Categorizing Motion: Story-Based Categorizations

Our most primary goal is to provide a motion categorization for moving entities. A motion categorization that is related to how humans categorize motion, i.e., that is cognitive plausible, and can take advantage of the computing power; for example, that we can use in reasoning algorithms. In short, a motion categorization that is both attractive to psychologists and computer researchers: both applicable in human cognition and artificial intelligence. We fulfill this goal not only by providing six of such motion categorizations, e.g., Motion-RCC and Motion-OPRA_1, but, even more, providing a method for generating eventually an endless number of motion categorizations; we call the categorizations created by our method ‘story-based categorizations’. The key of our method is that we generate the motion categorizations from already existing spatial categorizations—concretely, those called spatial representations—and, of such spatial categorizations, we find a great number in the literature. The very first motivation for motion categorizations that are cognitively plausible is to apply them in human-aware navigation. In that sense, the story-based categorizations bridge the gapbetween the human understanding of motion and the computer requirements for information processing. Beyond that, the motivation for motion categorization is, on its own, overwhelming; we see categorization as the first step in knowledge representation—the most basic cognitive task. Motion categorization provides a solution for processing the flood of motion data that is being currently generated by the myriad of navigation and motion control systems.Another important goal, which serves the primary one, is to relate the understanding and terminology of categorization in psychology with those in artificial intelligence. We also present the foundations of categorization according to psychology, and explain how computer scientists deal with spatial and motion categorization, even mentioning underlying linguistic aspects. In that way, we can later introduce our elementary categorization model under consideration of the most basic elements in categorization. We use this model to describe and deal with the story-based categorizations throughout this work.Finally, we verify that the story-based categorizations are not just categorizations, but display very convenient properties. Most importantly, they reflect categorization aspects of human cognition, and they are endowed with the powerful reasoning operations of the qualitative calculi. At the end, we present additional properties in a broad summary: story-based categorizations can achieve extraordinary variety of categorization criteria, they are applicable in high dimensional spaces, they can categorize any sort of trajectories (even when entities are motionless), they are suitable for decision-making and control, and they can categorize motions of multiple (more than two) entities.