Data processing: generic control systems or specific application – Specific application – apparatus or process – Robot control
Reexamination Certificate
2002-09-24
2004-04-06
Cuchlinski, Jr., William A. (Department: 3661)
Data processing: generic control systems or specific application
Specific application, apparatus or process
Robot control
C700S250000, C700S253000, C700S255000, C700S257000, C700S258000, C700S259000, C318S568110, C701S023000, C701S214000, C701S301000, C180S170000
Reexamination Certificate
active
06718232
ABSTRACT:
TECHNICAL FIELD
This invention relates to a robot apparatus and a method for controlling its behavior.
BACKGROUND ART
In recent years, a robot apparatus formed after the appearance of an animal, such as a dog, is being presented to the market. This robot apparatus acts like an animal, responsive to the information from outside or to its own internal state, such as its feeling state. Among these robot apparatus, there is such a one which makes such behavior as kicking balls, while also there is such a one which is provided with the learning functions. Among these learning functions, there is a language learning function.
One ultimate goals in a pet type robot apparatus is to create a life-like robot apparatus or, in other words, to construct an open-ended system.
Heretofore, it has been attempted to approach to the goal of constructing a life-like robot apparatus by increasing apparent complexity of the robot apparatus. This also takes account of elements such as learning or growth.
However, in mounting these functions, the conventional practice has been to change the probability of a probabilistic state machine, fixedly set for behavior generation, or to exchange the state machine in its entirety.
Although the frequency of occurrence of a behavior is changed by interactions with the user or with the outer world to increase the complexity proper to the robot apparatus, it cannot be said that such complexity is beyond the extent of the behavior or the reaction initially presupposed by a designer (programmer). Similarly, the objects that can be recognized by the robot apparatus are limited such that the robot apparatus cannot recognize more than is set by the programmer. This indicates that the current robot apparatus lacks in the ability to determine how to act against unknown stimuli.
The following is what is necessary in creating an open-ended system which may be beyond the extent of the designer's setting:
(1) the ability to recognize the unknown stimuli;
(2) the ability to create a new behavior; and
(3) the ability to select an appropriate behavior against unknown stimuli.
If the fact that the robot apparatus in question is a pet type robot is taken into account, the ability for a pet-type robot to have interactions with the human being is of particular importance. As a matter of fact, in learning various unknown things, the pet-type robot learns them through interaction with the human being.
What is most important in the interactions with the human being is the communication through a language. As for the ability in recognizing the unknown stimuli (1), optimum categorization, acquisition of the names as symbols and acquisition of the names of the behaviors are thought to be the first steps. This is a research field termed the language acquisition. As pointed out, it is crucial that these symbols be physically correlated or grounded.
This is pointed out by, for example, a report by Kaplan et al. (Kaplan, F. Talking AIBO: First experimentation of verbal interactions with an autonomous four-legged robot, in proceedings of the CELE-Twente workshop on interacting agents, October 2000, referred to below as reference material 1), a report by Roy et al. (Roy, D. and Pentland A. Learning words from natural audio-visual input, in proceedings of International Conference on Spoken language Processing, 1998, referred to below as reference material 2) and by a report by Steels et al. (Steels, L. Perceptually Grounded Meaning Creation, In Proceedings of the International Conference on Multi-Agent Systems, 1996, referred to below as reference material 3).
The behavior acquisition of (2) above may be exemplified by that through imitation, that by reinforced learning and that by evolutionary computing.
This is pointed out by, for example, a report by Damasio (Damasio, A. Descartes' Error: Emotion, and the Human Brain, Putman Publishing Group, 1994, referred to below as reference material 4), a report by Mataric (Mataric, M., Sensory-motor primitives as a basis for imitation: Linking perception to action and biology to robotics, Imitation in Animals and Artifacts, C. Nehniv and K. Dautenhalm (eds), The MIT Press, 2000, referred to below as reference material 5).
As for the appropriate behaviors against unknown stimuli (3), only extremely primitive behaviors in the real world, or a few correlated behaviors in the virtual world, have been reported.
What the above ability (3) connotes is how to acquire the meaning an object has on the robot apparatus, for example, whether it is food, a plaything or what is fearful. It becomes necessary in this connection not only to physically correlate or ground the objects for recognition but also to recognize how the object for recognition affects the internal state of the robot apparatus, such as, for example, the primary emotion or the secondary emotion.
DISCLOSURE OF THE INVENTION
In view of the above-described status of the art, it is an object of the present invention to provide a robot apparatus which is made more life-like, and a method for controlling the behavior of the robot apparatus.
For solving the above problem, a robot apparatus according to the present invention includes behavior control means for causing the robot apparatus to execute an information acquisition behavior as a behavior belonging to the autonomous behavior. The robot apparatus, thus constructed, manifests an information acquisition behavior as a behavior belonging to the autonomous behavior.
For solving the above problem, a behavior control method for a robot apparatus also causes the robot apparatus to execute an information acquisition behavior as a behavior belonging to the autonomous behavior. By this behavior control method, the robot apparatus manifests an information acquisition behavior as a behavior belonging to the autonomous behavior.
For solving the above problem, a robot apparatus according to the present invention also includes meaning acquisition means for acquiring the meaning of an object. The robot apparatus, thus constructed, acquires the meaning of the object.
For solving the above problem, a behavior control method for a robot apparatus according to the present invention also acquires changes in internal states of the robot apparatus that are produced when the robot apparatus acting based on the internal states has taken an action on the object as the meaning of the object.
With the behavior control method for the robot apparatus, the robot apparatus acts based on its internal states and acquires the changes in the internal states, that are produced when the robot apparatus acting based on the internal states has taken an action on the object, as the meaning of the object.
For solving the above object, the robot apparatus also includes speech inputting means, and a plurality of models characterizing a plurality of word sequences, in which the models have been classified based on characteristic values of the word sequences prevailing when the word sequences have been uttered. The robot apparatus further includes speech input evaluating means for evaluating speech inputs, made by the speech inputting means, in accordance with the word sequence characterizing models, and word sequence specifying means for specifying the word sequences of the speech input based on evaluation values of the speech input evaluating means.
With the robot apparatus, thus constructed, the speech input made by the speech inputting means is evaluated by speech input evaluating means, based on the word sequence characterizing models, classified on the basis of the characteristic values of the word sequence uttered, and the word sequence of the speech input is specified by word sequence specifying means, based on the evaluated value of the speech input evaluating means. This enables the robot apparatus to specify the input speech as optimum word sequence.
For solving the above problem, the behavior control method for the robot apparatus according to the present invention also includes a speech inputting step, a speech input evaluating step of evaluating speech inputs, made at t
Costa Gabriel
Fujita Masahiro
Horinaka Rika
Minamino Katsuki
Shimomura Hideki
Cuchlinski Jr. William A.
Frommer William S.
Frommer & Lawrence & Haug LLP
Kessler Gordon
Marc McDieunel
LandOfFree
Robot device and behavior control method for robot device does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Robot device and behavior control method for robot device, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Robot device and behavior control method for robot device will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3228042