Affiliations: Department of Mechanical and Aerospace Engineering, University of Texas at Arlington, Arlington, TX, USA | Department of Computer Science and Engineering, University of Texas at Arlington, Arlington, TX, USA
Abstract: This paper presents an exploration of the characteristics and structure of a cognitive architecture for control of assisted living systems. The aspects of cognition considered are self-organization, communication, and inherited knowledge. A cognitive solution for a related problem, function optimization, is developed because of the complexity and size of the assistive living problem. Support for this approach stems from the artificial intelligence field where optimization is considered to be a critical aspect of cognition, and from the similarity between the performance metrics for the two problem domains. A search algorithm is developed using the bracketing and gradient methods as inherited knowledge. A key finding is that using a cognitive structure caused the search to display aspects of the characteristics, behavior, and performance of human cognition. In terms of performance, the cognitive search converges faster than either the bracketing or gradient searches alone, and its feasible problem set is larger than the intersection of their individual sets. Similarly, human cognition acts quickly and can address a large set of dissimilar problems. This gives confidence that the guidelines distilled from the development of the cognitive search can produce a similar level of performance when applied to an assistive living system. However, this paper does not address the details involved in actually implementing these guidelines on an assistive living system.
Keywords: Cognition, cognitive control, optimization, search, assistive living system
