Embedded Software and Motor Control Libraries for PXR40xx

Function GFLIB_Lut2D_F32
• f[x,y] is the interpolated value
• f[x,y 1] and f[x,y 2] are the ordinate values at, respectively, the beginning and the end
of the final interpolating interval
• x 1, x 2, y 1 and y 2 are the area values, respectively, the interpolated area
• the x, y are the input values provided to the function in the f32In1 and f32In2
arguments
The interpolating intervals are defined in the table provided by the pf32Table member of
the parameters structure. The table contains ordinate values consecutively over the whole
interpolating range. The abscissa values are assumed to be defined implicitly by a single
interpolating interval length and a table index while the interpolating index zero is the
table element pointed to by the pf32Table parameter. The abscissa value is equal to the
multiplication of the interpolating index and the interpolating interval length. For
example let's consider the following interpolating table:
Table 4-90. GFLIB_Lut2D example table
ordinate (f[x,y]) interpolating index abscissa (x) abscissa (y)
-0.5 -1 -1*(2 -1 ) -1*(2 -1 )
pf32Table 0.0 0 0*(2 -1 ) 0*(2 -1 )
0.5 1 1*(2 -1 ) 1*(2 -1 )
1.0 N/A 2*(2 -1 ) 2*(2 -1 )
The Table 4-90 contains 4 interpolating points (note four rows). Interpolating interval
length in this example is equal to 2 -1 . The pf32Table parameter points to the second
row, defining also the interpolating index 0. The x-coordinates of the interpolating points
are calculated in the right column. It should be noticed that the pf32Table pointer does
not have to point to the start of the memory area with ordinate values. Therefore the
interpolating index can be positive or negative or, even, does not have to have zero in its
range.
Special algorithm is used to make the computation efficient, however under some
additional assumptions as provided below:
• the values of the interpolated function are 32-bit long
• the length of each interpolating interval is equal to 2 -n , where n is an integer in the
range of 1, 2, ... 29
• the provided abscissa for interpolation is 32-bit long
The algorithm performs the following steps:
1. Compute the index representing the interval, in which the bilinear interpolation will
be performed in each axis:
Embedded Software and Motor Control Libraries for PXR40xx, Rev. 1.0
370 Freescale Semiconductor, Inc.