A.H.R.D - Autonomous Human Recognition Device

During disasters like earthquakes, several people lose their lives in the debris, as they cannot be rescued spontaneously. Situations, where the rescue team cannot act immediately due to lack of the information pertaining to the existence of lives trapped in remains, urge an expeditious need of some technology/tool/device triggering an alarm and thus revealing the existence of living beings ambushed in the affected area. Certain devices already exist, but the extremely high cost of the devices refrain many countries to buy it and use during the catastrophe. As per NDMA (national disaster management authority), there is not a single registered technology in the country that can easily be implemented in small towns and villages to deal with such critical situations. So as a step forward, this paper is aimed at constructing a cost effective handy autonomous robot for sensing and recognizing human beings caught in calamitous situations, having ability to steer itself on uneven terrains independently of user control and avoiding obstacles coming its way, by building a predictive model to recognize images by imparting training using open source machine learning framework.