| 摘要 |
Controlled variables are inputted from sensors (ST11), it is judged whether or not the difference of every controlled variable is within an allowable range (ST12), it is judged whether the position in the controlled variable difference space is the success experience which has been already learned at this time point (ST13) if the judgment is "no" (ST13), the control is changed to a learning control in which the control input of the success experience is used if the experience has been gained (ST14), the control input is determined by adaptive control if the experience has not been gained (ST15), an operation is conducted according to the control input of the success experience or a control input calculated (ST16), it is judged again from the input of the controlled variable whether or not the difference is within the allowable range (ST12), and when the judgment is "yes", the controlled variable patterns at the time points from the initial difference to a test operation and all the measures for the amounts of change of the control inputs in the process from there to the time point when the judgment is "yes" about the controlled variable patterns are stored as success experience only when the control is changed to adaptive control (ST17).
|
