Mixed Robot-Sensor Networks for Search and Rescue

We explore how a system of wireless sensors augmented with mobile robots can effectively explore a physical environment, focusing on target detection/tracking and area mapping. Areas of potential application include urban search and rescue, mining detection, and planetary exploration.

We focus on how such systems may be built using the minimal possible hardware.

Particularly, we've developed distributed gradient-based techniques that allow for target tracking without the use of either Global Positioning System (GPS) equipment or digital compass, while allowing for graceful degradation and scalability of the system.

This applet demonstrates our gradient-based search algorithm for a robot-sensor network.
Sensors are scattered about the area. Those detecting a target (X) begin propagating the gradient outwards. The robots are able to follow this gradient without GPS, compass, etc. - they only need to know when they are getting closer or further from a given neighboring sensor.

To use this applet, first press "setup", then "go", feel free to play with the sliders.

This demo was written using NetLogo - demo copyright © Joshua Reich, 2005

Selected Publications:

Joshua Reich and Betsy Sklar, Robot-Sensor Networks for Search and Rescue, In IEEE International Workshop on Safety, Security and Rescue Robotics, Gaithersburg, MD, August, 2006. [Paper] [BibTeX]