WO1995001038A1

WO1995001038A1 - Paging system with voice recognition

Info

Publication number: WO1995001038A1
Application number: PCT/US1994/005723
Authority: WO
Inventors: Walter Lee Davis
Original assignee: Motorola, Inc.
Priority date: 1993-06-17
Filing date: 1994-04-23
Publication date: 1995-01-05
Also published as: TW307970B; AU6917394A

Abstract

A paging system (10) having a paging terminal (17) for receiving messages from at least one of a plurality of callers and transmitting messages to at least one of a plurality of users of selective call receivers (13) comprises an interface (16) between a telephone network (14) and the paging terminal (17), a processor (18) for processing messages received by at least one of the callers, the processor being coupled to the interface, voice recognition circuitry (20) for recognizing voice patterns in at least one of the callers, the voice recognition circuitry being coupled to the interface and coupled to the processor, and memory (22) having stored voice templates for at least one of the plurality of callers for correlating the voice patterns recognized by the voice recognition circuitry with a given stored voice template, the memory being coupled to the processor.

Description

PAGING SYSTEM WITH VOICE RECOGNITION

Technical Field

This invention relates generally to paging systems and in particular, to a paging system using voice recognition with voice templates.

Background

Paging services proliferating around the world are using technology in different phases of an evolutionary cycle that progresses from tone-only paging to numeric message paging to full alpha-numeric messaging. Each phase of this cycle provides more utility to the end user of the service, with alpha-numeric messaging providing the most benefits and features.

However, with alpha-numeric paging, a problem exists with the entry of messages into a system. Specifically, entering an alpha-numeric message into a paging system requires that the caller either enter messages via a laborious process of pushing buttons on a touch-tone phone or use a computer-like key board data entry device.

Voice recognition technology presents a potential efficient solution to the laborious problem of key entry messaging. Unfortunately, the problems associated with the current state of voice recognition art has precluded voice recognition as a viable practical solution. With literally thousands of people using a typical paging system, the use of speaker independent recognition technology to translate spoken words into alphanumeric messages transmittable over the system is practically impossible with today's technology. The present technology can not handle the large number of users and the correspondingly large number of individual accents and voice patterns. Thus, a need exists for a practical paging system solution that can best incorporate the voice recognition technology of today while overcoming the limitations of today's voice recognition technology. Summary of the Invention

A paging system having a paging terminal for receiving messages from at least one of a plurality of callers and transmitting messages to at least one of a plurality of users of selective call receivers comprise an interface between a telephone network and the paging terminal, a processor for processing messages received by at least one of the callers, the processor being coupled to the interface, voice recognition circuitry for recognizing voice patterns in at least one of the callers, the voice recognition circuitry being coupled to the interface and coupled to the processor, and memory having stored voice templates for at least one of the plurality of callers for correlating the voice patterns recognized by the voice recognition circuitry with a given stored voice template, the memory being coupled to the processor.

Brief Description of the Drawings

FIG. 1 is block diagram of a paging system with voice recognition capabilities in accordance with the present invention.

FIG. 2 is a flow chart of the operation of the paging system of FIG. 1 in accordance with the present invention. FIG. 3 is another block diagram of an alternative embodiment of a paging system in accordance with the present invention.

Detailed Description of the Preferred Embodiment

Referring to FIG. 1 , there is shown a block diagram of a paging system 10 using speech recognition and caller selectable voice templates 22 in accordance with the present invention. Preferably, a Public Switching Telephone Network (PSTN) 14 is coupled to a terminal 17 of the paging system 10. The paging terminal 17 preferably comprises of a conventional set of terminal elements including a means 16 for interfacing a telephone network with a paging terminal such as a set of telephone system interface circuits, a processing means for processing messages 18 such as a message or call processing computer and a radio frequency transmitter 26. Optionally, the system 10 includes a subscriber database 24 coupled to the processing means 18 for correlating a particular code with a particular user of a selective call receiver 13. In addition, the paging system 10 of the present invention includes voice recognition means 20 for recognizing voice pattems such as a set of voice recognition processors which are coupled to the processing means 18. Finally, the paging system 10 includes memory means 22 coupled to the processing means 18 wherein voice templates are stored for at least one of the plurality of callers. The templates 22 allow the paging system 10 to correlate the voice patterns recognized by the voice recognition means 20 with a given stored voice template. Referring to FIGs. 1 & 2, the paging system 10 operates by having a caller initiate a paging sequence (Step 32) by using a telephone 12 preferably having a touch-tone keypad 11 to enter an access number of a pager or selective call receiver 13 to be called. When the incoming call is received (step 34) and answered (step 36) by the paging terminal 17 via the Public Telephone Network 14, the system 10 preferably plays a pre¬ recorded message (step 38) that asks the caller to either identify him or herself by entering a special identification code if the caller is a registered user of the system 10, or to press a designated phone button such as zero to indicate that the caller is not a registered user. If the caller is not a registered user, i.e., one for whom voice characterization data has not been recorded and stored in the system, the call is processed using current techniques (step 42). That is, the user enters data using keystrokes entered on the telephone keyboard or a computer or other keyboard controlled device. The paging system would then typically wait for the end of the call (step 44) and subsequently wait for another incoming call (step 46).

If the caller is a registered user of the system 10 as indicated by the entry of a valid identification code (step 40), the computer controller or processing means 18 in the paging terminal 17 recalls the voice characterization data for the particular caller corresponding to the identification code entered, loads the data (step 50) into one of a number of speech recognition processors and connects the particular speech recognition processor to the appropriate incoming phone line. Preferably, a voice, prompt requests the caller to enter a voice message and depress an end message (step 50). Optionally, the message can end automatically when a predetermined amount of time passes with audible energy level registered below a predetermined threshold. The voice recognition processors process the voice input (step 52) until an "end" code is entered (step 54) or the message automatically terminates. The information the caller enters into the system via spoken words is converted into text by the voice recognition processor (step 56). The voice processor element converts the spoken word inputs into the corresponding text form by processing the input speech into spectral templates and comparing the templates of the spoken words with the templates stored in the library of stored templates that were previously entered by the caller. Subsequently, a paging address signal (identified by the code of the desired user, i.e., the receiver of the message) and message is preferably queued into a paging signal transmission (step 58) as is known in the art.

The process for recognizing spoken word inputs by comparing templates with a library of reference templates is well known in the speech recognition art, and U.S. Patent Nos. 4,400828 and 4,400788, hereby incorporated by reference, are but a few examples of the existing art. Optionally, it is envisioned that the caller would receive on-line feed¬ back from the voice recognition processors 20 to indicate if part of a message must be repeated, and that the entire message would be read back to the user for confirmation or editing before it is transmitted to the intended recipient, i.e., the pager user. Referring to FIGs. 1 and 3 a paging system in accordance with the present invention would allow frequent users of a paging system to enter text messages by simply talking into a telephone, and greatly enhance the ease and speed by which messages are entered. Frequent users of the system can enter their voice templates into the system by calling a special number that accesses the voice processor elements (20) of the terminal 17, and be prompted to recite a vocabulary of words that are frequently used in paging messages plus any other words the user expects to use. This process can be used to generate the speaker dependent word templates that can be stored in the terminal in a file associated with that specific individual. In addition, if a user encountered a problem while entering a voice message into the system, he could press a special button on the phone to engage the "learning" mode of the recognition system, and thus could build up and/or augment the stored library of templates in the course of using the system. In another aspect of the invention, the words most commonly used in paging messages could be encoded into coded binary sequences rather than in the full text representation of the word, significantly reducing the transmission air time required to send a message. This aspect of the invention is depicted in FIG. 3 where voice inputs from a telephone line 62 are received and processed by the voice recognition processors 64 in a paging system 60. The paging system 60 would further comprise a message processing computer 70 coupled between the voice recognition processors 64 and memory locations having the stored voice templates 72, the subscriber data base 74 and a message queue 76. The common word outputs from the voice recognition processors 64 are preferably stored in a temporary memory area or buffer 66 and are encoded by a common word substitution dictionary 68 into short binary coded sequences. Once again, as in system 10, the system 60 preferably queues a paging address signal and a message into a paging signal transmission. In the selective call receiver 72, these sequences are converted back using a decoder 74 into the original text representation of the word or phrase. What is claimed is:

Claims

1. A paging system having a paging terminal for receiving messages from at least one of a plurality of callers and transmitting messages to at least one of a plurality of users of selective call receivers, the paging system comprising: a means for interfacing a telephone network with the paging terminal; processing means for processing messages received by at least one of the callers, the processing means being coupled to the means for interfacing; voice recognition means for recognizing voice patterns in at least one of the callers, the voice recognition means being coupled to the means for interfacing and coupled to the processing means; memory means having stored voice templates for at least one of the plurality of callers for correlating the voice pattems recognized by the voice recognition means with a given stored voice template, the memory means being coupled to the processing means.

2. The paging system of claim 1 , wherein the system further comprises an on-line feed-back system that allows the caller into the paging system to replay and edit their message after processing by the voice recognition means but before the message is transmitted to the intended user of a selective call receiver.

3. The paging system of claim 1 , wherein the system further comprises a means for encoding words most commonly used in paging messages into coded binary sequences which are subsequently decoded by the selective call receiver.

4. The paging system of claim 1 , wherein the processing means, voice recognition means and memory means allows the caller to selectively identify their stored voice template before the caller provides a voice message to the paging system.

5. A paging terminal for receiving messages from at least one of a plurality of callers and transmitting messages to at least one of a plurality of users of selective call receivers, the paging terminal comprising: a means for interfacing a telephone network with the paging terminal; processing means for processing messages received by at least one of the callers, the processing means being coupled to the means for interfacing; voice recognition means for recognizing voice pattems in at least one of the callers, the voice recognition means being coupled to the means for interfacing and coupled to the processing means; memory means having stored voice templates for at least one of the plurality of callers for correlating the voice pattems recognized by the voice recognition means with a given stored voice template, the memory means being coupled to the processing means.

6. The paging terminal of claim 5, wherein the paging terminal further comprises a transmitter coupled to the processing means for transmitting messages to at least one of the plurality of users.

7. The paging terminal of claim 5 wherein the means for interfacing comprises a paging terminal telephone interface for coupling to a public telephone network.

8. The paging terminal of claim 5, wherein the paging terminal in a learning mode allows a user without a corresponding stored voice template to create a corresponding stored voice template and optionally augment the stored voice template in the course of using the paging terminal.

9. The paging terminal of claim 5, wherein the paging terminal in a learning mode allows a user with a corresponding voice template to augment the stored voice template in the course of using the paging terminal.

10. In a paging system having a paging terminal for receiving messages from at least one of a plurality of callers and transmitting messages to at least one of a plurality of users of selective call receivers, a method comprising the steps of: interfacing a telephone network with the paging terminal; processing messages received by at least one of the callers via a voice recognition processor coupled to a message processing computer; comparing voice pattems in at least one of the callers recognized by the voice recognition processor with voice templates stored in a memory location in the paging terminal; correlating the voice pattems recognized by the voice recognition processor with a given stored voice template in the memory location providing a correlated voice signal; transmitting the correlated voice signal to at least one of the plurality of users.