Pervasive Computing (a.k.a. Ubiquitous or Ambient Computing) is a new field of computer science, which is generally considered to lie at the intersection of desktop computing, networking and embedded systems. Related pervasive computing technologies are Intelligent Environment and Smart Spaces. Intelligence is related to the responsiveness, the ability to adapt and the ability to learn of the networked and computerized devices and appliances that populate the user's environment. Unfortunately, the ongoing efforts in the industry are mainly focused on applications and devices that explore short-term technical issues, and there is no standard definitions or even agreements on the common vocabulary associated with these technologies. Pervasive computing is mainly about creating computerized and networked environment that will help the user to perform technical tasks and to help with daily "chores". Human Computer Interaction is therefore another important research area of the Pervasive Computing field. For the NIST Aroma project, we have developed the Layered Pervasive Computing model, where users are considered as key elements of pervasive computing. We are also developing EXiST, the Experimental Simulation Tool. It is being created to explore pervasive computing requirements and use cases with the ultimate goal of using a case-based approach to defining this field. Artificial Intelligence will be part of this new world, and we are currently exploring the use of Intelligent Agents. They can learn or even teach how to use resources by interacting with human users, providing a higher level of intelligence to the environment, but they also can be used as human models in simulations or in expert systems. We are using COGNET, a cognitive agent modeling toolkit in conjunction with EXiST to create ?unit? agents to help define usability metrics in pervasive computing.