Return to Robotics Home page                                                            

CMAC Neural Network Code

Department of Electrical and Computer Engineering    
University of New Hampshire

Contact: Dr. W. Thomas Miller  (

    Much of the work on robot learning/adaptation within the UNH Robotics Laboratory has used variations on the traditional Albus CMAC for the nonlinear adaptive element(s) in the control. While we have used various fixed-point and floating-point, hardware and software CMAC implementations in our research over the years, the following files contain the fixed-point software version of CMAC which we use most-commonly today.

Available CMAC related files:

The CMAC C code.
The CMAC header file.
Documentation for the UNH CMAC code   (postscript format).
C code for a simple CMAC demonstration.
A more interesting Windows 95/NT demonstration   (executable binary only).