Data processing: generic control systems or specific application – Generic control system – apparatus or process – Optimization or adaptive control
Reexamination Certificate
2000-06-16
2004-10-19
Knight, Anthony (Department: 2121)
Data processing: generic control systems or specific application
Generic control system, apparatus or process
Optimization or adaptive control
C700S031000, C700S044000, C700S045000
Reexamination Certificate
active
06807448
ABSTRACT:
INCORPORATION BY REFERENCE
The disclosure of Japanese Patent Application No. HEI 10-175032 filed on Jun. 22, 1999 including the specification, drawings and abstract is incorporated herein by reference in its entirety.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a weight identification method and a feedback control method. More specifically, the invention relates to a weight identification method for identifying weights that are assigned individually to a plurality of models, each of which includes a group of parameters and which are used to construct a control model representing a control system, and a feedback control method, based on a target value for controlling the control system, for a value of response of the control system controlled based on the target value, and weights assigned individually to the plurality of models, each of which includes the group of parameters and which are used to construct the control system, feedback-controlling such an instruction value for the control system that the value of response becomes equal to the target value.
2. Description of the Related Art
A related-art vehicular continuously variable transmission (CVT) has a CVT unit that includes a speed-shift hydraulic pressure controller, an input-side pulley, a metallic belt, an output-side pulley, and the like, and a controller that controls the speed-shift hydraulic pressure controller. That is, the controller performs the speed-shift control by controlling the speed-shift hydraulic pressure controller. In this CVT, it is necessary to individually adjust various characteristics. Therefore, parameters of the controller need to be set in accordance with characteristics of the CVT unit. Specifically, parameters of the controller are determined by inputting target rotation speeds to the CVT unit stepwise, and measuring the response rotation speeds. Then, in order to realize the thus-determined parameters of the controller, the CVT unit is altered.
If the transmission coefficient of the CVT unit is unknown, it is necessary to measure step response rotation speeds under various conditions in order to determine the parameters of the controller. However, since the characteristics of the CVT unit vary depending on the pulley-pressing force or the location of the point of equilibrium, the setting of optimal parameters is not easy, and requires a great amount of time. After such cumbersome setting of parameters, it is still difficult to achieve satisfactory performance over the entire operation range.
Therefore, according to the conventional art, a control system having the CVT unit and the controller is approximated by equation (1):
Transfer function:
Gr
(
s
)
=
Krwn
2
S
2
+
2
ζ
wnS
+
wn
2
(
1
)
where Kr=a constant,
&ohgr;n=the natural angular frequency,
&zgr;=the damping coefficient, and
S=the Laplace operator
In equation (1), the constant Kr, the natural angular frequency &ohgr;n and the damping coefficient &zgr; are estimated directly from the target rotation speed determined from the throttle opening extent and the actual rotation speed of the output-side pulley, and the actually measured rotation speed. Japanese Patent Application Laid-Open No. HEI 7-35210 proposes that parameters of the controller expressable by the constant Kr, the natural angular frequency &ohgr;n and the damping coefficient &zgr; be adjusted so that constant Kr, the natural angular frequency &ohgr;n and the damping coefficient &zgr; assume desirable values that are determined by the target rotation speed and the actually measured rotation speed.
Normally, a control object has a non-linear characteristic. In a slip control system, as for example, characteristics of the controller vary depending on the engine speed, the vehicle speed, the engine load, the oil temperature, aging changes, or the like. Due to the compensation for such variations of characteristics, it is difficult in the stage of designing a controller to improve the responsiveness of the controller. Furthermore, the designing of the controller requires a great amount of time.
The use of the related art, which identifies and automatically adjusts control parameters in accordance with the characteristics of a model, improves the controllability and reduces the time required for the design.
However, in the related art, the constant Kr, the natural angular frequency &ohgr;n and the damping coefficient &zgr; of the control object are directly identified. Therefore, as the order of the Laplace operator S increases depending on the range of approximation of the control object, the number of parameters that need to be estimated increases. As a result, the amount of calculation, the program size, and the memory capacity required increase, thereby causing a packaging problem. Furthermore, there arises a danger that during the process of estimating a parameter, the estimated value may be incorrect.
SUMMARY OF THE INVENTION
Accordingly, it is an object of the invention to provide a weight identification method and a feedback control method that reduce the amount of calculation by estimating weights on models having groups of parameters that contain parameters that construct a control model, instead of directly estimating the control model-constructing parameters.
To achieve the aforementioned and other objects, one aspect of the invention provides a weight identification method that identifies weights assigned to a plurality of models, each of which includes a group of parameters. The method identifies a first weight assigned individually to a plurality of models each. Based on the first weight that is assigned to one of the plurality of models, the weight identification method specifies a second weight that is assigned to at least one model that is other than the one of the plurality of models.
That is, in the invention, first and second weights that are assigned individually to the plurality of models are identified. The plurality of models are used to form a control model representing a control system, and each includes a group of parameters.
In the invention, based on the first weight assigned to one of the models, the second weight that is assigned to at least one model that is other than the one of the models is specified. For example, the second weight that is assigned to at least one model that is other than the one of the models may decrease if the first weight assigned to the one of the model is increased. For example, if the number of the models is two, the first weight &thgr;
1
of one of the two models and the second weight &thgr;
2
of the other model may have a relationship of &thgr;
1
=1−&thgr;
2
. If three models are provided, the weights &thgr;
1
, &thgr;
2
and &thgr;
3
of the three models may have a relationship of &thgr;
3
=1−(&thgr;
1
+&thgr;
2
). The number of models is not limited to two or three. Thus, the weight assigned to each model may be a value that is greater than or equal to “0” and less than or equal to “1”, so that divergence of an estimated value can be prevented.
Therefore, since the first weight assigned to one of a plurality of models specifies a second weight that is assigned to at least one model that is other than the one of the plurality of models, the weight identification method eliminates the need to specify the weights of all the models, and therefore reduces the amount of calculation required.
The weight identification method of the invention is applicable to a feedback control method.
That is, another aspect of the invention provides a feedback control method, based on a target value for controlling a control system, for determining an amount of control of the control system controlled based on the target value, and determining weights assigned individually to a plurality of models each of which includes the group of parameters and which form a control model representing the control system, and feedback-controlling such an instruction value for the control sy
Asami Shu
Hibino Ryoichi
Kono Katsumi
Ono Eiichi
Osawa Masataka
Hartman Jr. Ronald D
Knight Anthony
Toyota Jidosha & Kabushiki Kaisha
LandOfFree
Weight identification method and feedback 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 Weight identification method and feedback control method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Weight identification method and feedback control method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3264389