Adaptive adjustment method for key performance indicators of data-driven system
A key performance index and self-adaptive adjustment technology, applied in the field of tracking control, can solve the problems of poor tracking control performance and high tracking error, and achieve the effect of high versatility, simple method process and good tracking control performance
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0022] Specific implementation mode one: combine figure 1 To describe this embodiment, a data-driven self-adaptive adjustment method for system key performance indicators provided in this embodiment specifically includes the following steps:
[0023] Step 1, set the system input as a random square wave signal to collect data, and carry out the initialization of the method:
[0024] Step A1, set the time window length N, collect the input and output data of the past time window length of the controlled object under the random square wave excitation signal, and construct the input matrix U up to the last time N (k-1) and the output matrix Y at the last moment N (k-1), and calculate the generalized inverse of the input matrix up to the last moment
[0025] Step A2, use the improved partial least squares method to calculate the initial value M(k-1) of the regression model and then obtain the prediction model
[0026] Step A3. Use the expected output of the controlled object...
specific Embodiment approach 2
[0034] Specific implementation mode 2: The difference between this implementation mode and specific implementation mode 1 is that the generalized inverse of the input matrix at the last moment as described in step A1 Specifically:
[0035]
[0036] Among them, N is the length of the time window, k represents the current moment, Indicates the input matrix at the last moment; Indicates the input of the controlled object at time k, represents a real vector space of dimension; Indicates the output matrix at the last moment, Indicates the output of the controlled object at time k, Indicates an n-dimensional real number vector space; the superscript "T" indicates a transpose, and the superscript "+" indicates a generalized inverse.
[0037] Other steps and parameters are the same as those in the first embodiment.
specific Embodiment approach 3
[0038] Embodiment 3: The difference between this embodiment and Embodiment 1 is that the specific calculation process of the prediction model includes:
[0039] Compute the regression model using modified partial least squares:
[0040]
[0041] Among them, M(k) represents the regression model at time k, Represents the generalized inverse of the input matrix at time k; Indicates the output matrix at time k; Indicates the output of the controlled object at time k, Represents an n-dimensional real vector space;
[0042] Given a prediction step size of n p , to compute the predictive model:
[0043]
[0044] Among them, n represents the dimension of the input data of the controlled object, and m represents the dimension of the output data of the controlled object.
[0045] Other steps and parameters are the same as those in Embodiment 1 or 2.
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


