A Dataset for Aftermath Victim Detection Behind Walls or Obstacles Using an UWB Radar Sensor

Abstract

First Responders (FRs) are the vanguard in cases of natural or manmade disasters by saving lives, preventing the spread of panic and generic aftermath crisis management. For this reason, they need to be equipped with a whole arsenal of tools which can assist their senses during field operations. To augment FRs’ vision, in this work we employ a commercially available Ultra WideBand (UWB) radar sensor for the purposes of victim detection behind large obstacles, such as walls and doors. In particular, we have created and present here a dataset, which incorporates about 15 hours of data records, for a number of different scenarios. This dataset has been uploaded in an openly accessible database to give the opportunity to the research community to apply and further develop methods for human detection behind large obstacles. Finally, we introduce a novel and of low complexity method which is applied in the collected dataset managing to attain a more than 95% accuracy in victim detection.