On the usage of resonate and fire dynamics in the complex oscillation-based test approach

Abstract

Resonate and Fire (R+F) models were introduced to account for many phenomena occurring in biological neurons showing sub-threshold oscillations of the membrane potential. In information technology, they are at the basis of Chaotic Spiking Oscillators (CSOs), exploitable in Pulse Coupled Neural Networks (PCNNs). This paper illustrates how the R + F paradigm can also be used for the testing of analog signal processing structures (and specifically filters), extending the Oscillation Based Test (OBT) framework. The rich dynamics of the R + F model is used to encode the block under test features and faults into pulse trains directly processable at the digital level. Means to achieve a precise characterization of firing times are provided and used for parametric testing. Considerations about the trade-off between testing times and accuracy are provided together with a practical example and simulation data. Copyright © 2012 John Wiley & Sons, Ltd.