Philips Research - Technologies
The technologies of Ambient Intelligence |
|
 |
|
|
The ability to implement ‘ubiquitous computing’, a situation where
processing power is embedded into everything, is certain to arrive
within the next ten to fifteen years. When it does, Philips will be
able to fully realize its vision of Ambient Intelligence, using
technology that is embedded into the fabric of our surroundings to
create an environment that is sensitive and responsive to our
presence and sympathetic to our needs. In the meantime, Philips
Research is already exploiting key characteristics of Ambient
Intelligence to develop product concepts that will allow us to
experience some of the benefits much sooner than you may think.
These can be typified by the following characteristics:
+ Context Awareness + Personalized + Immersive + Adaptive Context awareness One of the key characteristics of devices in an Ambient Intelligence environment is that they must know the context in which they are being used. Screens, for example, become very general-purpose devices, capable of displaying highquality video and graphics when needed, or transforming themselves into transparent windows or mirrors. To configure themselves correctly without user intervention, they must know what they are being used for. This context awareness is even more important in mobile devices because the convergence of productivity tools with communication and entertainment devices (for example: PDA, mobile phone and audio jukebox combined) will make mobiles highly multifunctional. Yet the ways in which these multi-functional devices respond must be very different when we are sitting in a business meeting to when we are walking through a shopping mall, and different again when we are watching a movie at the cinema. Because location is a valuable indicator of the context in which mobile devices must operate, Philips Research is developing ways of locating them to within a few metres, and feeding them with information that is relevant to their environment. User interface, connectivity and cost constraints mean that mobile users do not have the luxury of being able to browse the Internet the way that PC users do. As a result, the information that is made available to them must be much more highly personalized and relevant to what they are doing. Philips Research’s CAMP (Context Aware Messaging Platform) demonstrator uses Bluetooth shortrange wireless transmissions between stationary ‘beacons’ and mobile devices to deliver personalized context-driven information to users. Beacons and users both emanate digital ‘auras’ that detail their interests and offerings, and wherever these coincide meaningful information can be exchanged. For example, if you are attend a meeting with people that are not familiar to you, a beacon in the room could relay their electronic business cards to your mobile device so that you could readily identify them. If you enter a shopping mall, a beacon could alert you via your mobile that several of your friends are also there, and even tell you which stores they are in. It could also alert you to special offers on products that you are interested in buying. The challenges in this work involve the development of conceptual models for the mutual exchange of information, software platforms that support this new communication paradigm in a technology-independent way, and methods of ensuring privacy and trust in the exchanged data. Philips Research also investigates the type of profiles needed in different situations, for example, in business or social environments, at home or on the move. Personalized This new class of mobile devices, typified by a 10-to-15-cm full-colour screen, large amounts of non-volatile memory, wireless connectivity and considerable media processing power, will soon allow us to enjoy the same experiences we enjoy at home while we are on the move, albeit on a smaller scale. Such devices will be able to store and playback entire music collections and several full-length movies, but they will also act as an electronic repository for all those things we normally take with us or collect while we are on our travels, such as money, receipts, credit cards, business cards and photographs. The distinction between PDA, mobile phone, audio jukebox and portable DVD player will disappear as these functions converge in a single device that also becomes an electronic wallet. And the development of body-area networks will quickly consign this device permanently to a clip on your belt, connected wirelessly to headsets and ultra-slim pocket-size displays. It will be the one thing that you take with you wherever you go and it will be a true reflection of your own personality and identity. Philips Research is also working in conjunction with sportswear manufacturer Nike on an even more personal application for these devices. By developing biosensors that can be built into clothing to detect parameters such as heart rate, respiration rate and blood oxygen levels, it will be possible to add personal fitness training and healthcare-monitoring capabilities to your universal personal infotainment companion. This very high level of personalization and the ability to emanate a digital aura that invites and enhances the opportunity for new experiences are both key characteristics of Ambient Intelligence. For mobile devices, however, they must be implemented in a way that allows free interaction in the same way that people on the move make acquaintances and share experiences in a largely spontaneous way. The problem today is that many of the devices that are beginning to show convergent functionality remain slaves to the PC, relying on it for synchronization, content formatting and exchange. A major research track for Philip Research is therefore to break this master-slave relationship by developing peer-to-peer communication mechanisms that mirror natural human interaction. To facilitate meaningful interaction you need to know about personalization, profiling, recommendation and content management as well as the content that is being exchanged, and these are all areas where Philips already has a significant IPR (Intellectual-Property Rights) portfolio. However, to guarantee the interoperability needed for free exchange, this IPR must be implemented within the framework of open standards, which is why Philips is basing peer-to-peer consumer product networking on open industry standards such as UPnP (Universal Plug and Play). Philips Research has developed a software stack around UPnP’s Content Directory Service (a service that allows user-interface devices to browse the content on servers and obtain detailed information about individual content objects). Utilizing this, a demonstrator has been produced in which a mobile audio/video jukebox automatically sets itself up as a server to stream audio and video to a WiFi-connected TV client as soon as you bring them within range of each other. Because of the multiplicity of formats in which content is still created, peerto- peer connections such as this will require a considerable amount of transcoding between formats, which may have to be done on the client or the server. Although transcoding can be performed by software on a computationally powerful enough processor, this approach consumes far too much electrical energy to be compatible with battery-powered mobile devices. In cooperation with Philips’ Semiconductors division, Philips Research is therefore developing hardware accelerators (co-processors) that can be integrated into low-power, low-voltage system-on-chip solutions to perform these transcoding functions. Immersive If one role of Ambient Intelligence is to make the delivery and sharing of experiences much easier, another role is to enhance the quality of the experience itself. And being ‘in the mood’ for something goes a long way to making the experience truly enjoyable. Mood, however, is not something that is created entirely within us. It can also be invoked by our surroundings, in the form of lighting levels, the colours and smells around us, the vibrations that we feel and the sounds that we hear. For Ambient Intelligence to heighten our experiences by reflecting or invoking our moods, it must therefore be able to control our environment. Enabling this mood control is the subject of Philips Research’s Physical Mark-up Language (PML) project. In much the same way that HTML (Hypertext Mark-up Language) informs web browsers about how web pages should be rendered onto our computer screen, PML informs physical objects in our surrounding how experiences should be rendered into our rooms. Unlike HTML, however, which has the benefit of only one or two web browsers to contend with, PML has to be suitable for many different home situations, where the context for the experience, for example the layout of the room, the ambient conditions, the availability of equipment and the intentions of the user, are dynamic and unclear to the author. As a result, it does not define the rendering explicitly in the way that HTML defines the point size of a font or the pixel size of a graphic. Instead, it aims to define experiences in terms of the mood that should be created, leaving ‘experience browsers’ to render the experience using whatever PML-enabled devices exist in a particular environment. Adaptive Context awareness, personalization and immersion are some of the important steps towards Ambient Intelligence that are already beginning to appear in consumer products. The next major challenge for Philips Research is to add the ability for environments to become highly adaptive – a characteristic that relates very much to how people will interact with Ambient Intelligence. One of the main themes of Philips Research’s current work is to make this interaction more natural, by developing technologies that will allow us to use modalities such as speech and gesture. However, identifying who is speaking and recognizing what they are saying is a highly complex task in a home environment. Two or more people may be speaking at the same time, and they may be moving around while doing so. Other sound sources may also be present, such as radio or TV sets. As a result, microphones will often be bombarded by sound, sometimes receiving it directly but sometimes receiving it largely via reflections from the walls of the room. Depending on the situation, the algorithms needed to optimally condition the sound signal for the speech recognizer will need to vary, and this will have to be done dynamically as conditions change. To overcome these problems, Philips Research is using beam-forming microphone techniques that can focus on sound sources, together with image recognition systems that can analyse lip movements to identify and track individual speakers. The same image recognition system will be used to experiment with video-chat links that automatically identify speakers and capture their images irrespective of where they are sitting. By giving us the freedom to access content and to control our environment in a natural and intuitive way, and by providing the means to immerse ourselves in the resulting experiences and the ability to create a real sense of community in a physically fragmented world, Philips Research has already gone a long way to making Ambient Intelligence a reality. In the process, it is gaining a much greater understanding of what experiences, not technology, mean to real people.
More
+ What is Ambient Intelligence? + Technology breakthroughs |
 |