The goal of this project is to develop field-suitable robotic technologies to assist first-responders in the aftermath of natural and/or man-made disasters. After an earthquake, for example, robots can provide a great deal of help to minimize the operational risks for rescuers, while at the same time increasing the chances to locate survivors quickly and save human lives.
UC Merced professor Stefano Carpin plans to develop a software/hardware module that can be plugged into mobile robots to solve the localization problem in unstructured environments typical of rescue scenarios. Specifically, his team plans to use off-the-shelf hardware components to implement and improve an innovative algorithm they have developed. This practical study will not only lead to a first working prototype of the system, but will also provide extensive experimental data to better understand the problem and stimulate further theoretical investigations to produce more accurate localization results.
Traditional approaches to localization assume the availability of known and detectable landmarks. Such landmarks, however, cannot be assumed to be present in the aftermath of a natural disaster. To overcome this, the engineers propose to use a fleet of robots that is split in two groups. A first group of robots acts as detectable landmarks, while the second group moves around and relies on the first group to solve the localization problem. The roles of the two groups are swapped from time to time, so that all robots move.
Carpin and colleagues will Implement and evaluate the proposed algorithm in a realistic search and rescue scenario to identify further needs for theoretical investigations. They will then develop ready-to-use hardware/software module to outfit mobile rescue robots.