Data processing: artificial intelligence – Adaptive system
Reexamination Certificate
1999-09-10
2003-08-05
Follansbee, John (Department: 2121)
Data processing: artificial intelligence
Adaptive system
Reexamination Certificate
active
06604091
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method for adjusting behavior of a device, and particularly to a method for adjusting the behavior based on the device's experience.
2. Description of the Related Art
Conventionally, various controlling methods have been available for controlling a device in accordance with a user's demand.
In such controlling methods, normally, the user sets a target value at output which the user wants, and the device is controlled in such a way that the output matches the target value, while feeding the output back to a control system which compares the feedback and the target value to adjust the output. In the above, by feeding the output back to the system to adjust the output, the output of the device to be controlled can approach the target value, thereby achieving control satisfying the user's preference.
However, the above control system may not be appropriate in an area where the user can develop emotions such as companionship toward a device. Such a device includes toys, games, communication tools, search tools, and other tools and devices subjected to personal use. Further, if a target value is not determinable due to the system's complexity or lack of information, the conventional control system does not work.
Furthermore, if a device senses numerous pieces of information, it is difficult to sort them out and use them as useful information to act thereon. If the device is operated solely in accordance with the user's command, the above may not be a problem. However, for an autonomic device which acts without the user's command, it is important to obtain meaningful information.
SUMMARY OF THE INVENTION
A general objective of the present invention is to provide a control system which enables a device to autonomously modify its behavior or performance through interaction with its external environment, i.e., based on its experience. That is a self-development system. The control system can generate pseudo-emotions which are used as a parameter for controlling behavior. The control system can collect information on an object of its own interest, and can store the information and update it by itself. This autonomous behavior system is advantageous, especially when applied to robots, toys, or games. The present invention has exploited a real-time basis behavior-adjusting system.
An embodiment of the present invention is a method for adjusting behavior of a device based on the device's experience. The device comprises: (i) a sensing unit for sensing signals; (ii) an interest-generating unit programmed to generate interest-parameters; (iii) an emotion-generating unit programmed to generate emotion-parameters; and (iv) an actuating unit for actuating the device. The above method comprises the steps of: (a) selecting an object (S) of interest based on interest-parameters (I
j
) programmed into the device, wherein S=f(I
j
), said device having emotion-parameters (E
i
); (b) upon sensing the object, extracting information, if any, on variations of emotions (&Dgr;E
i−1
) from a memory under at least the index of the object, said memory storing under the index of the object, information on variations of emotions (&Dgr;E
i−1
) previously generated by the device through past interaction with the object; (c) modifying the initial emotions (E
i
) of the device with the variations of emotions (&Dgr;E
i−1
); (d) actuating the device based on the modified emotions (E
i
+&Dgr;E
i−1
); and (e) inputting in the memory, under the index of the object, variations of emotions (&Dgr;E
i
) generated by the device during interaction with the object, thereby updating the memory. Accordingly, when the device observes the same object, the device can modify or improve its behavior based on the information obtained in the past.
In the above, in an embodiment, the updated memory stores variations of emotions (&Dgr;E′
i
) under the index of the object, wherein &Dgr;E′
i
=&agr;&Dgr;E
i
+(1−&agr;) &Dgr;E
i−1
(0<&agr;<1). Accordingly, the memory needs one column for one object and can reduce its capacity. Further, the influence of new experience can be controlled to moderate the behavior changes.
In an embodiment, in step (b), if no information on the object is stored in the memory, steps (d) and (e) can be performed by regarding &Dgr;E
i−1
as zero. Further, in step (a), if no object of interest is selected because there is no record, an object sensed by the sensing unit can be selected as an object of interest. Accordingly, the device can learn various objects' characteristics and can improve its behavior.
In another embodiment, the interest-generating unit comprises a plurality of interest modules, each differently programmed to generate and output the interest-parameter, and an object of interest (S) is selected based on the interest-parameter (max I
j
) which has been assigned the highest score at the time the object is selected (S=f(max I
j
).
Preferably, the control system further comprises a working memory which temporarily pools and stores information from the sensing unit, the interest-generating unit, the emotion-generating unit, in the first-mentioned memory until the device completes its action, and which outputs information to the interest-generating unit, the actuating unit, and the first-mentioned memory. Accordingly, processing becomes efficient.
The interest-parameters and the emotion-parameters can have any labels and definitions. For example, the interest-parameters can represent “safety”, “affection”, “hunger”, and “play”, respectively, and the emotion-parameters can represent “happy”, “angry”, “surprised”, “sad”, “fearful”, and “disgust”, respectively. The definition of each label can be predetermined. The number of interest-parameters and the number of emotion-parameters are not limited. The larger the number of parameters, the more complex the behavior becomes. If an object-oriented language program, such as JAVA and C++, is used for parameter control, new labels can be automatically created, so that new behavior can be generated accordingly.
A device controlled by the above control system can behave autonomously and improve behavior based on its experience.
The present invention can be applied to a method as well as an apparatus. Further, the system need not be an integrated system, but can be composed of plural separate units. That is, by networking plural devices or by using a separable medium (CD or IC card), an individual device can be downsized without losing memory and programming capacities.
REFERENCES:
patent: 5682469 (1997-10-01), Linnett et al.
patent: 5852819 (1998-12-01), Beller
patent: 5870730 (1999-02-01), Furuya et al.
patent: 5999895 (1999-12-01), Forest
patent: 5999908 (1999-12-01), Abelow
patent: 6005549 (1999-12-01), Forest
patent: 6034652 (2000-03-01), Freiberger et al.
patent: 6085183 (2000-07-01), Horn et al.
patent: 6195651 (2001-02-01), Handel et al.
patent: 6219657 (2001-04-01), Hatayama
patent: 10-328422 (1998-12-01), None
patent: 11-126017 (1999-05-01), None
Takanori Shibata et al; Artificial Emotional Creature for Human-Machine Interaction; Jan. 1997; IEEE; 0-7803-4053;2269-2274.*
Hiroshi Kobayashi et al; Facial Interaction between Animated 3D Face Robot and Human Beings; Jan. 1997; IEEE; 0-7803-4053; 3732-3737.
Follansbee John
Hirl Joseph P.
Knobbe Martens Olson & Bear LLP
Yamaha Hatsudoki Kabushiki Kaisha
LandOfFree
Interactive artificial intelligence does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Interactive artificial intelligence, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Interactive artificial intelligence will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3092509