Today, the Internet of Things (IOT) concept is gaining much attention and popularity; The related technologies as spacesuits and embedded ECG acquisition device is already existed. However, there are important issues to be resolved when an application is in a space environment. The ECG signal may be measured by different mobile conditions when embedded in spacesuits, requiring a more robust algorithm to remove exercise and noise issues. Thus, we propose a more complete architecture with a new storing polymorphic average template (SPAT) and a multistage identification algorithm (MIA) to improve the robustness of ECG identification in motion. In addition, we select better combinations of de-noising and feature extractions to create a better and more complete architecture. According to our experimental results, our proposed architecture offers better performance than previous adaptive boosting (AdaBoost) methods; thus, it is also suitable for application in astronaut spacesuits.