Mechanics
RATAR makes use of a mobile robotic platform in order to explore disaster zones. The robot uses a reliable electric four wheel drive system to allow it to traverse
uneven terrain at high speed. The drive train provides 1.77hp in order to also allow the robot to clear obstacles and move objects out of the way through brute force if necessary.
Sensing
The RATAR solution uses a single Microsoft® Kinect® to view its surroundings, as well as additional sensors to keep track of its own state. The Kinect® provides an image
called a
depth-map
that is used in the three-dimensional map building. The Kinect® can also see in the dark through its IR camera, as well as listen to the environment through an array of four microphones.
Brains
The RATAR system uses an approach called
SLAM to build
a three-dimensional map of its surroundings in real time as it explores them using an exploration-based decision making algorithm. This map can then be provided to on-site rescuers to facilitate
safe "virtual" explorations of the disaster area.
The microphone array is used to provide further information through a process called sound-source localization. Through the use of its on-board microphone array, that RATAR robot is able
to determine the approximate direction of any sound that it hears. This allows its search alogrithm to operate with increased efficacy by allowing it to zero in on human voices and search
areas it suspects contain injured victims more thoroughly. Additionally, this allows the robot to "see through" walls in a manner of speaking. By hearing what is occuring on the other side
of a door or wall, the RATAR robot can direct rescue efforts in the direction that it suspects a person may be trapped.