Evolution of a Performance Metric for Urban Search and Rescue Robots (2003)

This paper discusses the advancement and proliferation of the Reference Test Arenas for Urban Search and Rescue Robots as representative search and rescue environments, which allow objective performance evaluation of mobile robot capabilities. These arenas have hosted annual rescue robot competitions sponsored by the American Association for Artificial Intelligence since 2000 and the RoboCupRescue Robot League since 2001. The rules of these competitions have evolved each year to encourage robots to negotiate complex and collapsed structures, find simulated victims, determine their condition and location, and generate human readable maps to enable victim recovery. The associated performance metric has also evolved as it attempts to quantify and encourage these and other robot capabilities pertinent to urban search and rescue applications. This paper presents the competition rules, performance metric, and generalized results of the 2003 competitions which included some inspiring robotic implementations. Performance data captured during these competitions is discussed along with recently available development tools which can quicken the pace of innovation in the field of search and rescue robotics.