System for and method of creating and browsing a voice web

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

Technical field of invention The present invention relates to the field of information systems. More specifically, the present invention relates to the field of interactive voice response systems. [0002]

Background technology of the invention Various services are available via the telephone network. Initially, these services required a telephone operator. With the introduction of touch-tone telephones, callers could select and provide information using telephone buttons. Recent developments have allowed users to make choices and provide information in natural language. In general, such an interface makes it much easier for users to gain access to such services. An example of a technique for implementing such a speech system is US Patent Application No. 09 / 039,203, entitled "System Construction and Speech Processing Methods for Speech Processing," filed March 31, 1998. , US Patent Application No. 09 / 105,837, entitled "How to Analyze Dialogs with Natural Language Speech Recognition Systems," filed January 26, 1998, and filed July 29, 1998. , Found in US Provisional Patent Application No. 60 / 091,047 entitled "Methods and Devices for Processing and Translating Natural Language with Speech Activation Applications". These three patent documents are incorporated herein by reference in their entirety. [0003]

With the advent of natural language recognition systems, users can answer interactive telephone systems with more natural alternating responses. Such systems are used for a variety of applications and are known as interactive voice response (IVR) systems. One known example is to provide information and services regarding flight availability, flight times, flight reservations, and the like for predetermined airlines. Another well-known use of such a system involves obtaining information about stocks, bonds, and other securities, buying and selling such securities, and obtaining information about the user's stock account. Similarly, systems exist to control transactions with customers in banks. Other applications are also available. [0004]

While such systems are used to provide dramatic advances through other voice information and voice service systems, there are still obstacles. Each such system accessed by the user requires the user to make separate calls. Often, the information is on the relevant topic. For example, if a user communicates with a voice service to obtain airline information or travel tickets, they may also want to book a hotel room or dinner in the destination city. Even if the hotel is located in the destination city that provides a voice system for room rates and availability information and the caller can book the room automatically or manually, the user will call while making the airline reservation. You must somehow find the phone number of the hotel in your destination city and then dial the desired number. This procedure is at best cumbersome. The procedure can be dangerous when done from a car during commuting hours. [0005]

Other automatic information and service systems are also available. The World Wide Web (also known as the "Internet", referred to below) is a rapidly expanding network of computers that provides users with many services and a wealth of information. Unlike the audio systems described above, the Internet is primarily a vision-based system that allows users to interact realistically with images and continuous images on display screens. [0006]

The Internet was originally created as a non-profit Koichi to provide communication links between government agencies as well as higher education facilities. Today, the Internet has transformed into a universal network of computers, including private businesses, as well as government agencies. The Internet has become accessible to many people from computers located in homes, offices, or public libraries. People can find up-to-date information on weather, stock prices, news and many other topics. In addition, people can find a wide variety of information about products and services. [0007]

The Internet offers many advantages through other media. The Internet smoothly links information stored on geographically distant servers at the same time. Therefore, the user can smoothly access the information stored in the geographically distant server. Similarly, information on the server can be updated remotely from any geographic point that accesses the Internet. [0008]

When a user accesses information on a server, the user interfaces with the server through a website. Many websites provide hyperlinks to other websites that are easy to use the Internet. When the current website has a hyperlink to another website, the user can jump directly from the current website to this other website without entering the address of this other website. it can. In use, hyperlinks are a visually recognizable notation. The user activates the hyperlink by "clicking" on an icon called hyperlink notation or point and click. The user's computer is programmed to automatically access the website identified by the hyperlink as a result of the user's point-and-click operation. [0009]

Unfortunately, Internet technology is not easily applicable to voice systems. In a visual internet system, a realistic image remains on the display screen until modified by the user. This gives the user ample opportunity to carefully read all the images on the display screen for as much time as desired to make the right point-and-click selection. In voice systems, once a message is spoken, it cannot be easily reviewed by the user. Therefore, voice systems do not have long-known similar operations for point-and-click. In addition, hyperlinking is not available in voice systems. Telephone calls are made through the central office on a call-by-call basis. In contrast, on the Internet, once connected, the computer is functionally connected to all Internet addresses at the same time. Different sites are accessed by requesting information found at different addresses. At least these differences make ordinary Internet technology inapplicable to voice systems. What is needed is a system for browsing voice networks. [0010]

The PSTN (Public Switched Telephone Network) is an individual "stations" of more than 800 million to make any two-way connection by one party (caller) dialing the phone number of another party (receiver). Provide means for. The station can be anyone by telephone, IVR system or especially information services. The current approach has two disadvantages. First, the sender must be aware of the recipient's existence. There is no easy way to browse or find information or recipients that are important to the caller and sell. Second, the caller must know the recipient's phone number. Moreover, there is no convenient way to browse web pages that can or cannot play audio from the phone. In addition, there is no integration between the PSTN and the World Wide Web, which allows you to browse both smoothly as an integrated web. [0011]

Outline of the invention The present invention is a system and method that allows a user to browse audibly and interactively through a network of audio information. The system and method preferably include a browser controller that allows the user to receive audio information and convey verbal instructions. The browser controller preferably links the user to the telephone office, voice-enabled worldwide web pages, and regular worldwide web pages in response to voice instructions. In a link to an audio page, telephone office or the World Wide Web, some information is processed, along with an audio display of the link capacity of that information. For example, information can be separated using earcons. Earcons consist of special sounds played before or after the link text. Displaying other audio than the earphones, such as speaking a link with a voice different from the main text, setting a link text away from the surrounding (salounding) text with a pause, or linking a mixed background sound. Can be used, such as playing with. If the user repeats the information initiated by earcon or other means, the voice page conveys the phone number or URL of the selected link to the browser controller. The browser controller establishes a new link with a confirmed new phone number or worldwide web page. If the new link is successfully created, the first call will be disconnected. [0012]

The present invention originates by providing a way to browser the World Wide Web with audio, to browser all telephone networks as well, and to contact recipients by saying the name of a hyperlink. Overcome the limitation that the originator must also know the existence of the recipient and the recipient's phone number or URL. This brings the power of the World Wide Web to the telephone network. In practice, the present invention takes over the PSTN as a collection of over 800,000,000 nodes from its current situation, including means for making bidirectional connections, and makes it a highly interconnected browser. Change to a web that can be used. In addition, the present invention unifies all PSTNs with the entire World Wide Web into a network that allows a browser with one large audio. [0013]

One object of the present invention is to provide a system that enables a user to browse an audio network by listening. [0014]

The accompanying drawings exemplify embodiments of the present invention. Other embodiments are also possible and are described herein. [0015]

Detailed description of preferred embodiments The present invention allows users to solicit, navigate, search, and store information from networks of telephone stations, interactive voice response (IVR) stations, voice-enabled worldwide web pages, and regular worldwide web pages. With respect to a voice activation system intended to be possible to do. This complete set of phone numbers and URLs is called a voice page. All ordinary worldwide web pages, as well as audio pages intended to work in concert with the present invention, can take advantage of the properties of the present invention, including hyperlinks. Traditional telephone stations or currently existing IVR stations currently utilize hyperlinks, but other browsing features are still available. These audio pages will link to other audio pages to form a network. All traditional telephone stations are also part of this network. Various voice web pages together make it a phone number on PSIN or a URL on the World Wide Web to form a pseudo-network by having a browser that can connect to any voice web page. Be connected. [0016]

The present invention considers some major uses of the system that include these lessons. It is first considered that the user will produce a particular application, including audio pages or worldwide web pages that are specially configured to take advantage of the audio browsing capabilities of the present invention. In addition, the invention also includes the ability to allow access to over 800 million existing traditional telephone nodes and voice pages, as well as worldwide web pages that can be read by a text-speech converter. This is true whether or not the original recipient of the call is aware of the existence of the present invention. That way, the user can access multiple well-known phone numbers, IVRs, voice information services or worldwide web pages that were not intended to use the invention, and the user can also access the IVR voice page or with the invention in mind. You can access the designed worldwide web page. When accessing a voice page designed according to the present invention, the user is provided with the option of hyperlinking to another phone number or URL. Hyperlink performance will not be given when accessing other types of telephone numbers, including traditional telephones, IVRs or voice information services. Nevertheless, certain properties of the invention will still be available to the caller. For example, the caller can return to the start page, return to the previous page, or visit the bookmarked voice page. [0017]

According to the present invention, a user can access the personal start page by dialing the telephone number assigned to the browser controller. The browser controller can continuously connect to each desired audio page. The browser controller maintains its connection to the user as much as possible, connects to the desired voice page, and ties those phones together. This causes the browser controller to monitor the call between the user and each voice page. If the current audio page contains an audio link to another audio page chosen by the user, the browser controller will connect to the selected audio page, and then if this connection is successful. , Maintain the connection to the user all the time and disconnect the call to the current voice page. A voice-activated browser controller system thus forces the user to search for useful information or services in the collection of voice pages available. [0018]

Desirable embodiments of the present invention include a browser controller and various audio pages. The user can command the browser controller by voice, thus freeing the user's hands for other tasks such as driving a car. [0019]

Further, the present invention conveys information to the user by means of a voice source. The browser controller is also configured to contact any audio page. [0020]

FIG. 1 shows an exemplary network containing desirable embodiments of the present invention. This embodiment of the invention is not intended to be limited to any particular number or type of system. Users can access the system using any conventional 100 telephone system, including stand-alone analog telephones, digital telephones, nodes on PBXs, and so on. The system includes a browser controller 102 accessed by a user using a conventional telephone 100. It is expected that the browser controller 102 will be provided as a service to many users who need similar access to it. Corporations can also use the invention to allow customers voice access to their websites and to provide them with the ability to link to sites that they choose to link to. Thus, a typical user would access the browser controller 102 via PSIN 104. However, in certain enterprises or in a standardized environment, the browser controller 102 can be made available to the user in a facility such as a PBX, thereby making the user's traditional phone 100 PSIN 104. The need to connect to the browser controller 102 via is eliminated, and a direct connection can be made instead. In addition, the browser controller 102 can be implemented in a personal computer in hardware and / or software, eliminating the need to connect the user's traditional phone 100 to the browser controller 102 via PSIN 104, and instead. I was able to connect directly to my computer via the Internet. [0021] [0021]

Browser controller 102 includes a pointer to start page 106 for each user of the system. The start page 106 is a personal home page for the user and can work in connection with the browser controller 102. The start page could also be any audio page on the audio web. The browser controller 102 has a static grammar that assists the user with navigation and other browser functions. [0022]

In addition, the browser controller 102 generates a dynamic grammar that browses and links to each voice page visited. [0023]

Browser controller 102 also contains many dynamic grammars that are modified according to the needs of each particular user. These grammars are described in more detail below. [0024]

The originator user can use the traditional phone 100 and call another user 110's second traditional phone 108 directly via PSIN 104 in the usual way. This is nothing unusual and customary. Alternatively, using the present invention, the origin user calls the browser controller 102 using those conventional telephones 100. Once the link is established with the browser controller 102, the origin user is recognized and then instructed the browser controller 102 to call the conventional phone 108 via PSIN 104. Originating users are identified using well-known methods. The browser controller 102 calls the telephone number of the second conventional telephone 108 to establish the link. The origin user is linked to the receiving user through the browser controller 102. In this way, the browser controller 102 has two links via PSIN 104. One is for the origin user and the other is for the receiving user. [0025]

This link through the browser controller 102 allows the starting point user an advantage on a traditional phone. The browser controller 102 includes a natural speech recognition engine to "listen" to the starting user. Each origin user speaks a well-known assigned "browser wakeup" word to feed the browser controller 102. Browser Wake Up Words are, if possible, less commonly used words. Under certain circumstances, users may choose their own browser wakeup word, which is not preferred. When the browser controller 102 recognizes the browser wakeup word spoken by the origin user, the browser returns to the command mode described in more detail below. The browser controller 102 can be configured to just wait for a command after the browser wake up word, or the browser controller l02 can be configured, for example, by saying, "How can I help you?" Can respond. Depending on the nature of the command, the link to the receiving user is maintained or broken. Other calls are then made, as described below. [0026]

The origin user can use the browser controller 102 to establish other types of communication links. For example, the origin user may want to receive audio information such as time or weather. Other types of voice information are also available. As is well known, the originator user can call the information service directly using the telephone number for the information service. [0027]

Alternatively, the origin user can call the browser controller 102 and instruct it to call a predetermined information service 112. When the browser controller 102 establishes a call to the information service 112, the originator listens for the desired information in the usual way. At any time, the origin user can recite the browser wake up word and disconnect the information service 112 call to have the browser controller 102 make another call. [0028]

The origin user can call it the IVR system 114, which recognizes only dtmf tones using the browser controller 102. When the browser controller 102 connects the starting point user to the IVR dtmf system 114, the user can use the keypad of a conventional telephone 100 to retrieve or supply the required information. As soon as the desired transaction or communication is completed, or whenever the user speaks the browser wakeup word, the origin user speaks the browser wakeup word and control is returned to the browser controller 102. The connection to the IVR dtmf system 114 may then be disconnected or reclaimed. [0029]

Similarly, the origin user can use the browser controller 102 to call all IVR systems 116, including the natural language speech recognition system. When the browser controller 102 connects the starting point user to the IVR speech system 116, the user can extract or supply as much information as needed using natural language. When the user speaks the browser wake up word as soon as the desired transaction or communication is completed, or at any time, the origin user states the browser wake up word and control is returned to the browser controller 102. [0030]

The connection to the IVR dtmf system 116 can then be disconnected or reclaimed. For example, a user can speak a browser wake up word to return control to the browser controller 102, but can still wish to return to the current phone link. [0031]

When control is returned to the browser controller 102, proper action is taken. As an example, a user could request that a bookmark be created for the current page. Then, as soon as you say the appropriate command, the browser controller 102 returns the user to the pending link. All of the links described above are accessed via PSIN using a traditional phone number to initiate contact. [0032]

As another example, the origin user can use the browser controller 102 to call audio-enabled and voice-enabled World Wide Web pages. The composition of such audio-enabled Worldwide Web Pages 120 will vary depending on the needs or aspirations of the developer. For example, developers could include the ability to determine if a contact originated via the Computer World Wide Web or from the PSIN and Voice Web. Developers can set up pages that include audio content in the same way as audio page 118. Any hyperlink present on a worldwide web page could be identified by audio display in the manner described here. This hyperlink could also occur on worldwide web pages that are not voice enabled. The browser controller 102 included a text speech converter that allowed PSIN to read the content of the page to the origin user contacted by the browser controller 102. The same voice instructions can be used to indicate hyperlinks on worldwide web pages. [0033]

Unlike the links described above, traditional pages on the World Wide Web are not accessed using phone numbers on PSIN. If anything, pages on the World Wide Web are accessed using internet addresses. In addition, on the Internet World Wide Web, information is generally transferred directly using traditional protocols such as http. Communication over the Internet is generally not performed using data signals exchanged between a pair of modems on PSIN. This is true despite the fact that many users access the Internet through IPS (Internet Service Providers). [0034]

Communication between the user and the ISP is performed using data signals exchanged between a pair of modems, but communication for the same information as is the case for transactions from the ISP to sites on the Internet is TCP / It is executed using an internet protocol such as IP or HTTP. [0035]

At least for this reason, the browser controller 102 described above cannot interact directly with traditional pages on the World Wide Web without a direct internet connection. To overcome this obstacle, the browser controller 102 includes as much as possible a second dedicated internet connection to interface the browser controller 102 to the internet. The browser controller 102 is configured to transmit data bidirectionally between the browser controller 102 and the Internet. In addition, the browser controller 102 is also configured as a gateway for bidirectionally coupling audio-voice information between one user and PSIN and between the other user and a worldwide web page over the Internet. Alternatively, each server that is configured to interact with the browser controller 102 of the invention and provides a World Wide Web page containing voice information is configured to be accessed by telephone and modem, and even its own. Could include a gateway. [0036]

However, it is clear that such a build would require a considerable replication of equipment and software across all suitable worldwide web servers and pages. [0037]

Further alternatives would require providing PSIN access to worldwide web pages. This approach overcomes the well-known delay problem on the Internet. This approach would be even less desirable as it would solve the Internet latency problem. [0038]

For other methods of providing access to the World Wide Web page or IVR system from the browser controller 102, it may include an interface using what is called an IP telephony protocol. As is well known, IP telephony technology allows simultaneous transmission of both voice and digital data. Alternatively, parallel telephone lines and internet connections can be provided to emulate IP telephone communication technology. Yet another method utilizes XML or other similar voice / data protocols (Motorola VoxML or Microsoft's HTML extensions) to provide Internet access to PSTN applications such as Browser Controller 102. be able to. [0039]

As is clear from the above, all the features of the present invention, except for hyperlinks, can also be used to access conventional telephone communication services. It provides outgoing user access to over 800 million existing ones by using the improved features of the present invention. The greatest capabilities of the present invention can be achieved by combining with audio pages 118 specially configured to accommodate all the advantages of the present invention, including hyperlinks as defined herein. The audio page 118 can be formed on an IVR system or a worldwide web page. When the information is presented to the calling user, a given audio item is specifically associated with that user. For example, a special voice information segment is configured to tell users with the latest inventory price, "The latest transaction price for <Apple> is $ xx.xx per share. The latest transaction price for <IBM> is the share. Information is provided by stating, "Every is yy.yy dollars." The "less than" symbol ("<") represents a voice start marker, such as an earcon (as defined below), which informs the user that a custom audio link has started. Similarly, the "larger" symbol (">") represents an end-of-speech marker, which informs the user that the custom audio link has ended. Following this example, if a user wants to know more about the company's Apple, he or she can say "Apple." If the user wants to know more about the "transaction price" by verbally saying "transaction price" only in the above voice information segment, the "transaction price" does not exist and the user "transactions". Do not receive details regarding "price". Before the user reads the word "Apple", the start marker ("< The user will know that "Apple" is a valid audio link because he can hear ") and also hear the end marker after the word" Apple ". As an example, the start marker (<) can be represented as three consecutive ascending notes configured so that the pitch of each note rises in sequence. In addition, the end marker (">") can be represented as three consecutive descending notes configured so that the pitch of each note is continuously lowered. The term "earcon" is a custom glamor o Used for audible marking of Diolink. The above examples are for illustration purposes only and should not limit the scope of the invention of the present application. It will be clear to those skilled in the art that there is no reason to limit the way custom glamor audio links are marked with audible sound. For example, the text of an audio link can be spoken in separate voices, the background sound can be mixed with the audio link, and the text of the audio link can be paused. [0040]

When the browser controller 102 hears the calling user repeat the audio link, a new phone number is dialed or the World Wide Web is accessed according to the repeated audio link. If successful, the connection with the most recently accessed voice page 118 will be disconnected. The browser controller 102 knows the phone number or the URL of the World Wide Web corresponding to the repeated audio link. That is because the information was sent to the browser controller 102 by voice page 118. Sufficient bandwidth is also present in the PSTN104 so that such information is clearly communicated to the outgoing user between such devices without loss of normal conversational function or quality. [0041]

Other types of connections can be established on worldwide web pages, including plain text that has voice performance or can be read through a text-to-speech converter. Such pages are configured to provide one or both of graphic data and audio data, depending on the type of device accessing it. There, it can be shown as a hyperlink in the usual way, or as a voice link with an earcon or other voice instruction, or both. The predetermined link is available only to computer users logged on the Internet and provides only graphic information. Such a link is not presented to the originating user of the present invention using the earcon. Other links are for voice services only and provide audio information only. Such links are not provided using hypertext links on Graphic World Wide Web pages. Yet another link is available for both audio and hypertext links with respect to data providers that provide both graphics and audio information. [0042]

Links to worldwide web pages are made by the browser controller 102 via a modem over a PSTN or a well-known gateway, although gateways are clearly desirable. In any case, such a connection is utilized to provide the advantages of the present invention. [0043]

The outgoing user can perform a number of functions by using the browser controller 102. All outgoing users have a default suit of features and commands, which are available to the user when connected to their respective browser controller 102. Below is a list of such features Table 1 below. The list is exemplary and the execution of the present invention may include more or less commands. [0044]

Glamour Resident in browser / controller Static dynamic Next page Bookmark 1 Previous page Bookmark 2 Return Go home Go to the start page Bookmark n My choice Phone number 1 Help phone number 2 Where i am Add it to my bookmarks Remove it from my bookmarks Phone number n Go to my bookmarks Favorite things 1 Go to bookmarks Favorite things 2 search personal information Favorite thing n Phone number n table 1 [0045]

In addition, each outgoing user can refine a set of personal tasks for each browser controller 102 to perform what is stored on their personal start page. Such dynamic information allows the calling user to call or connect to a known service without having to remember the phone number. For example, at work, the calling user can access each browser controller 102 and say the command "weather". The browser controller 102 dials the phone to get the local weather report so that the user can hear the report. The browser controller 102 maintains its connection until it "listens" to the browser's rising word. Upon hearing the browser startup word, the browser controller 102 waits for a command. An exemplary outgoing user looks for a list in stock. The connection to the weather report will be disconnected and a new connection will be established for the service that provides stock information. The connection is maintained until the browser controller 102 hears the browser startup word again. The example calling user then commands "call mother". Browser controller 102 disconnects from the stock list and calls the desired person. The illustrated calling user hangs up the call and accesses the news report on audio page 118. During the advertisement, voice instructions announce the local <restaurant>. The example outgoing user is then <Restaurant> Say the name of. The browser controller 102 automatically connects the exemplary calling user to the restaurant and hangs up the current phone, at which time the calling user makes a lunch reservation. Transactions on all these communications occurred without the example calling user dialing except for the first call to browser controller 102. In addition, the user accessed both traditional telephones, IVRs, voice information services and voice pages with a single call to the user's browser controller 102. [0046]

There is a set of static and dynamic glamor that works on each audio page 118. Depending on the execution, speech recognition for the item in those glamor can reside as part of either the browser controller 102 or the speech page 118. Table 2 describes what those glamors do. It is clear to those skilled in the art that more or less items can be included in those glamors. [0047]

Glamour Features on audio pages Static dynamic Help dynamic link What i chose Static link Table 2 [0048]

Since items can be changed frequently, there is a dynamic grammar on the audio page. For example, on the news audio page, you can see that the news changes frequently. The news report includes audio links to other audio pages, phone numbers or audio information and links corresponding to the news report. Therefore, those links must be dynamic. Both the voice page 118 and the browser controller 102 generate a dynamic grammar link. For example, if the audio page 118 is a worldwide web page, the dynamic grammar is generated by the text of the link directed by an audio queue such as Earcon. [0049]

FIG. 2 shows a flowchart of the operation of the browser controller 102 (FIG. 1). The calling user makes a call to the browser controller 102, which identifies the caller by using the known method shown in block 200. Once the outgoing user is identified, the browser controller 102 can load the start page 106 (FIG. 1) for that outgoing user, as shown in block 202. The browser controller 102 executes a dialog (interaction) with the calling user to receive the command. For example, immediate "Hello Steve, or something you might you serve you." Is started. Functions are performed in response to interactions involving commands that depend on the calling user. For example, if the calling user describes his favorite things 206, the browser controller 102 shows a program for those favorite things. The calling user can then add 208 and remove 210 or Listing 212 to perform those functions. When the function is terminated, the browser controller 102 returns to the execution dialog block 204, and a new dialog replacement occurs between the outgoing user and the browser controller 102. [0050]

As in the other examples, the user can command "bookmark 214". The calling user can then add 216 and remove 218 or Listing 220 to perform those functions. When the function is terminated, the browser controller 102 returns to the execution dialog block 204, and a new dialog replacement occurs between the outgoing user and the browser controller 102. In another example, the calling user can provide a "go" command or request an audio link requesting that a new phone number be recorded. The browser controller 102 then enters the execution link block 222. The action allows the calling user to connect to another phone number or worldwide web page via the browser controller. When the link is complete, the browser controller 102 returns to the execute dialog block 204 via the return from link block 224. [0051]

From the execution dialog block 204, the outgoing user can instruct the browser controller 102 to play the outgoing user's start page. If the audio link is not listed, its control returns to run dialog block 204. If the audio links are listed, the execution link block 222 makes the appropriate connections. As mentioned above, the audio link can be set away from the rest of the audio page by the earcon, but there are other ways to identify the audio link. [0052]

The calling user can instruct the browser controller 102 to investigate the voice page 118 within the execute dialog block. The search request can be carried out by a suitable search engine. Finally, the dialogue ends and the call is hung up at exit block 226. [0053]

FIG. 3 shows a flowchart of the operation of the execution link block 222 (FIG. 2). A list of calls made during the operating period is maintained. This allows the calling user to return to the previous call. Once in the execution link block 222, the forward / backward list is updated in block 300 with the link information that communicated with the command to execute the link. The call is made to the linked phone number in block 302. The phone is connected to the desired phone number in block 304. Thus, during the phone call, the browser controller 102 (FIG. 1) hears either dtmf or the browser launch word in block 306. When the dtmf command is executed in block 308, the link is broken in block 310, the forward / backward list is updated in block 300, and a new call is made in block 302 as before. Instead of dtmf, as mentioned above, the phone station of the browser controller 102 and the phone station of the voice page can communicate via IP phone communication, or IP It can include a parallel internet connection that emulates telephone communications. In that case, instead of using dtmf, the recipient's phone number or World Wide Web URL can communicate via that data channel. In addition, additional information such as the state of user interaction can be communicated. If the browser launch word is heard in block 306, recognition command block 312 identifies the command to be executed in execution command block 314. If the command is not for a new link, control returns to block 306 and continues listening to dtmf or the browser launch word. If the command is for a new link, the latest link is broken in block 316, the link is broken in block 310, the forward / backward list is updated in block 300, and a new one. The call is made at block 302 as before. Alternatively, instead of making a phone call, a worldwide web page is downloaded from the Internet. [0054]

FIG. 4 shows a flowchart (flow chart) of the return operation from the link block 224 (FIG. 2). First, the phone call is cut off by disconnecting the line from link block 400. The forward / reverse list is then updated with the update forward / reverse list block 402. [0055]

FIG. 5 shows a flowchart of voice page 118 (FIG. 1) operation. Upon access, voice page 118 plays audio text or prompts with play text 500. The prompt may include a link name or a list of link names. The speech of the initiating user is recognized by recognition block 502. When the command is recognized, it is executed in the action (execution) block 502. If the action mentions a hyperlink name, the phone number of that link is dtmf transferred to browser controller 102 (Figure 1) in block 506. Instead, as shown in block 506, the link is reported to the browser controller 102 via LP telephone method or internet connection. Therefore, voice page 118 is issued to block 508 to return control to browser controller 102. If the action is a browser wake-up call, control is returned to browser controller 102 in block 510. The line is held in block 512. If the browser controller 102 returns control to voice page 118, the operation returns to play text block 500. If the browser controller 102 breaks the link, voice page 118 appears in block 514. [0056]

FIG. 6 shows a flow diagram of the recognition operation, and each stage of the browser controller 102 (FIG. 1) and the voice page 118 (FIG. 1) is translated. The memory 600 includes a dictionary memory 602, an acoustic model memory 604, its grammar, and their pronunciation memory 606. The dictionary memory 602 contains all the words in the grammar. The acoustic model memory 604 includes all statistical models of all speech units (units) that make up a word. The input signal 608 for the digitized speech is the input to the front edge analysis module 610. The front edge analysis module 610 separates the feature vector from the digitized speech, each containing a speech of a given length. In a preferred embodiment, the feature vector is the output for each 10 mS of speech signal length. The feature vector is given to the search engine 612, which compares the feature vector with the language model. The search engine 612 uses grammatical memory that defines all of the word strings (a series of words) that the starting user may say, dictionary memory defines how these words are spoken, and acoustic memory defines the words. Memorize the voice classification for the dictionary. The best guess is made for each word. [0057]

The present invention may be executed and used by a user who does not have an access account (qualification) for his / her own browser controller or the browser controller of the provider. As an example, consider airlines, rental car agencies and hotel chains that co-recognize the market. The user may call the airline to make travel arrangements to the city. Airline arrangements can be made and tickets can be purchased using an autonomous system. The automated system may include a browser controller. In such cases, the user can be prompted by the appropriate car connection method or other audio queue and then book the rental car in cooperation with the rental car agency. The airline's automation system browser controller will then automatically connect the user to the rental car agency exactly as described above. Once the car is rented, the agency's browser device then connects the user to the hotel chain to reserve a room. [0058] [0058]

As will be easily understood, the users of this example are daisy-chained to the hotel chain by both the airline browser controller and the rental car browser controller. When the user is daisy-chained, each call in the chain remains active and is billed by the telephone service provider. Therefore, it is desirable for the browser controller to work as described above, where it listens to the user's audio link iterations and at the same time sets up a new call and disconnects the previous call rather than daisy-chaining each other's calls. .. [0059]

As an example of how the invention is used by users who do not have their own browser controller 102, consider that the airlines do not want links to hotel and rental car voice pages. Even so, it is advantageous for airlines to use the present invention. The browser controller 102 can read the airline information as a voiceable worldwide web page, thereby eliminating the need for a separate IVR system with a separate database integration for the airline. If the user has his own browser controller, the airline does not need to give telephone access to its World Wide Web page. But if the user does not have his own browser controller 102, the airline can give it for them. The airline can also eliminate the need for the airline to have its own calling center for telephone access to its World Wide Web page by renting the browser controller 102 at the external calling center for hours. This is in preparation for a significant savings. Potential can still be kept to a minimum by putting airline voice data, prompts and grammar into the intelligent cache. [0060]

The present invention is described in a particular embodiment incorporating details to facilitate understanding of its configuration and principles of operation. References to such particular embodiments and their details are not intended to limit the scope of the claims attached herein. It will be apparent to those skilled in the art that modifications to the illustrated embodiments can be made without departing from the spirit and scope of the invention. For example, the browser controller 102 may be configured to first break the current link before setting the new link. [0061]

It will be apparent to those skilled in the art that the devices of the present invention are performed in several different ways and the devices disclosed above merely illustrate and by no means limit the desired embodiments of the present invention. It is clear that the various aspects of the invention can be used alone or in combination with one or more of the other aspects of the invention. Moreover, the various elements of the invention can be replaced with other elements.

[Simple explanation of drawings]

[Figure 1]

An example of a comprehensive block diagram of the present invention. [Figure 2]

The operation flowchart of the browser controller of this invention is shown. [Fig. 3]

The operation flowchart of the execution link block of FIG. 2 is shown. [Fig. 4]

The operation flowchart of the return from the link block of FIG. 2 is shown. [Fig. 5]

The operation flowchart of the voice page of this invention is shown. [Fig. 6]

A simplified block diagram of the speech recognition and translation system of the present invention is shown.