Data processing: speech signal processing – linguistics – language – Speech signal processing – Application
Reexamination Certificate
1999-04-30
2002-05-07
Knepper, David D. (Department: 2645)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Application
C379S010010
Reexamination Certificate
active
06385584
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to telecommunications and particularly to voice response platforms and methods.
2. Description of the Related Technology
Voice response units (VRUs) and systems are widely used to replace or supplement human operators and attendants. VRUs provide user access to information using a conventional telephone by providing voice prompts and information. The user accesses the platform using a conventional telephone and interacts using, typically, either the touch tone pad and DTMF signaling or by speaking requests and responses. The VRU, in response, greets callers and provides instructions, prompts, menus, and information to the user using prerecorded or synthesized speech. These voice response units are extensively used by service providers, such as banks, government offices, customer service organizations, etc. to automate telephone call answering, routing and responding to inquiries. VRUs are also used to interface users with other applications, such as voice mail systems and telephone directory services.
VRUs are programmed to provide set responses to particular events and requests. Thus, a VRU may answer an incoming telephone call with a standard greeting followed by a menu or other information describing available choices and sub-menus available to the user together with the appropriate access method. For example, the user may be prompted to enter the number “1” using the telephone touch tone pad to indicate that a first option is desired, the number “2” to select a second option, etc. Alternatively, if the VRU includes a speech recognition capability, the VRU may prompt the user to speak a desired option or to input other information such as the name of a called party. As the VRU interacts with the user, it will provide information and prompting using appropriate prestored or synthesized voice messages.
As part of the scripting process, some VRUs take into account additional information in selecting an appropriate voice message. For example, some VRUs will greet a caller with a message appropriate to the time of day, day of week, etc. Certain VRUs also take into consideration information obtained about the caller to tailor an appropriate response. For example, once a caller has been identified either through the user entering identifying information such as a Social Security number or automatically through the use of automatic number identification (ANI), the VRU may tailor its scripts and messaging to that user by, for example, incorporating the user's name into the script as appropriate. Other VRUs allow the user to select the types of messages used or automatically attempt to select one of several available scripts appropriate to a particular user. For example, an inexperienced user may manually select detailed instructions for using an unfamiliar system while more experienced users may elect to receive more abbreviated prompts. Alternatively, the system may automatically select an appropriate level of detail based on user response time or the number of erroneous or incomplete responses received from the caller to prior prompts. However, while the system will adapt to a user by providing an appropriate script, the wording of each script is fixed and invariable.
While VRUs have consistently improved the speech quality of prompts and information provided to a user, present systems still fail to fully simulate the speech patterns of a human operator or attendant. By providing the same prompts over and over to a user, user frustration using the automated platform may be heightened. Still further, the user has no indication that a prior response has been acknowledged by the system when the system provides the same wording for a prompt as previously used. Further, if a user does not understand a prompt, mere repetitions of the prompt will most likely not be helpful.
Accordingly, a need exists for a more user friendly, human sounding VRU to enhance user interaction and minimize frustration. A further need exists for a VRU which reduces the monotony characteristic of conventional automated response systems.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a voice response system and method of responding to user requests with spoken responses and prompts simulating human operator speech patterns and variations.
It is another object of the invention to avoid speech qualities and characteristics attributable to mechanized attendant systems.
A still further object of the invention is to provide a system and method of providing voiced information and prompts which assists a user's understanding by providing alternative wording of and for system voice scripts.
According to an aspect of the invention, a voice response system provides variations in verbal prompts provided to a user to simulate the wording and sentence structure variations characteristic of a human operator. These variations are introduced by randomly selecting from among several content-equivalent messages or prompts in response to a pseudorandom number generated at the time each prompt or message is to be retrieved and used. Since each of the content equivalent prompts varies slightly in wording or sentence structure from one another, random wording variations are included in the prompts as would be characteristic of a human operator.
A voice response unit according to an aspect of the invention responds to spoken commands and includes a memory storing a plurality of voice messages arranged in groups of content equivalent messages, each of the messages of a group having a different wording. A speech recognition engine receives a speech input signal and, in response, provides an output signal indicative of the speech input signal. A processor responds to both the input signal for selecting one of the groups of content equivalent messages and to a pseudorandom number for selecting one of the voice messages within the selected group. A speech output device provides an audio signal corresponding to the selected voice message stored in the memory.
According to a feature of the invention, the processor further includes a pseudorandom number generator for supplying a pseudorandom number. The pseudorandom number generator calculates numeric values falling within a predetermined range, the range being divided into a plurality of subranges. Within each of the groups of voice messages, each voice message is associated with a different one of the subranges. The subranges may be substantially coextensive and equal in number to a number of voice messages within each of the groups of content equivalent messages. Further, each of the numeric values may comprise a uniform deviate within the predetermined range.
According to another feature of the invention, the voice messages are stored in a variety of formats including WAV, compressed audio and MP3 formats.
According to another feature of the invention, the voice response unit further includes an interface connected to a telephone switch which receives the speech input signal. An electronic telephone directory lists subscribers and their respective telephone numbers. A dialer is responsive to a selected one of the telephone numbers for completing a telephone call to the corresponding subscriber, wherein the processor is responsive to the output signal from the speech recognition engine for selecting one of the subscribers and an associated one of the telephone numbers from the electronic telephone directory.
According to another aspect of the invention, a method of responding to spoken commands includes a step of storing a plurality of voice messages arranged in groups of content equivalent messages, each of the messages of a group having a different wording. The method further includes the steps of receiving a speech signal, processing the speech signal to recognize a content thereof, and selecting, in response to the content, one of the groups of content equivalent messages. Further included are the steps of generating a pseudorandom number, selecting one of the
Curry James E.
McAllister Alexander I.
Fulbright & Jaworski LLP
Knepper David D.
Verizon Services Corp.
LandOfFree
Providing automated voice responses with variable user... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Providing automated voice responses with variable user..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Providing automated voice responses with variable user... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2905797