Personal Interactive Computing Objects

Technology is usually the main driving force behind the capabilities of our computer systems and as a result a change in technology can have far reaching consequences for the architecture of future systems. There has been a trend over recent years for processors to be made that are small with low-power consumption, the smart card has to some extent encouraged their development. This paper discusses the variety of Personal Computing Objects that can now be made, some of their uses, and how they may change our view of the traditional computing environment. Modern day distributed computer systems are often referred to in the context of a computing environment. As personal computing is becoming possible with smaller and smaller hand-held devices, there is the potential for the now familiar computing environment to become more tightly integrated into our daily lives, particularly in the context of an o ce building. Portability is an essential factor providing the opportunity for us to carry a computer of reasonable capability with us for most of the day. The key component to integration is communication: by restricting the local computation requirements of a portable device to a basic level its size and power consumption can be kept to a minimum while the `real' computing power is utilised elsewhere. A portable communication link of this kind is most suited to radio telemetry but infrared communication may also have application. At Olivetti these devices have been called Personal Interactive Computing Objects or PiCOs and this class of device may take many forms. The speed of the communications link limits the application available on any remote machine. Fig 1. illustrates the application limits for many computer systems (using known techniques) at a variety of communication speeds, and also demonstrates the application area of a PiCO device. Insert speed/application graph here. Figure 1: Application Graph (speed against device type) In the commercial world `Laptop Computers' are becoming particularly prevalent and there are also a growing number of products described as Palmtops and Personal Electronic Diaries. All of these have limited use in practice but if attached to a communications link providing direct access to online databases along with some of the facilities normally available though a workstation, the integrated system would allow computer facilities to be used in most o ce situations and possibly at home. Such a system has been called `Ubiquitous Computing' by Mark Weiser at Xerox PARC. The Xerox vision is that the computing environment can be extended into the walls and surfaces of an o ce, and be further enhanced by an array of portable computing technologies including electronic writing tablets and paper replacing technologies. It is quite possible that many of these devices may be carried in a pocket or attached to the outside of clothing. The `Active Badge System' developed at Olivetti Research Ltd (ORL) is an example of Ubiquitous Computing: the Active Badge is a PiCO device worn as an identity badge. The original application of the Active Badge began as a location aid for a receptionist in an organisation, providing up-todate information about the physical location and the nearest telephone available to all sta in that company (even when distributed over a number of di erent sites). Although this application is of considerable importance in itself, when these facilities are combined with a computing environment there are many other application possibilities. A distributed computer-system can use an Active Badge network to obtain location information, and as a result, Workstations and Network services can take the location of personnel into account when performing tasks within the environment. The Active Badge also opens up other possibilities for system integration. All o ce equipment can potentially be designed to respond to the signal from a badge and would then register the identity of the person operating it. In itself this is useful when accounting for resources but it also means that electronic equipment can customise itself to the settings last de ned by that user. An o ce full of equipment, all of which is con gured to a users way of working may well be a feature of `The o ce of the future'. 1 System Communication Portable computing objects fall into three communication catagories: transmit-only, receive-only and transponding devices. The Active Badge falls into the transmit-only catagory. It is clearly desireable that all PiCOs should have the ability for two-way communications, however by restricting the communication to one direction in some cases can still result in a useful device but with the advantage its size and power-consumption can be reduced. 1.1 Simplex Devices (transmit-only) The Active Badge [6] operates as a beacon, regularly signalling a unique code to a network of sensors distributed around the area to be monitored. Location information is gathered by using a master-processor to poll the sensors through the sensor network. Unique codes that are periodically signalled by the badges and bu ered in the Sensor units are returned to the master, the name and location of the badge carriers can then be ascertained by looking up the badge ID, to determine a name, and the station address, to determine the position. The badge was designed in a case 55x55x7mm with a weight of about 40g. Pulse-width modulated infrared (IR) signals were used for signaling between the badge and sensor [3], mainly because: IR solid-state emitters and detectors can be made very small and very cheaply (unlike ultrasonic transducers). They can be made to operate with a 6m range and the signals are re ected by walls and are not particularly directional when used inside a small room. Moreover, the signals will not travel through walls unlike radio signals that can penetrate the partitions found in o ce buildings. Infrared communication has been used in a number commercial applications ranging from, the remote control of domestic appliances to, data back-up for programmable calculators and personal organisers [4], and at the more novel end of the market IR-based local area networks [5]. Because IR technology has already been exploited commercially, it is inexpensive and readily available for developing new applications such as the `Active Badge'. An active signaling unit will consume power; therefore the signaling rate is an important design issue. Firstly, by only emitting a signal every 15 seconds the mean current consumption can be very small with the result that `badge sized' batteries will last for about one year. Secondly, it is a requirement that several people in the same locality must be detectable by the system. Because the unique signals have a duration of only one tenth of a second there is, approximately, a 1/150 chance that two signals will collide when two badges are placed in the same location. For a small number Figure 2: The ORL Active Badge Design of people there is a good probability they will all be detected. Even so, to improve this chance the beacon oscillator has been deliberately designed around low-tolerance components: it is very likely that two badges, which by chance are synchronised, will have slightly di ering frequencies and lose synchronisation in a few minutes. A disadvantage of an infrequent signal from the badge is that the location of a badge is only known, at best, to a 15 second granularity. However, because in general a person tends to move relatively slowly in an o ce building, the information the Active Badge system provides is very accurate. The Active Badge also incorporates a light-dependent component that with reduced lighting increases the period of the beacon signal to an interval greater than 15 seconds. In ambient lighting conditions for a room this e ect only slightly modi es the period, but adds su ciently random components to the beacon period to remove badge synchronisation problems. However, in a signi cantly dark room e.g. at night, or in a closed drawer; the period increases to a point where the badge is e ectively turned o . If the badge is placed in a drawer out of o ce hours, at weekends and during vacation, the e ective lifetime of the batteries is increased by a factor of 4. The more obvious solution of a `switch' was considered a bad idea as it is likely that a `badge user' would forget to turn it on. Other options for switching the device on included a tilt switch and an accelerometer although the size limitation of a badge precluded using these techniques in the initial experimental system. The initial application of this system (the demonstration system) has been designed as an aid for a telephone receptionist. The system provides a table of names against a dynamically updating eld containing the nearest telephone extension and a description of that location. The format ts onto a standard PC display and is shown in more detail in gure 3. A third eld shows the likelihood of nding somebody at that location in the form of a percentage. If it is below 100% it indicates the person is moving around, and if they have not been sighted for 5 minutes it displays the last time and location at which they were sighted. The last sighted location is still the best clue a receptionist may have to locate somebody and indeed there may be other work colleagues in that area who will know why that person is no longer there. Beyond 24 hours the last day a badge is sighted is shown in abbreviation and if there are no sightings detected for a week or more, the person is indicated to be `AWAY'. This format was found to be useful and did not overload the display with too much information. In addition to the display a command interpreter allows simple investigations to be performed on the system. A simple extension to the receptionist's system has been to distribute the badge-display through the network ling system at ORL, allowing all lab members to use the location mechanism. By pressing a badge's test-button its signal can be manually triggered and hence its signalling rate will be increased (a mechanism call double clicking). It is this process by which a command can be initiated by a badge carrier. At rst such a mechanism seems extremely limited, however because a command can be executed at di erent locations or even with di erent people present its e ect can vary and be completely context sensitive. If other portable devices were developed with more buttons but were still the size of a badge, the command capability would be considerably improved, and yet still achieve a relatively simple interface. It is the simplicity of this command interface combined with location context that is believed to be important in making these devices useful and accepted. The exact e ect a command from a PiCO can also be de ned by a pro le edited within the usual workstation environment. It is here that these commands can be bound to the various command signals with the appropriate context attributes. Examples of context-sensitive (location, people present and time) applications include the following: