System and method for adaptive filtering
A method in an adaptive filter system is provided. The method comprises obtaining parameters for a plurality of branches of the adaptive filter system. The method further comprises computing gradient-based information for a selected one of the plurality of branches. The method further comprises updating the parameters for the plurality of branches based on the gradient-based information for the selected branch. An adaptive filter system is also provided.
1. A method in an adaptive filter system, the method comprising:
obtaining a plurality of parameters for a plurality of branches of the adaptive filter system, each parameter of the plurality of parameters corresponding to one of the plurality of branches;
computing gradient-based information for a selected one of the plurality of branches; and
updating the plurality of parameters for the plurality of branches based on the gradient-based information for the selected branch,
wherein the updating the plurality of parameters comprises, for a parameter of the plurality of parameters that corresponds to the selected one of the plurality of branches, updating the parameter by adding the parameter to a product of the gradient-based information and a first ratio, the first ratio being larger than zero.
2. The method according to claim 1 , wherein a respective parameter of the plurality of parameters for a respective branch of the plurality of branches other than the selected branch is updated by adding the respective parameter for the respective branch to a product of the gradient-based information and a second ratio, the second ratio being larger than zero and dependent on a similarity between the respective branch and the selected branch.
3. The method according to claim 2 , wherein the similarity is based on an angle between gradient vectors for the respective branch and the selected branch.
4. The method according to claim 2 , wherein the similarity is based on hardware structure and working condition of the respective branch and the selected branch.
5. The method according to claim 2 , wherein the similarity is based on a model for describing the respective branch and the selected branch and model parameters thereof.
6. The method according to claim 1 , wherein the updating is performed by using a gradient based algorithm.
7. The method according to claim 6 , wherein the gradient based algorithm comprises any of Least Mean Square (LMS) algorithm, Normalized LMS (NLMS) algorithm, Stochastic Gradient Descent (SGD) algorithm, recursive least square (RLS) algorithm, and Simultaneous Perturbation Stochastic Approximation (SPS A) algorithm.
8. The method according to claim 1 , wherein the selected branch is selected from the plurality of branches in circular order without priority.
9. The method according to claim 1 , wherein the adaptive filter system includes a digital pre-distortion system.
10. An adaptive filter system, comprising:
one or more processors; and
a memory having stored thereon instructions which, when run on the one or more processors, cause the one or more processors to execute operations including:
obtain a plurality of parameters for a plurality of branches of the adaptive filter system, each parameter of the plurality of parameters corresponding to one of the plurality of branches;
compute gradient-based information for a selected one of the plurality of branches; and
update the plurality of parameters for the plurality of branches based on the gradient-based information for the selected branch,
wherein the updating the plurality of parameters comprises, for a parameter of the plurality of parameters that corresponds to the selected one of the plurality of branches, updating the parameter by adding the parameter to a product of the gradient-based information and a first ratio, the first ratio being larger than zero.
11. The adaptive filter system according to claim 10 , wherein a respective parameter of the plurality of parameters for a respective branch of the plurality of branches other than the selected branch is updated by adding the respective parameter for the respective branch to a product of the gradient-based information and a second ratio, the second ratio being larger than zero and depending on a similarity between the respective branch and the selected branch.
12. The adaptive filter system according to claim 11 , wherein the similarity is based on an angle between gradient vectors for the respective branch and the selected branch.
13. The adaptive filter system according to claim 11 , wherein the similarity is based on hardware structure and working condition of the respective branch and the selected branch.
14. The adaptive filter system according to claim 11 , wherein the similarity is based on a model for describing the respective branch and the selected branch and model parameters thereof.
15. The adaptive filter system according to claim 10 , wherein the updating is performed by using a gradient based algorithm.
16. The adaptive filter system according to claim 15 , wherein the gradient based algorithm comprises any of Least Mean Square (LMS) algorithm, Normalized LMS (NLMS) algorithm, Stochastic Gradient Descent (SGD) algorithm, recursive least square (RLS) algorithm, and Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm.
17. The adaptive filter system according to claim 10 , wherein the selected branch is selected from the plurality of branches in circular order without priority.
18. The adaptive filter system according to claim 10 , wherein the adaptive filter system includes a digital pre-distortion system.
19. A method in an adaptive filter system, the method comprising:
obtaining a plurality of parameters for a plurality of branches of the adaptive filter system, each parameter of the plurality of parameters corresponding to one of the plurality of branches;
computing gradient-based information for a selected one of the plurality of branches; and
updating the plurality of parameters for the plurality of branches based on the gradient-based information for the selected branch,
wherein the selected branch is selected from the plurality of branches in circular order without priority.
20. The method according to claim 19 , wherein a respective parameter of the plurality of parameters for a respective branch of the plurality of branches other than the selected branch is updated by adding the respective parameter for the respective branch to a product of the gradient-based information and a second ratio, the second ratio being larger than zero and dependent on a similarity between the respective branch and the selected branch.