Selective speaker adaptation for an in-vehicle speech recognition system

• US 6,587,824 B1
• Filed: 05/04/2000
• Issued: 07/01/2003
• Est. Priority Date: 05/04/2000
• Status: Expired due to Term

• Current Assignee: Visteon Global Technologies Incorporated (Visteon Corporation)
• Original Assignee: Visteon Global Technologies Incorporated (Visteon Corporation)
• Inventors: Everhart, Charles Allen, Thompson, Scott Alan
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/565,304
• Time in Patent Office: 1,153 Days
• Field of Search: 704/275, 704/246, 704/234, 704/251, 704/256, 704/270, 704/272, 704/243, 704/244, 704/278, 704/255
• US Class Current: 704/275
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

Disclosed herein is a method of improving the recognition accuracy of an in-vehicle speech recognition system. The method of the present invention selectively adapts the system'"'"'s speech engine to a speaker'"'"'s voice characteristics using an N-best matching technique. In this method, the speech recognition system receives and processes a spoken utterance relating to a car command and having particular speaker-dependent speech characteristics so as to select a set of N-best voice commands matching the spoken utterance. Upon receiving a training mode input from the speaker, the system outputs the N-best command set to the speaker who selects the correct car command. The system then adapts the speech engine to recognize a spoken utterance having the received speech characteristics as the user-selected car command.

Claims

1. A method for speaker adaptation of selected car commands in an in-vehicle speech command system for operating vehicle accessories by voice in moving and non-moving vehicles, comprising the steps of:

(A) receiving from a speaker a spoken utterance having speaker-dependent speech characteristics corresponding to one of a set of known car commands for operating one of said vehicle accessories;

(B) processing said spoken car command utterance according to a recognition parameter;

(C) identifying an N-best set of known car commands matching said spoken car command utterance;

(D) identifying whether said spoken car command utterance presents a training mode input by evaluating the state of said vehicle accessories during which said spoken car command utterance is spoken;

(E) outputting said N-best command set to said speaker only if said training mode input is identified;

(F) receiving speaker input selecting a correct car command from said N-best command set; and

(G) adjusting said recognition parameter so that said speech command system adapts to said speaker by recognizing as said correct car command said spoken car command utterance having said speech characteristics.

2. The method of claim 1, wherein said recognition parameter is an acoustic waveform model and wherein said spoken utterance speech characteristics include a speaker-dependent acoustic waveform signature, wherein said adapting step (G) further includes correlating said waveform signature to said acoustic waveform model for said correct car command.

3. The method of claim 1, wherein said recognition parameter is a phonetic classification set and said adapting step (G) includes altering said phonetic classification set for the correct command according to said spoken utterance.

4. The method of claim 1, wherein said step (E) includes displaying said N-best command set on a display.

5. The method of claim 1, wherein said step (E) includes processing said N-best command set and outputting said N-best command set audibly via a vehicle audio system.

6. The method of claim 5, wherein said audibly outputted N-best command set includes selection identifiers, and wherein said speaker selection input is spoken utterances containing said selection identifiers.

7. The method of claim 1, wherein said identifying step (C) includes assigning a match probability weighting to each of said known car commands in said N-best command set.

8. The method of claim 1, wherein one of said car commands in said N-best command set has a highest match probability weighting, and wherein said adapting step (G) is performed only if said speaker does not select said highest match probability command as said correct car command.

9. The method of claim 1, further comprising the step of:

performing an accessory operation corresponding to said correct car command.

10. The method of claim 1, wherein said N-best command set is comprised of at least three known car commands.

11. The method of claim 1, wherein said training mode input is selected from a group consisting of a repetitive spoken car command utterance and a spoken car command utterance corresponding to a car command for activating a vehicle accessory to be in its present state.