to all types of access systems and can be easily applied to TDMA and CDMA systems, or a combination thereof. Remote terminals are widely known and already on the market. Therefore, only the aspect of the remote terminal 200 relating to the present invention will be described in detail. For additional information on remote terminals, see US Pat. No. 5,745,52 by Dent et al.
See issue 3.

In FIG. 2, the remote terminal 200 has a microphone 210 for receiving voice input from a telephone user, as a portion directly related to the present invention. The microphone 210 is connected to the conversion module 220. The conversion module 220 is an analog-to-digital (A / D) converter 2 for converting an analog voice input into a digital signal.
With 24. The conversion module 220 further includes an automatic speech recognition (ASR) module 228 for recognizing a user's voice. In addition, the remote terminal 200
A decision module 230 is included to determine whether the characters spoken by the user have been recognized by the ASR module 228 with the desired accuracy. Remote terminal 200
Further includes a memory module 250 for storing a string indicating a valid telephone number.
And a search module for searching the memory module 250. The remote terminal 200 also includes a connection module 260 for establishing a communication connection with a communication network, such as the GSM network as shown in FIG. In addition, the remote terminal 200 may include an appropriate indicator 270 (eg, L
ED or LCD display). The T28 is available from Ericsson as one of the terminals with a suitable voice recognition module.

Some or all of the modules 220-260 may include any suitable application specific integrated circuit (
ASIC), a programmed digital signal processor, or a chipset containing multiple ASICs. The modules 220-260 and other components of the remote terminal are electrically connected. For example, the determination module 230 and the search module 240 are electrically connected to the display unit 270, the speaker 280, and the connection module 260.

In addition, in the preferred embodiment, the memory module 250 and the connection module 2
The electrical connection of 60 allows the telephone number associated with the connection established by the remote terminal to be stored in the memory module 250. For example, each time the user inputs a telephone number at the remote terminal 200, the number is stored in the memory module 250. In this way, the memory module 250 maintains a list of previously called telephone numbers that can be used as a priori information to enhance the accuracy of voice dialing, as described below.

FIG. 3 shows a method of voice dialing in the embodiment of the present invention. As an outline of this method, as shown in FIG. 3, a character spoken by the user is received, the character is converted into a digital signal, and it is determined whether or not the character string is completed. If the string is not complete, the system repeatedly receives the added characters and converts them into a digital signal. After receiving the complete string, the system determines if the string contains one or more unrecognized characters. If the string does not contain unrecognized characters, it is sent to a module that allows the telephone to dial the number corresponding to the recognized string (eg, a connection module). If the string contains one or more unrecognized characters, call the search module. The search module compares the recognized number in the string with the corresponding number in the associated in-memory string, and the in-memory string appears to match the string entered by the user. Determine if If it is determined that there is a match, the string is sent to the module that allows the phone to dial the number corresponding to the recognized string. Alternatively, the string may be displayed or voiced to inform the user of the phone, in which case the user may indicate whether the string actually matches the desired string. Hereinafter, this process will be described in detail.

In one embodiment, the process shown in FIG. 3 is implemented on a telecommunications terminal, such as a mobile telephone capable of voice dialing. In step 310 of Figure 3, the voice dialing feature is enabled and the remote terminal receives a voice input indicating the first character of the string. In the United States, the letters preferably refer to a single digit in the well-known ten digit telephone number format (eg, XXX-XXX-XXXX). However, the character string may be in a format for a telephone number system of a different area, or may indicate a network address (eg, URL or IP address) of a data network in a data application. Alternatively, the character string may be a command indicating a command for a remote terminal or a memory location including a number for high speed dialing.

In step 320, the received characters are converted into a digital signal that represents the characters spoken by the user. The conversion can be performed using an analog / digital (A / D) converter in combination with a suitable ASR module. The use of many ASR modules makes it possible to carry out statistical processes for reporting a measure of the reliability of the decisions made on a character. The desired reliability rate may be programmed into the logic of the ASR module or it may be user selectable and entered into the system as a parameter. The ASR module is a known technique, and details of the ASR module do not influence the present invention.

In step 330, a test is performed to determine if the character string input is complete. For example, in the US telephone system using the 10 character format, 10
It is determined that the character string is completed by the input of the character of the number. As another example, a time-out process is used in the determination step, and it is considered that the character string is completed when a predetermined time elapses after the specific character is input. As another example, the user may positively indicate the end of the character string by pressing a designated key or speaking a designated code. One of ordinary skill in the art will recognize many ways to detect the end of an input string. If the character string is not completed, until the character string is completed, or until the user instructs to stop the voice input process, step 3
Repeat from 10 to 330.

Once the string is determined to be complete, a test is performed at step 340 to determine if the string contains one or more unrecognized characters. The "unrecognized character" here refers to a character in a character string that is not confirmed by the ASR module. In one embodiment, the system performs a test to determine if the confidence metric for one or more characters in the string is below a predetermined threshold (eg, 95% or 90%), and if so, The character string may have unrecognized characters. Additional tests may be added. For example, if the reliability metric for two characters is less than a predetermined threshold, then the string may have unrecognized characters.

If the string does not contain any unrecognized characters, the string is dialed in step 380 and the remote terminal 200 attempts to connect to the network.

If the character string includes unrecognized characters, the memory module attached to the remote terminal is searched in step 350, and the character string in the memory module is the recognized character of the character string input by the user. To see if it matches.
If there is a match in step 360, the string is retrieved from memory and presented to the user in step 370, although this need not be the case. In one embodiment, the string is visually presented to the user, such as by being displayed on an LCD or other suitable display. In another form, the character string is presented to the user by voice using a voice synthesizer. Upon receiving the approval instruction from the user, the character string is dialed in step S380.

Some or all of steps 310-380 may be performed by logic instructions running on a suitable ASIC, DSC or chipset, or general purpose processor.

The present invention has been described in detail based on a few embodiments. However, it will be apparent to those skilled in the art that various modifications can be made without departing from the present invention. Therefore, the present invention is defined only by the appended claims, and all equivalent constructions are included in the present invention.

[Brief description of drawings]

The objects, features and advantages of the present invention will become more apparent when the above detailed description is read in conjunction with the following drawings.

[Figure 1]   FIG. 1 is a block diagram showing an example of GSM communication suitable for implementing the present invention.

FIG. 2 is a flowchart showing a method for facilitating a voice call in a communication device according to an embodiment of the present invention.

[Figure 3]   FIG. 3 is a schematic diagram of a telecommunications terminal in the embodiment of the present invention.

─────────────────────────────────────────────────── ─── Continued front page    (81) Designated countries EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, I T, LU, MC, NL, PT, SE), OA (BF, BJ , CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP (GH, GM, K E, LS, MW, MZ, SD, SL, SZ, TZ, UG , ZW), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AE, AG, AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, BZ, C A, CH, CN, CR, CU, CZ, DE, DK, DM , DZ, EE, ES, FI, GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, K E, KG, KP, KR, KZ, LC, LK, LR, LS , LT, LU, LV, MA, MD, MG, MK, MN, MW, MX, MZ, NO, NZ, PL, PT, RO, R U, SD, SE, SG, SI, SK, SL, TJ, TM , TR, TT, TZ, UA, UG, UZ, VN, YU, ZA, ZW

Claims (16)

[Claims] 1. A system for facilitating voice dialing of a communication device, comprising: a conversion module that receives voice input indicating an input character string and generates a signal indicating each character of the input character string. A determination module for determining whether the input character string includes unrecognized characters, a memory module including a plurality of character strings corresponding to network addresses, and a memory module corresponding to the recognized characters in the input character string. A search module for searching a character string containing characters from the memory module, wherein the search module recognizes the input character string when the conversion module cannot convert one or more characters in the input character string. A system for retrieving from the memory module one or more character strings having characters that match the retrieved characters. 2. The system of claim 1, wherein the conversion module includes an A / D converter for digitizing the received audio input signal. 3. The system according to claim 1, wherein the conversion module includes a voice recognition module that analyzes a digital signal and generates a signal indicating a character string represented by the digital signal. 4. The conversion module generates a signal indicating a reliability level regarding the accuracy of conversion, and the determination module generates a signal indicating whether the reliability level is higher than a predetermined threshold value. The system of claim 1, wherein: 5. The system of claim 1, wherein the conversion module and the decision module are implemented in a digital signal processor. 6. The system according to claim 1, further comprising an output module for generating a signal indicating a character string in the memory. 7. The system according to claim 6, further comprising a display module for displaying the character string indicated by the signal generated by the output module. 8. The system according to claim 6, further comprising a module for audibly notifying a character string indicated by a signal generated by the output module. 9. The system according to claim 1, further comprising a connection module for making a connection with the character string indicated by the signal generated by the output module. 10. A method for facilitating a voice call in a communication device, wherein a voice input indicating a desired character string is received, a signal indicating each character of the character string is generated, and the character string is recognized. It is determined whether or not the character that does not exist is included, and if it does, a matching character string having a character corresponding to the recognized character in the input character string is searched in the memory module, and the matching character string is searched. A method comprising: producing a signal indicative. 11. The step of generating a signal indicating each character in the character string comprises:
The method of claim 1 including digitizing the received audio input signal.
The method described in 0. 12. The step of generating a signal indicating each character in the character string comprises:
12. The method according to claim 11, comprising the step of analyzing the digital signal and generating a signal indicating a character string represented by the digital signal. 13. The step of generating a signal indicating each character in the character string comprises:
The method of claim 10 including the step of generating a first signal indicative of a confidence level for the accuracy of the transformation. 14. The step of determining whether the character string includes an unrecognized character includes comparing the reliability level with a predetermined threshold value and indicating whether the reliability level is greater than a predetermined threshold value. 14. The method of claim 13 including the step of generating a signal of 15. The method of claim 10, further comprising displaying the character string represented by the signal generated by the output module. 16. The method according to claim 10, further comprising the step of audibly notifying the character string indicated by the signal generated by the output module.