Phonetic voice activated dialing

• US 5,991,364 A
• Filed: 03/27/1997
• Issued: 11/23/1999
• Est. Priority Date: 03/27/1997
• Status: Expired due to Term

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Bell Atlantic Network Services, Inc. (Verizon Communications Inc.)
• Inventors: Farris, Robert D., Curry, James E., McAllister, Alexander I.
• Primary Examiner(s): SMITH, CREIGHTON H
• Application Number:
  US08/828,781
• Time in Patent Office: 971 Days
• Field of Search: 379/88.03, 379/88.02, 379/88.01, 379/230, 379/229
• US Class Current: 379/88.01
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2207/12 intelligent networks
• CPC Class Codes: H04M 3/42204 Arrangements at the exchange for service or number selection by voice at the terminal H04M1/27
• CPC Class Codes: H04M 7/06 using auxiliary connections for control or supervision , e.g. where the auxiliary connection is a signalling system number 7 link

Abstract

A telephone communications system Advanced Intelligent Network (AIN) platform provides a voice activated call dialing functionality through speaker independent phoneme speech recognition having a minimum volume of storage without requiring user template training. Speaker independent phoneme recognition identifies phoneme strings of caller spoken utterances which are then compared to phoneme string representations that previously have been stored in respective caller processing records (CPRs) for those subscribers listed in the ISCP database, or stored in an equivalent peripheral database with which the ISCP can communicate. Each stored phoneme string representation is associated in the CPR with a destination telephone number that may then be extracted to route a call.

Claims

1. In a telephone switching system having a plurality of subscriber lines, a method for routing outgoing calls comprising the steps of:

sensing an off-hook condition on a subscriber line at an associated switching location and identifying the calling subscriber line;

speaker independently recognizing, at said associated switching location, a string of phonemes of which a voiced utterance from a caller on said line is comprised;

accessing, at a service control point that serves a plurality of system switches, a call processing record (CPR) associated with the identified subscriber line from a database containing a plurality of subscriber call processing records, each call processing record having at least one entry;

comparing, at said service control point, the recognized string of phonemes with phoneme strings contained in the call processing record;

transporting a destination telephone number contained in an entry of the CPR that corresponds to the recognized string of phonemes, in response to a comparison match in said comparing step, from the service control point to said switching location;

routing the call to the destination in response to receipt of the transported destination telephone number at the switching location;

obtaining from the caller a destination telephone number for the voiced utterance in response to a non-match in said comparing step; and

routing the call to the obtained destination telephone number;

wherein said step of obtaining includes receiving DTMF input from the caller; and

further including the steps of;

prompting the caller to indicate whether the voiced utterance and associated destination telephone number is to be placed in a new entry in the CPR; and

in response to a positive indication from said prompting step, adding the new entry in the CPR.

2. In a telephone system having a plurality of subscriber lines having respective telephone numbers, a method for processing a telephone call comprising the steps of:

sensing an off-hook condition on a calling subscriber line;

recognizing a voiced utterance from a caller on said line;

accessing a call processing record (CPR) corresponding to the subscriber line from a database containing a plurality of subscriber call processing records, each call processing record having at least one telephone number entry associated with a voice pattern;

routing the call to a destination having a telephone number that corresponds to an entry of the accessed CPR in response to a match thereof with the recognized utterance;

obtaining from the caller a destination telephone number for the voiced utterance in response to a non-match with the accessed CPR; and

adding a new entry in the CPR that associates the obtained destination telephone number with the recognized voiced utterance;

wherein said obtaining step comprises receiving voice input that identifies the telephone number.