Supervisor: Prof. Per Gunningberg

Uppsala University

Computer Networks II

Supervisor: Prof. Per Gunningberg

2000-06-07

Smart Spaces

Magnus Berggren,

Sven Montán,

Henrik Nord,

Mattias Östergren,

Table of Contents

Abstract 2

1. Introduction 2

2. Smart Spaces 3

3. Some Computer Science Issues of Smart Spaces 5

3.1 Human-Computer Interaction in Smart Spaces 6

3.2 Smart Space Networking 9

3.3 Security and Privacy in Smart Spaces 11

4. Conclusion 13

References 14

Abstract

Recent advances in hardware engineering are now providing us with small, yet powerful networking computing devices. When many of the physical things surrounding us in our own environment are being equipped with these devices, we can exploit them for aiding us in our daily work. Hence our environment will become intelligent. Our homes and offices will become Smart Spaces. This paper will give a definition of Smart Spaces and explore some of the computer science issues of them.

1. Introduction

A man enters his apartment. As he opens the door the lighting turns on and the stereo starts playing his favourite group. The man says: “Open my work-address book. Show it on the wall screen and read it out using this loudspeaker.” After a couple of seconds the address book appears at the flat-screen attached to the wall. The system informs the person that his request is fulfilled with a mellow gong. The man tells the system to send a barbecue-sauce recipe to his boss at her weekend cottage where she usually spends her weekends. The system begins with trying to find such a recipe in the man’s recipe collection. Since this man hates cooking there is no such document. The system then tries to locate it by searching the Web. After locating a suitable receipt, he commands the system to send to the boss, who receives them just in time to put the chicken in its marinade.

This scenario is an example of a Smart Space system and it exemplifies several issues that we will cover in this paper. The first is that the system has the ability to accept spoken language as input. The second thing is that the system has the ability to identify the person saying the sentence. A Smart Space user does not want any non-trusted persons accessing your private information. Furthermore, a Smart Space has several ways of determining the identity of a user, for instance, by computer vision or voice recognition. However, Computer vision gives rise to ethical questions concerning privacy. A lot of people do not want to be monitored by computers. The third ability has to do with finding the location of the address book. In this case the wanted file might be located at the office at the other side of the town. Finally, after the file has been fetched from the office it is presented in two different displays with drastically different capabilities. In this case the man wanted to interact with the address book both visually and vocally. This requires that the data have to be presentable in different formats depending on the medium to which it is sent.

The rest of the paper is outlined as follows. In section 2 we give a definition of Smart Spaces. In section 3 we describe some computer science issues of Smart Spaces. The issues presented are Human-Computer Interaction in Smart Spaces, Networking in Smart Spaces, and Security and Privacy in Smart Spaces. In section 4 we have a conclusion where we summarise the main points of this paper.

2. Smart Spaces

A Smart Space is not a particular technique to solve a specific problem. It is closer to a vision of what the future of computing will be. The envisioned Smart Space system will be comprised of several devices. All devices in a Smart Space will contain some means of performing computations and to offer connectivity, they will be attached to some kind of network. They are envisioned to be heterogeneous in the sense that their computational, network access capabilities and interfaces will differ. The devices may be situated in a limited physical area, hence the word Space. However, separate Smart Space systems are not limited with the physical boundaries. The system is connected to the virtual world and combined with other local systems forming to a larger global entity. The system is Smart in the sense it will provide seamless and transparent interaction to the user as well as to devices. Also, the devices are envisioned to be concealed, perhaps embedded in walls and furniture, making the system transparent and non-intrusive to the user. Other devices are envisioned to be personal and carried at all times. Consequently, when a person carrying a personal device enters a foreign Smart Space it must immediately be incorporated into it. No configuration should be needed. Finally, some Smart Spaces will demand a device to interact with it others will be intelligent and accept camera and microphone input.

The research community have come up with several concepts that address different aspects of Smart Spaces. The vision of a Smart Spaces system incorporate ideas found in Nomadic Computing, Ubiquitous Computing, Wearable Computing, Intelligent Environments and Co-operative Buildings. The fundamental ideas behind Nomadic Computing [2] are derived from the observation that, today, computer users can be considered to be nomads, in that they carry portable computers and communication devices on their travels between office, home, hotel, car and so on. These portable devices range from laptops and Personal Digital Assistants (PDA) to smart cards and wristwatch computers. The nomad is expecting his devices to work adequately wherever he may travel, even though environment and communication capabilities may change dramatically as he moves. Access to computing and communications is necessary, not only from “home base” of the user but also during travels and when the nomad reaches his final destination. This new user behaviour implies fundamental changes in the requirements on our computing and communication systems. The key functions in Nomadic Computing is independence, referring to perception of computing environment and ability to automatically adjust to momentary capabilities of communication and access in a transparent and integrated fashion. The concept of Nomad Computing is rather wide in scope and it partly comprises many of today’s computer and telecom systems such as GSM and Wireless LAN. Nevertheless, the introduction of Nomadic Computing stipulates the major shift of paradigm in how we use computers in our daily life. It also points out many important issues and research areas that are central for establishing Smart Space Systems.

Ubiquitous Computing [4] extends the ideas of Nomadic Computing. In any Ubiquitous Computing system many computers are located in the environment surrounding the user. The key ideas are that they should be abundant, invisible to the user and networked together. The user, with help from a set of small devices, interacts with hundreds of computers located in the physical world around him or her. However, these personal devices are supposed to be small and inexpensive, so that every one can afford them [1, 3]. In order to create a working structure for Ubiquitous computing, computers of all sizes and shapes are needed, not only those computers equipped with a mouse, keyboard and display [5]. Finally, Ubiquitous computing is about facilitating system interaction. Interaction is not necessarily limited to typing at a keyboard or clicking in a display. A Ubiquitous computing system accepts user voice input.

Wearable computer systems oppose to the Ubiquitous Computing approach by having computers everywhere. A user in a wearable computer system carries all necessary data and equipment with him or her, worn like clothing. Data is kept in the worn device and is not shared with the environment [6]. If two different wearable want to share the same resource, they need the capability to communicate with each other [7]. In a wearable computer system it is up to the wearable device to solve the contention for accessing a common resource. The resource itself is highly specialised and it has very little extra computational power. This produces high demands on flexibility of the software of the wearable device.

An Intelligent Environment resembles a Ubiquitous Computing system. However, in the user needs no device to enable interaction with the Intelligent Environment system. The system accepts only camera and microphone input. Intelligent environments, as defined by Michael H. Cohen, are “spaces in which computation is seamlessly used to enhance ordinary activity” [17]. By embedded Cohen means that the main input for theses systems are cameras, microphones and speakers coupled with computer vision, speech recognition and speech synthesis algorithms. These functions allows the user to perform interaction as he or she naturally should feel comfortable with, by performing a dialogue with the system as if the system was another person. The aim is to make the user forget that he or she is interacting with a software system.

Finally, A Co-operative Building as defined by Norbert Streitz et al. is ”flexible and dynamic environment that provides co-operative workspaces supporting and augmenting human communication and collaboration” [18]. The Co-operative Building is a kind of Ubiquitous Computing system where the emphasis is in enabling computer-aided co-operative work. However, the co-operative building should not only augment co-operation among the users but also work in a co-operative manner with them. Furthermore, a Co-operative Building should not be limited by the physical dimensions of it, but also provide seamless access and interaction with the virtual space. The parts of a system that constitute a Co-operative Building may not be physically located at the same place. The system on a whole brings together local environments to a seamless entity. Within this entity users can interact as easily as being in the same location.

3. Some Computer Science Issues of Smart Spaces

The concept of Smart Space touches several fields of computer science. Smart Spaces are envisioned to be brought in our daily life where computing traditionally have not been before. Since the Smart Space system will be all around us at all times, it will create challenging demands on how users interact with it, how the devices communicate with each other and how it protects sensitive information. Having a user centric view, Smart Spaces provides an easy interaction with services and resources within it. Also, means of interaction should be non-intimidating and non-intrusive. Currently there are two major approaches to interacting with a Smart Space system. In the first case, the user does need to have any artificial device in order to interact. It suffices for the user to be present and have functional vocal cords. Such a system accepts camera and microphone input. In the second, the user must be provided with some kind of device. Human-computer interaction issues here are concerned with limited GUI space and efficient distribution of resource capabilities.

A Smart Space system is populated with several heterogeneous devices. It must enable a generic way of locating devices and provide a minimal set of common communication primitives. There are basically two ways of solving resource locating. One is by taking a client/server approach, where the server has a list of all available resources and where they can be found. The other is to keep a decentralised approach and let the resources keep a small execution environment where devices can download code. When this code is executed, the resource performs the desired effect.

Finally, users may make sensitive information available without being conscious of it. How should the Smart Space protect the user from this? Also, in a Smart Space there are potentially several different kinds of sensors such as cameras or microphones. A user may want to remain incognito or be locatable. Finally, Smart Spaces also need to protect information that a user has given away consciously.

3.1 Human-Computer Interaction in Smart Spaces

Smart spaces aim at bringing computing in where it had no place before. Also Smart Spaces aims at centring technology around the human being and not the other way around. Interaction with a device in the Smart Space must be conducted effortless and potentially demand little or no prior training. Consequently, a large research and developing area of Smart Spaces focuses on the human-computer interaction issues.

The research and developing efforts in this area are mainly about understanding the requirements of how humans interact and co-operate in a particular situation. The idea is then to aid this situation by introducing computing, without violating the constraints of effortless interaction. A typical challenge for a human-computer interaction researcher is to come up with mechanisms that will aid a team in a creative process [9, 18]. Such a process takes for instance place by a whiteboard where members of a programming team discusses first-attempt solutions by drawing diagrams and pieces of code on the board. Which kind of computing could aid this situation without hampering the already straight forward usage of a pen and a whiteboard?

Inspired by how we interact with each other, researchers are exploring the fields of computer vision and speech recognition as means of unencumbered user input to the Smart Space system. With computer vision comes the ability to, for instance, accept gestures as cues to commands and recognition of faces to improve a systems accommodation to personal needs. Computer vision is used in a wide row of research projects and commercial systems. The main usage of this technique is recognition of people frequenting the Smart Space, gesture recognition and various forms of symbolic interaction based on computer vision.