Creating an Ambient Intelligence Environment Using Embedded Agents





Ambient intelligence is an exciting new information technology paradigm in which people are empowered through a digital environment that is aware of their presence and context and is sensitive, adaptive, and responsive to their needs.1 Ambient-intelligence environments are characterized by their ubiquity, transparency, and intelligence. In these environments, a multitude of interconnected, invisible embedded systems, seamlessly integrated into the background, surround the user. The system recognizes the people that live in it and programs itself to meet their needs by learning from their behavior.1 To realize the ambient-intelligence vision, people must be able to seamlessly and unobtrusively use and configure the computer-based artifacts and systems in their ubiquitous-computing environments without being cognitively overloaded.1 The user shouldn’t have to program each device or connect them together to achieve the required functionality. The complexity associated with the number, varieties, and uses of computer-based artifacts requires that we design a system that lets intelligence disappear into the infrastructure of active spaces (such as buildings, shopping malls, theaters, and homes),2 automatically learning to carry out everyday tasks based on the users’ habitual behavior. Our work focuses on developing learning and adaptation techniques for embedded agents. We seek to provide online, lifelong, personalized learning of anticipatory adaptive control to realize the ambientintelligence vision in ubiquitous-computing environments. We developed the Essex intelligent dormitory, or iDorm, as a test bed for this work and an exemplar of this approach. Intelligent embedded agents Embedded intelligence refers to including some capacity for reasoning, planning, and learning in an artifact. Embedded computers that contain such an intelligent capability are normally referred to as embedded agents2 and are intrinsic parts of intelligent artifacts. These autonomous entities typically have a network connection, thereby facilitating communication and cooperation with other embedded agents to form multi-embedded-agent systems. Embedded agents in the form of mobile robotic agents can learn and adapt their navigation behaviors online.3 However, we concentrate on embedded agents in ubiquitous-computing environments that will help us realize the ambient-intelligence vision. Each embedded agent is connected to sensors and effectors, comprising a ubiquitous-computing environment. The agent uses our fuzzy-logic-based Incremental Synchronous Learning (ISL) system to learn and predict the user’s needs, adjusting the agent controller automatically, nonintrusively, and invisibly on the basis of a wide set of parameters (which is one requirement for ambient intelligence).4 Thus, we need to modify effectors for environmental variables (such as heat and light) on the basis of a complex, multidimensional input vector. An added control difficulty is that people are essentially nondeterministic and highly individual. Because the embedded agents are located on small embedded computers with limited processor and memory abilities, any learning and adaptation system must deal with these computational limitations. Most automation systems, which involve minimal intelligence, use mechanisms that generalize actions across a population—for example, setting temperature or loudness to the average of many peoples’needs

Free download research paper



- -

FREE IEEE PAPER