In recent years we have been witnessing a number of natural and man-made disasters such as hurricane Katrina and the London bombings. Such disaster pose major situational awareness and coordination problems for emergency responders. The ALADDIN (Autonomous Learning Agents For Decentralised Data and Information Networks - www.aladdinproject.org) project, which started a year ago, has been investigating agent-based technologies that could be used to tackle such problems. In particular, researchers at the University of Southampton have been looking at the use of sensor networks to try and make sense of the data gathered about the environment and decentralised/centralised coordination mechanisms that could enable agents to achieve their goals in such environments. In this talk I will present a testbed, developed in the ALADDIN project, based on the RoboCup Rescue (www.robocuprescue.org) platform, that provides a near-realistic simulation of a disaster scenario. In developing this testbed, our aim is to evaluate and benchmark technologies developed by agent researchers in more realistic settings and showcase the effectiveness and efficiency of these. I will also discuss some of the techniques we aim to apply/have applied in such contexts.