Technologies for Smart Environments: Capacitive User Tracking and Proactive Fuzzy Control

Smart environments are built upon many different kinds of technologies. These technologies, which are utilised in a variety of sensors, actuators, displays and computational elements within the environment, together define the smartness of the environment. Therefore, to support the users of smart environments in the most effective way, each technology must be perfected separately for its particular purpose. Moreover, new technologies must be under constant scrutiny in order to achieve the aims of enriching society in a more efficient way. In this work, we present two different types of new technologies for the creation of smarter living environments such as smart homes. More specifically, we will concentrate on unobtrusive methods for measuring user activity and for learning user routines in smart environments. Once an individual's routine has been learned, we will present techniques to control the environment in a proactive way to diminish the user load and to create a calm living environment. First we introduce several methods of capacitive measurement that can be used to passively track an individual indoors in three dimensions and in a non-intrusive way. In addition to providing a privacy-preserving way to monitor a person's position and posture, the methods presented here can be used to deduce how the user interacts with the environment. The methods have been tested in a real smart home environment and have proved to be a feasible solution for tracking an individual without the user having to wear any electronic tags. Therefore, these methods promote Mark Weiser's concept of calm technology and are suitable for both residents and visitors to a smart home. Second, the development of new methods for learning user routines and controlling the environment are presented. Inspired by the concept of a calm environment, these computational methods enable people's routines to be learned without any conscious input from the users, other than their daily use of everyday household devices such as light switches or curtain cords. Moreover, the algorithms which have been developed can adapt to changes in the user's lifestyle without the need for any user maintenance or adjustment. Thus, they can always be used for the proactive control of domestic appliances without anyone, including the users, needing to change or even understand the inner workings of the system. Nevertheless, because the control and learning algorithms are built with simple fuzzy control rules using common linguistic terms, the rules are self-evident to most people, which means that specific rules to modify how the system operates can be added if necessary. Since the methods developed here have been built into an actual (rather than virtual) test system and used over long periods of time, these technologies have the potential to be developed into commercial products without too much effort. Nevertheless, there is still plenty of scope for further research, and this work also indicates several useful directions for further study. Taken together, the technologies developed here are a step along the road to feasible intelligent environments for the future