Method and apparatus for adaptive gravity search
AssigneeElectronics and Telecommunications Research Institute
InventorsByung-Jae Kwak, Bum-Soo Park
Life of this patent
Estimating the term…
Abstract
Provided are a gravity search adaptive apparatus and method. The gravity search adaptive apparatus includes: a system configured to process and output a received signal; a system modeling unit configured to receive the same signal as a signal input to the system, to convert the signal using a system modeling parameter, and to output the converted signal; and an adaptive controller configured to use a gravity search adaptive algorithm to detect the system modeling parameter so that an error signal which is a difference between an output signal of the system and an output signal of the system modeling unit converges on a minimum value when the system modeling parameter is applied to the system modeling unit.
Claims17
A gravity search adaptive apparatus for modeling a system, the apparatus comprising:
- a system modeling circuit configured to receive an input signal input to the system, to convert the input signal using a model parameter vector, and to output the converted signal;
- and
- an adaptive controller configured to calculate, using a gravity search adaptive algorithm for solving a problem of finding a global minimum of a cost function which is in the form of a quadratic function, the model parameter vector when an error signal, which is a difference between an output signal of the system and an output signal of the system modeling circuit, converges on a minimum power value, and to apply the calculated model parameter vector to the system modeling circuit,
- wherein the adaptive controller calculates a power value ε of the error signal, the power value ε being expressed as equation 104 below:
- ε=σy2−2pw[k]+σx2w[k]2 (104),
- where w[k] is the model parameter vector resulting from k-th iteration of the calculation of the model parameter vector, p=E{x[k]y[k]}, σx2=E{x2[k]}, x[k] is the input signal, y[k] is the output signal of the system, σy2 is E{y2[k]}, and E{•} represents an ensemble average.
A method comprising:
- calculating, by an adaptive controller using a gravity search adaptive algorithm for solving a problem of finding a global minimum of a cost function which is in the form of a quadratic function, a model parameter vector of a system modeling circuit corresponding to a minimum power value of an error signal which is a difference between an output signal of a system and an output signal of the system modeling circuit when an input signal is input to the system and the system modeling circuit;
- and
- applying, by the adaptive controller, the model parameter vector to the system modeling circuit,
- wherein the calculating further comprises:
- calculating a power value ε of the error signal that is expressed as follows:
- ε=σy2−2pw[k]+σx2w[k]2,
- where w[k] is the model parameter vector resulting from k-th iteration of the calculation of the model parameter vector, p is E{x[k]y[k]}, σx2 is E{x2[k]}, x[k] is the input signal, y[k] is the output signal of the system, σy2 is E{y2[k]}, and E{•} represents an ensemble average.
Dependencies are inferred from the claim text.