Method of controlling devices via speech signals, more particularly, in motorcars

• US 7,165,027 B2
• Filed: 08/22/2001
• Issued: 01/16/2007
• Est. Priority Date: 08/23/2000
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Kellner, Andreas, Fischer, Alexander
• Primary Examiner(s): Chawan; Vijay B.
• Application Number:
  US09/935,235
• Publication Number:
  US 20020065584A1
• Time in Patent Office: 1,973 Days
• Field of Search: 704/233, 704226-228, 704/243, 704/244, 704/211, 704/222, 704/256.5, 704/258, 704/256.8, 704/246, 704/234, 381/92, 381/328, 381/356, 381/380, 381/94.1, 379/406.07, 379/430, 379/433.08, 379/202.01
• US Class Current: 704/233
• 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

Abstract

The invention relates to a method of controlling function units of a motorcar or of devices (1a, 1b) installed in a motorcar, via speech signals, in which acoustic signals occurring in the motorcar, which contain noise signal portions that depend on the operating state and/or operation environment of the motorcar, and speech signal portions, as the case may be, are applied to a speech recognition system (3) and the speech recognition system (3) uses acoustic references (8) which are selected and/or adapted in dependence on detected data of the operating state and/or operation environment. The invention is not restricted to speech control in the domain of motorcars.

Claims

1. A method of controlling function units of a motorcar, or of devices (1a, 1b) installed in a motorcar, by means of speech signals, the method comprising the steps of:

receiving acoustic signals occurring in the motorcar, which contain noise signal portions and speech signal portions,

applying the received acoustic signals to a speech recognition system (3), the speech recognition system using acoustic references (8) which are selected in dependence on the operating state and/or operation environment to determine an estimate of noise signal portions, and

adapting the estimate of noise signal portions during at least one period containing a noise signal portion but no speech signal portion.

2. A method as claimed in claim 1, wherein an adaptation is provided (22, 32-1, . . . 32-n) of the selected acoustic references in dependence on the operating state and/or operation environment of the motorcar.

3. A method as claimed in claim 1, wherein acoustic basic references are combined (33) in dependence on the operating state and/or operation environment of the motorcar.

4. A method as claimed in claim 1, further comprising the step of:

determining operating state and/or operation environment of the motorcar by reading from an on-board computer (11) of the motorcar and/or by means of one or more detectors (13) installed in the motorcar.

5. A method as claimed in claim 1, wherein parts of a vocabulary (9) of the speech recognition system (3) are determined (13) that represent speech control signals that have their effect on the control of function units of the motorcar or on devices installed inside the motorcar.

6. An arrangement for controlling function units of a motorcar, or of devices (1a, 1b) installed in a motorcar by means of speech signals, the arrangement comprising:

at least one microphone (2) for converting acoustic signals occurring in the motorcar, which acoustic signals contain noise signal portions and speech signal portions, and

a speech recognition system (3) coupled to the microphone (2) for receiving the acoustic signals and at least one of a plurality of acoustic references (8) in dependence on the operating state and/or operation environment to determine an estimate of noise signal portions, and adapting said estimate of noise signal portions during at least one period containing a noise signal portion but no speech signal portion.

7. A method of controlling a device via speech signals, in which

acoustic signals, which contain noise signal portions and, speech signal portions, are applied to a speech recognition system and

the speech recognition system uses acoustic references which are selected in dependence on the operating state and/or operation environment of the device to determine an estimate of the noise signal portions, and

adapting said the estimate of the noise signal portions during at least one period containing a noise signal portion but no speech signal portion.

8. An arrangement comprising a device controllable via speech signals, wherein

acoustic signals which contain at least noise signals portions are applied to a speech recognition system and

the speech recognition system uses acoustic references which are selected in dependence on the operating state and/or operation environment of the device to determine an estimate of the noise signal portions, and

adapting said estimated noise signal portions during at least one period containing a noise signal portion but no speech signal portion.