CUDA implementation of an optimal online Gaussian-Signal-in-Gaussian-Noise detector

We address the computationally demanding task of real time optimal detection of a Gaussian Signal in Gaussian Noise. The mathematical principles of such a detector were formulated in 1965, but a full real-time implementation of these principles was not possible for decades mainly due to technological barriers. We present a CUDA based implementation of such an optimal detector and study its decision making speed (or throughput) as function of target signal duration and signal filter length. We also compare the throughput results to those of a CPU based design. We report on detection rates ranging from 3.5 KHz for a target duration of 10756 samples up to 15.6 KHz for target duration of 92 samples. The CUDA based detector running on 384 parallel cores had a superior throughput comparing to a pure CPU implementation when target duration was longer than 600 samples.