Breathtaking advances in computer and communication technology open a window on a new and exciting world, where flocks of electronic devices become an integral and natural part of our everyday lives. A world that is full of smart objects because of embedded microcomputers with adequate communication capabilities. Where integrated sensors create a “contextual aware” environment by sensing and communicating events and personal data to the system, that combines them in meaningful responses to the end-user. With electronics integrated into clothing, furniture, and mirrors and probably even in the construction materials of our houses and offices. And where the electronic functions spontaneously organize themselves in networks to provide their human companions with communication and entertainment services and with access to rich content wherever they are. A world referred to as “Ambient Intelligence”.
Many devices or sub-systems working together in a network create Ambient Intelligence environments. All Ambient Intelligence devices have on the one hand communication capabilities for the interaction with other devices and on the other hand computational capabilities to implement the interaction protocols and to process the information.
Energy constraints are a determinative factor for the distribution of functionality over the devices in the Ambient Intelligence network, when developing the architecture. Because of these power constraints, ambient intelligence environments will consist of a hierarchy of devices with optimized functionality. This hierarchy of devices constitutes of the following;
Autonomous micro devices (so-called micro-Watt nodes)
Portable mini devices (so-called milli-Watt nodes)
Static maxi devices (so-called Watt nodes)
The class of autonomous micro devices is meant to collect and disseminate a range of environmental data like light, temperature, sound pressure, vibrations, humidity, heart frequency, etc. Thus, these devices harvest data in the process of perceiving the environment, making them an essential element in the “awareness” part of Ambient Intelligence. Networks of these devices might implement applications like environmental control in homes and offices; person or health monitoring; identification; diagnostics; smart home; security; robot control; etc. These devices are lightweight, extremely small and very cheap. They have to operate autonomously without any supervision over their entire lifetime and need to organize themselves in networks without any manual interaction.
Portable mini devices will be the workhorses of the Ambient Intelligence environment. They are wireless infotainment pocket consoles; battery or micro fuel cell powered and small enough to be carried on the body. In this category one will find the current and future GSM and UMTS phones with built-in cameras and Audio/Video (A/V) storage, wireless video tablets, portable access devices to audio and video information, personal digital assistants and so on. In terms of functionality they are the “A/V everywhere” part of the Ambient Intelligence environment.
Static maxi devices or “Watt nodes” constitute the highest layer in the Ambient Intelligence device hierarchy. Watt nodes have neither power (besides heat and fan noise) nor volume constraints and consequently can lodge large amounts of processing horsepower and bulk storage. Examples of “Watt nodes” in the Ambient Intelligence environment are residential gateways, home servers, large (plasma) displays, etc. Their functional architecture resembles that of milli-Watt nodes but with additional interfaces to (wired) broadband access networks and of course much higher performance figures. Watt nodes also implement the processing-intensive multi-modal interfaces and User Interface (UI) applications that ultimately will allow the user to interact in a natural way with his/her Ambient Intelligence environment. They give access to personal audio and video databases that are managed via automatic feature extraction and retrieval. For instance the user can search the audio database via “query by humming”. Because of their almost unlimited processing power Watt nodes will play an important role in the control of the total Ambient Intelligence environment and in the format conversion of the audio/video material for consumption on the milli-Watt nodes with their limited internal processing power and different display technologies and resolution. The various nodes in the Ambient Intelligence architecture organize themselves spontaneously in networks as shown in the illustration below.
Micro-Watt nodes are, due to their power limitations, very dependent on neighboring nodes for reliable communication with the control center of the Ambient Intelligence environment they are in. We should expect the milli-Watt nodes to act as a kind of relay station between the micro-Watt nodes and the Watt-node that acts as control station. There are two possibilities. Micro-Watt nodes could use adjacent micro-Watt nodes as “hopping” devices to bridge some distance to the milli-Watt node or are triggered by a wake-up signal from the milli-Watt node when the latter is in reach. This requires the (ultra-low power) receiver to be on for most of the time which is unattractive from an energy point of view.
