Fixed-Point ATAN2 using CORDIC

This demo shows how a fixed-point cordic algorithm to calculate a phase from polar coordinates (arct

Description

This demo consists of a m-file script (fixed_point_atan2_using_cordic.m) and a m-file function (atan2_fixpt.m). The script contains a step-by-step explanation of how a four quadrant arctan can be calculated using a CORDIC (COordinate Rotation DIgital Computer) algorithm. The first part shows the CORDIC algorithm in floating-point double precision, whereas the second part of the m-file explains how to implement the CORDIC in 16-bit fixed-point with bit shifts. The third part shows how to generate C code for the m-file function to speed-up the algorithm.

The m-file function contains an implementation of the CORDIC in Embedded MATLAB™ compliant m-code. This function can be used to generate C-code or HDL (VHDL or Verilog) in MATLAB® of Simulink®.

• Required Products: Fixed-Point Toolbox
• MATLAB release: MATLAB 7.5 (R2007b)

[Ali]

can you please elaborate how you came across the trade off between fixed point and algorithmic precision.
Thanks