Avinash Ravichandran
 
 

Home > distances > calculatePartialMaxSVKernel.m

calculatePartialMaxSVKernel


PURPOSE

Compute the Maximum Singular value kernel between dynamical systems

SYNOPSIS

function kernel = calculatePartialMaxSVKernel(varargin)

 

DESCRIPTION

 Compute the Maximum Singular value kernel between dynamical systems

 INPUTS
 Takes variable inputs of the type:
   option 1 -  (system1, system2) - two system parameters
   option 2 -  (system1, system2, metricParams) - two system parameters
                and custom values for metric properties
   option 3 -  (M,sigma,system,metricParams) - all-pair solution matrix for
               corresponding sylvester equations, inital state covariance
               estimates, all system parmaeters and metric parameters
               metricParams  -   parameters for computing the Binet-Cauchy kernels
                   .Lambda - Dampening factor for the Binet-Cauchy kernels [Default: 0.9]
                   .Cov_U  - Weight factor for noise covariance [Default: 1]
                   .Alpha  - Weighting Heuristic between two parts of the Binet-Cauchy
                             kernels [Default: 0.5 for no heuristic weighting] 

 OUTPUTS
   kernel  -    kernel between 2 systems or all-pair Partial determinant
   kernel matrix [N x N]

 EXAMPLE

  kernel = calculatePartialMaxSVKernel(sysParamsCellArray);

% Written by : Rizwan Chaudhry and Avinash Ravichandran
% $DATE      : 27-Jul-2011 12:05:37 $
% $REVISION  : 1.00 $

 

CROSS-REFERENCE INFORMATION

This function calls:

  • preprocessSystems This is a helper function for the distance calulation functions. This allows

This function is called by:

  • calculateMetricLDS This is the main distance computing function. Use either default metric