Drake
TaylorVar Class Reference

does inline autodiff overloads operators: http://www.mathworks.com/help/techdoc/matlab_oop/br02znk-1.html More...

Collaboration diagram for TaylorVar:

## Public Member Functions

function TaylorVar (f, df)

function eval (obj)

function double (obj)

function getmsspoly (obj, p_xbar)
returns the taylor approximation as an msspoly More...

function getOrder (obj)

function superiorfloat (varargin)

function isnan (obj)

function isinf (obj)

function size (obj, dim)

function length (obj)

function numel (obj, varargin)

function end (obj, k, n)
end called as part of the kth index of n indices 'help end' was the only way i found documentation on this More...

function repmat (obj, M, N)

function reshape (obj, varargin)

function bsxfun (fun, a, b)

function plus (a, b)

function minus (a, b)

function uminus (a)

function uplus (a)
intentionally empty More...

function times (a, b)

function mtimes (a, b)

function rdivide (a, b)

function ldivide (a, b)

function mrdivide (a, b)

function inv (a)
use d/dq inv(H(q)) = - inv(H(q)) [d/dq H(q)] inv(H(q)) More...

function pinv (A)

function mldivide (a, b)

function power (a, k)

function mpower (a, k)

function lt (a, b)

function gt (a, b)

function le (a, b)

function ge (a, b)

function ne (a, b)

function eq (a, b)

function sign (a)

function and (a, b)

function or (a, b)

function not (a)

function any (a)

function ctranspose (a)

function transpose (a)

function horzcat (varargin)
find the index of the first TaylorVar More...

function vertcat (varargin)
find the index of the first TaylorVar More...

function subsref (obj, s)
figure out indices corresponding to s More...

function subsasgn (a, s, b)

function cleanup_subs ()
turn all subsrefs into A(ind)=B(:), instead of A(sub1,sub2)=B. More...

function subsindex (a)

function sin (obj)

function cos (obj)

function sec (obj)

function tan (obj)

function asin (obj)

function acos (obj)

function atan (obj)

function atan2 (y, x)

function acot (obj)

function tanh (obj)

function exp (obj)

function abs (a)

function diag (v, k)
v is a TaylorVar More...

function mod (X, Y)

function conj (a)
intentionally left blank. i've assumed real everywhere in here. More...

function sqrt (a)

function sum (a, d)

function max (a, b, dim)

function min (a, b, dim)

function cumprod (A, dim)

function norm (A, p)

function diff (x, n, dim)

function ppval (pp, xx)
equivalent to ppval.m, but made taylorvar compatible More...

## Static Public Member Functions

static function init (x, order)

Property f

Property df

Property dim

Property nX

## Detailed Description

does inline autodiff overloads operators: http://www.mathworks.com/help/techdoc/matlab_oop/br02znk-1.html

Use TaylorVar.init() to create a new TaylorVar. % Use getmsspoly() to get the approximation around a point close to where % you initialized the TaylorVar.

the internal representation is m=prod(size(f)), n=prod(size(x)); df{o} is a m x n^o sparse matrix , where e.g. d^2 f(i,j)/dx_k dx_l = df{o}(sub2ind(obj.dim,i,j),sub2ind([n n],k,l))

## Constructor & Destructor Documentation

 function TaylorVar ( f , df )
Return values
 obj

## Member Function Documentation

 function abs ( a )
Return values
 a
 function acos ( obj )
Return values
 tv
 function acot ( obj )
Return values
 tv
 function and ( a , b )
Return values
 l
 function any ( a )
Return values
 b
 function asin ( obj )
Return values
 tv
 function atan ( obj )
Return values
 tv
 function atan2 ( y , x )
Return values
 tv
 function bsxfun ( fun , a , b )
Return values
 c
 function cleanup_subs ( )

turn all subsrefs into A(ind)=B(:), instead of A(sub1,sub2)=B.

 function conj ( a )

intentionally left blank. i've assumed real everywhere in here.

Return values
 a
 function cos ( obj )
Return values
 tv
 function ctranspose ( a )
Return values
 a
 function cumprod ( A , dim )
Return values
 B
 function diag ( v , k )

v is a TaylorVar

Return values
 tv
 function diff ( x , n , dim )
Return values
 y
 function double ( obj )
Return values
 d
 function end ( obj , k , n )

end called as part of the kth index of n indices 'help end' was the only way i found documentation on this

Return values
 lastind
 function eq ( a , b )
Return values
 l
 function eval ( obj )
Return values
 varargout
 function exp ( obj )
Return values
 tv
 function ge ( a , b )
Return values
 l
 function getmsspoly ( obj , p_xbar )

returns the taylor approximation as an msspoly

Parameters
 p_xbar should be relative to the original value of x used in TaylorVar.init()
Return values
 p
 function getOrder ( obj )
Return values
 o
 function gt ( a , b )
Return values
 l
 function horzcat ( varargin )

find the index of the first TaylorVar

Return values
 tv
 static function init ( x , order )
static
Return values
 tv
 function inv ( a )

use d/dq inv(H(q)) = - inv(H(q)) [d/dq H(q)] inv(H(q))

Return values
 tv
 function isinf ( obj )
Return values
 tf
 function isnan ( obj )
Return values
 tf
 function ldivide ( a , b )
Return values
 tv
 function le ( a , b )
Return values
 l
 function length ( obj )
Return values
 numberOfElements
 function lt ( a , b )
Return values
 l
 function max ( a , b , dim )
Return values
 a i
 function min ( a , b , dim )
Return values
 a i
 function minus ( a , b )
Return values
 tv
 function mldivide ( a , b )
Return values
 tv
 function mod ( X , Y )
Return values
 X
 function mpower ( a , k )
Return values
 a
 function mrdivide ( a , b )
Return values
 a
 function mtimes ( a , b )
Return values
 a
 function ne ( a , b )
Return values
 l
 function norm ( A , p )
Return values
 n
 function not ( a )
Return values
 l
 function numel ( obj , varargin )
Return values
 numberOfElements
 function or ( a , b )
Return values
 l
 function pinv ( A )
Return values
 tv
 function plus ( a , b )
Return values
 a
 function power ( a , k )
Return values
 tv
 function ppval ( pp , xx )

equivalent to ppval.m, but made taylorvar compatible

Return values
 v
 function rdivide ( a , b )
Return values
 tv
 function repmat ( obj , M , N )
Return values
 obj
 function reshape ( obj , varargin )
Return values
 obj
 function sec ( obj )
Return values
 tv
 function sign ( a )
Return values
 l
 function sin ( obj )
Return values
 tv
 function size ( obj , dim )
Return values
 varargout
 function sqrt ( a )
Return values
 a
 function subsasgn ( a , s , b )
Return values
 a
 function subsindex ( a )
Return values
 ind
 function subsref ( obj , s )

figure out indices corresponding to s

Return values
 obj
 function sum ( a , d )
Return values
 a
 function superiorfloat ( varargin )
Return values
 classname
 function tan ( obj )
Return values
 tv
 function tanh ( obj )
Return values
 tv
 function times ( a , b )
Return values
 a
 function transpose ( a )
Return values
 a
 function uminus ( a )
Return values
 a
 function uplus ( a )

intentionally empty

Return values
 a
 function vertcat ( varargin )

find the index of the first TaylorVar

Return values
 tv

## Member Data Documentation

 Property df
 Property dim
 Property f
 Property nX

The documentation for this class was generated from the following file:
• /home/ubuntu/workspace/linux-gcc-continuous-matlab-documentation/drake/matlab/util/TaylorVar.m