Data processing: artificial intelligence – Adaptive system
Reexamination Certificate
2000-09-07
2003-09-23
Patel, Ramesh (Department: 2121)
Data processing: artificial intelligence
Adaptive system
C706S045000, C706S012000
Reexamination Certificate
active
06625586
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an adaptive type automatic control method in which even if a trouble or variation of control system characteristics especially not predicted to a control system occurs in an automatic control system, the control system can be stabilized.
2. Description of the Related Art
In an automatic control system, a fail safe system is adopted to cope with a trouble of a control system or an erroneous operation. As a conventional fail safe system, there is known, for example, a redundancy system in which controllers, actuators, sensors, and the like are provided double, triply, or more, or a system in which internal models corresponding to control object models of predicted troubles or the like are stored in a control system, the control object model is identified, and a control rule is reconfigured by the internal model corresponding to that.
However, in the former redundancy system, since the controllers, actuators, sensors, and the like are multiply provided, there is a fear that the hardware increases, and the costs, weight, and volume increase.
On the other hand, in the latter system of reconfiguring the control rule, since the control rule is reconfigured by the internal model prepared in advance correspondingly to the identified control object model, the hardware is not increased. However, conventionally, since the set of the internal models is finite, in the case where the identified control object model exceeds the category of the set of the predicted internal models, there is a fear that it becomes difficult to reconfigure the control rule.
SUMMARY OF THE INVENTION
Therefore, the present invention has been made in view of such circumstances and has an object to provide a model optimization adaptive control method which does not cause an increase in hardware and is most suitable for a control purpose when a normal function is still used under an identified control object state even if any state change occurs.
To achieve the above object, in a model optimization adaptive control method of the present invention as shown in an embodiment, a state of a control object is detected; the detected state of the control object is made an initial value; and a simulation with a control rule, a control gain, and an instruction signal value, which control the control object, as parameters is sequentially carried out by using an internal model expressing its own control system and by a computation method to optimize a predetermined evaluation standard, so that the control rule, the control gain, and the instruction signal value to optimize the evaluation standard are obtained and the control object is controlled.
According to the present invention of the embodiment, the control rule, the control gain, and the instruction signal value to control the control object are obtained by using the internal model expressing its own control system, and by making the simulation with the detected present state of the control object as the initial value and with the computation method to optimize the previously determined evaluation standard, so that the hardware is not increased, and it becomes possible to make control most suitable for the control purpose when the normal function is still used.
REFERENCES:
Michael J. A. Berry et al; Data Mining Techniques; 1997; John Wiley & Sons; 335-349.
Ishida Shintaro
Kurotaki Mitsuru
Yokoi Hiroshi
Fuji Jukogyo Kabushiki Kaisha
Hirl Joseph P.
Patel Ramesh
Smith , Gambrell & Russell, LLP
LandOfFree
Model optimization adaptive control method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Model optimization adaptive control method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Model optimization adaptive control method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3056755