In this chapter, we explore existing robot agility research efforts, while focusing on key technologies that help to enable agility, such as the ones mentioned in the preceding paragraphs. In particular, we will look at perception, knowledge representation, task planning, motion planning, and various artificial intelligence approaches. We then show an example of how these technologies can be put together into an overall architecture that enables robot agility in a small set of sample implementations, focusing on task failure (dropped part) and robot failure (robot breaking down). Lastly, we describe initial efforts in developing test methods to assess robot agility, which will explore sample agility challenges and associated metrics in a robot kitting operation. We show how these test methods are being validated via the Agile Robotics for Industrial Automation Competition (ARIAC) (http://www.nist.gov/ariac/
), which is a simulation-based competition. ARIAC allows teams to develop agile artificial intelligence (AI)-based systems that will control robots in a simulated manufacturing factory floor. Agility challenges are introduced to the system during the production operations to see how well the systems respond.