Real Time Sequential Detection for Diffusion Signals
Date: December 09 - December 11, 1987
We present the sequential detection problem for diffusion type signals in both the fixed probability of error and the Bayesian formulations. The optimal strategy in both cases is a threshold policy with explicitły computable thresholds. We provide numerical schemes for approximating the relevant likelihood ratio and provide an architecture for real-time signal processing.