Gauging Intelligence of Mobile Robots
Ka C. · 2006
The paper illustrates three means for gauging the intelligence of mobile robots: The qualitative perception of intelligence based on extending Turing test to the systems at hand; the quantitative measure that gauges task-specific intelligence from its performance in uncertain environments; the comparative measure that calibrates difficulty levels of what the system has to do against levels of human skills. The framework of an agent, an environment and a goal was applied to two case studies was helpful in gauging the intelligence of the mobile robots. An important field experiment reveals that the autonomous obstacle course challenge in the IGVC requires the sensory-motor-decision skill of at least a four-year old. As frontiers of and education efforts on IUVS are being pursued, it is necessary and worthwhile to consider means for improving as well as gauging the intelligence of the systems.