An analytical framework and implementation architecture for a reconfigurable signal classification system

We present an analytical framework for the mathematical development of signal classification systems and a closely associated implementation architecture for the instantiation of the classifier to an embedded system. The analytical framework provides a context for the development of a classification system that address those design challenges that are fundamental to the operation of the system and those that are related to its practical deployment. The implementation architecture is closely associated with this analytical framework and permits the rapid construction of efficient, accurate systems for deployment to embedded hardware. We present the results of several system development activities which have used this approach and which demonstrate the dramatic reduction in development effort resulting from the approach.