Voice recognition based on a growth state of a robot

• US 7,313,524 B1
• Filed: 11/28/2000
• Issued: 12/25/2007
• Est. Priority Date: 11/30/1999
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Minamino, Katsuki
• Primary Examiner(s): OPSASNICK, MICHAEL N
• Application Number:
  US09/723,512
• Time in Patent Office: 2,583 Days
• Field of Search: 704/260, 704/231, 704272-270, 704/258, 704/275, 700/1, 382/103, 3185681-56825
• US Class Current: 704/270
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A voice recognition apparatus used in a robot performs voice recognition in different manners depending upon the growth state of the robot thereby allowing the robot to provide better entertainment to a user. A growth score setting unit sets growth scores to respective words described in a word dictionary stored in a dictionary memory. A matching unit performs voice recognition for the words included in the word dictionary, taking into account the growth scores defined for the respective words.

Claims

1. A speech recognition apparatus disposed in a device, comprising:

speech recognition means for recognizing speech including a dictionary in which words to be recognized in speech recognition are described;

control means for controlling said speech recognition means in accordance with a growth state of said device, wherein said growth state is comprised of a plurality of nodes corresponding to increasing maturity levels for said device; and

action decision means for determining and performing a predetermined action in accordance with the speech recognized by said speech recognition means and an occurrence probability of the predetermined action as determined by the growth state,

wherein said control means controls said speech recognition means such that the words described in said dictionary are weighted in accordance with the growth state of said device and speech recognition is performed using the weighted words,

wherein coefficients for said weighted words are controlled by the growth state, and

wherein the occurrence probability is based on data collected from distinct behavior and environmental models comprised of a plurality of nodes representing unique actions.

2. A speech recognition apparatus according to claim 1, wherein said control means changes the recognition accuracy of said speech recognition means in accordance with the growth state of said device.

3. A speech recognition apparatus according to claim 1, wherein:

said speech recognition means includes dictionary storage means for storing a plurality of dictionaries in which words to be recognized in speech recognition are described such that the words to be recognized are divided into groups and the respective groups of words are stored in different dictionaries; and

said control means controls said speech recognition means such that the words described in the respective dictionaries are weighted in accordance with the growth state of said device and speech recognition is performed using the weighted words.

4. A speech recognition apparatus according to claim 1, wherein:

said control means controls the maximum number of words allowed to be described in said dictionary, in accordance with the growth state of said device.

5. A speech recognition apparatus according to claim 1, wherein:

said speech recognition means includes a dictionary in which words to be recognized in speech recognition are described such that other words are linked to said words to be recognized; and

said control means controls said speech recognition means such that another word linked to a word, which is included in the dictionary and which is obtained as a speech recognition result, is output as a final speech recognition word depending upon the growth state of the device.

6. A speech recognition apparatus according to claim 5, wherein words to be recognized in speech recognition are described in said dictionary such that said words are linked to other acoustically or semantically similar words.

7. A speech recognition method for a speech recognition apparatus disposed in a device, comprising the steps of:

recognizing speech using a dictionary in which words to be recognized in speech recognition are described;

controlling said speech recognition step in accordance with a growth state of said device, wherein said growth state is comprised of a plurality of nodes corresponding to increasing maturity levels for said device; and

determining and performing a predetermined action in accordance with the speech recognized in said speech recognition step and an occurrence probability of the predetermined action as determined by the growth state,

wherein said control step controls said speech recognition means such that the words described in said dictionary are weighted in accordance with the growth state of said device and speech recognition is performed using the weighted words,

wherein coefficients for said weighted words are controlled by the growth state, and

wherein the occurrence probability is based on data collected from distinct behavior and environmental models comprised of a plurality of nodes representing unique actions.

8. A computer-readable medium storing a computer program capable of being executed by a computer, causing a device comprising the computer to perform speech recognition, comprising the steps of:

recognizing speech using a dictionary in which words to be recognized in speech recognition are described;

controlling said speech recognition step in accordance with a growth state of said device, wherein said growth state is comprised of a plurality of nodes corresponding to increasing maturity levels for said device; and

determining and performing a predetermined action in accordance with the speech recognized in said speech recognition step and an occurrence probability of the predetermined action as determined by the growth state,

wherein said control step controls said speech recognition means such that the words described in said dictionary are weighted in accordance with the growth state of said device and speech recognition is performed using the weighted words,

wherein coefficients for said weighted words are controlled by the growth state, and

wherein the occurrence probability is based on data collected from distinct behavior and environmental models comprised of a plurality of nodes representing unique actions.