IBM®
Skip to main content
    Country/region change    Terms of use
 
 
 
    Home    Products    Services & solutions    Support & downloads    My account    
IBM Research

Think Research


 


Featured Concept
Mobile Computing on the Move
COVER STORY: Mobile computing:creating new solutions

By Micheal Sinclair

Mobile Computing Makes Its Move

By Michael Sinclair

Several research efforts at IBM promise to extend the reach of mobile computing with innovations ranging from smaller devices with more functions to mobile middleware and applications.

In Brief:

The technology needed to realize the goal of fully mobile computing - ready access to all sources of information and the ability to work anywhere, anytime and on a variety of convenient devices - is rapidly coming together. Its emergence represents a tremendous opportunity, points out Peter Hortensius, senior manager, personal systems, at IBM's Thomas J. Watson Research Center. "A recent IBM taskforce," says Hortensius, "concluded that the global market for mobile computing, which stood at $75 billion in 1995, is expected to grow at an annual rate of 20 percent through the year 2000."

Over the past few years, Research has contributed to all aspects of mobile computing, from innovative features of the award-winning IBM ThinkPad to mobile networking, middleware and applications. Some of this work has been featured in previous issues and some will be the subject of future stories, as the field is rapidly progressing.

Already, says IBM Fellow Ted Selker, "We can put desktop computing power into a pocket-size, lightweight package." In fact, he and other researchers have prototyped several innovative mobile devices that stretch the conceptual envelope using technology that's essentially available today (see below).

Selker makes the point that technology has brought us to the point where a computer's dimensions and form are simply determined by its interface surfaces. "How I need to interact with the computer ultimately defines its size and shape," he says.

Certainly interface design has become an area in which exciting things are happening. Witness Selker's work on a desk-top grade control surface and viewing for a pocket-size, full-function PC. "We've already designed the keyboard and the display and it fits right here," he says, patting his hip pocket.

Pocket-size PCs have yet to be broadly successful. IBM ThinkPads are the current leaders in mobile computer technology, and IBM Research is working on several fronts to extend that leadership and make mobile users more efficient and productive. Several groups are working on advanced user interfaces, innovative wireless technology and new enabling software tailored to mobile needs. The following pages feature several of these advances.

Advanced User Interfaces

Paperlike Displays Flat-panel liquid-crystal technology, like the thin-film transistor active matrix displays on IBM ThinkPads, is evolving to higher resolutions. ThinkPads available today have 800 x 600 pixels, totaling about half a million pixels. A team at IBM's Thomas J. Watson Research Center has built a flat-panel array of 1280 x 1024 pixels. The result is a total of 1.3 million pixels in a 10.4-inch diagonal display.

"While pixel count has been used as a way of distinguishing between the resolution of displays, we now also consider the spatial resolution, that is, the number of pixels, or color dots, per inch," says Paul Alt, manager of display systems at Watson. "What we've developed is unprecedented in flat- panel displays: 157 color dots per inch (dpi) compared with 80 dpi in today's best displays."

The display, moreover, can represent 262,000 colors. "That, combined with high resolution, creates an extremely sharp image that approaches laser printing quality," says Alt. He is confident that his group soon will achieve 200 dpi resolution.

Where is the work going? "The goal is to make flat-panel computer displays that look like paper," Alt explains. Studies show that individuals make fewer mistakes and report better comprehension, higher productivity and less eye strain when they use paper, in part because the printed page has higher resolution than most computer displays. Paperlike displays will increase the scope of mobile applications. For example, mobile users with high-resolution displays will be able to easily read faxes without the need for scrolling and zooming; fine details will be fully legible in electronic maps; and architectural drawings and detailed maintenance diagrams for aircraft, automobiles and buildings will become immediately accessible as part of portable CD-ROM catalogs and maintenance manuals.

Lower Power, Longer Life In addition, the group is working to reduce the new display's power requirements. Indeed, lowering power consumption and thereby increasing battery life are key goals for mobile computing. In April 1996, following a decision of the Corporate Technical Committee, the Low-Power Systems Institute was established at the Tokyo Research Laboratory (TRL) with the aim of increasing battery life at least twofold. The choice of TRL reflects the lab's experience in this area. "We use a system approach," says team leader Akihiro Kuroda. "We first measure the power consumption of portable PCs in actual use to understand where and how power is being expended. Based on that knowledge, we can then make improvements in the individual hardware components, as well as in the overall system design and power management software."

As a result of their previous work, significant improvements were made to extend the battery life of the ThinkPad 760CD, which was announced in September 1995. The result did not go unnoticed. According to PC Magazine, the battery life of the 760CD was ranked number 1 among various Pentium-based, full-featured, super-VGA notebook PCs that had been announced by that time.

TrackPoint Technology TrackPoint, the innovative red pointing stub on each ThinkPad keyboard developed five years ago by Selker, Joe Rutledge and Bob Olyha at Watson, continues to evolve. They and their colleagues are working on several new ideas to make the TrackPoint even more useful.

  • Press to select. "As often as not," notes Rutledge, "a person meeting the TrackPoint for the first time will guide the cursor to a target icon and then push down on the TrackPoint as if it were a selection button. They then wonder why the icon isn't selected. We felt that, if users regard that as a natural way of using the TrackPoint, it would be a good thing to let them do it." "Allowing them to press with the same finger that they used to press the mouse button simplifies the task of getting people started with the TrackPoint," adds Selker. "We want to include this function to make the TrackPoint as simple to operate as possible." The result will be a TrackPoint that lets the user "point and pick" in a natural way, without having to change focus from fingertip to thumb and back again.
  • Tactile feedback. Fingertips are exquisitely sensitive and Selker's work on a dynamic TrackPoint stub gives a whole new meaning to human interface. "It bumps and vibrates depending on what is underneath the cursor as it moves over window edges, text, graphics and other display elements," Selker explains. Thiscould be a big help to mobile users who often have to access their ThinkPads under less than ideal conditions.
  • Two hands. Selker is obtaining interesting results with a two-TrackPoint keyboard, that contains one stub for the left hand and one for the right. "We're getting productivity improvements when the nondominant hand makes tool selections while the dominant hand does the accurate work," he says.

Innovative Wireless Technology

The Cordless Modem If you are a laptop user, you know the scene: before you can start an online session, you have to find a telephone cable, then crawl under the desk to plug it in. "I came to the conclusion one day in 1994 that I was really getting sick of that," says William E. Pence, senior manager, communications technology, at Watson. "I thought: 'why can't I just take a cordless phone and hook my ThinkPad modem up to it ?'"

It turns out that several reasons prevent you from doing that. Chiefly, cordless phones are designed to support voice rather than data. Nevertheless, Pence thought that he was on to something: a marketing need with no existing solution. "Cordless phones are made by the millions and are relatively inexpensive," he points out. "So we set out to adapt that technology to link laptops to the telephone wall outlet without wires."

Pence knew that it would be difficult, if not impossible, to handle the information density of today's high-speed 14.4 and 28.8 kbs modems over the inexpensive radio frequency links he wanted to employ. "The modulation is very complex and it is difficult to build a reliable radio link using components found in 900 MHz cordless phones," he explains. So he and his associates developed a fundamentally different approach to FM modulation. It achieves a very high signal-to-noise ratio to reduce signal problems within the 100-foot transmission range of the device. The resulting cordless modem prototype works on a plug-and-play basis - with any modem and computing platform, without changes in system software - and plugs into a standard RJ-11 telephone jack.

The biggest advantage of the cordless modem, which wowed attendees at the PC Expo in New York in June of this year, is that it handles anything that runs across plain old telephone copper wires. "The Internet phone, Internet video, emerging simultaneous voice and data modems, even videoconferencing, will run over this without any change," explains Pence.

Smaller, Cheaper, More Robust Infrared When it was introduced, infrared (IR) technology required a hump-backed PCMCIA card to accommodate electronics, light-emitting diodes, an IR- sensitive photodiode and deep pockets to pay its $300 price tag. Things have changed in the three years since then. IBM Microelectronics recently announced a new IR transceiver chip family, called Firefly, designed by Pence's team at Watson. Based on established BiCMOS technology and compatible with the Infrared Data Association standards - some of which were proposed by IBM Research - it has three unique features: a very small size, which makes it look like a licorice cough lozenge; a low price of about $5; and lower power requirements than previous devices.

The transceiver circuit designed by Pence's group and integrated on the chip handles three industry-standard IR transmission speeds: 115 kilobits per second, 1 megabit per second and 4 megabits per second. A major challenge was assuring that the device operates uniformly at all distances up to 1 meter. "It's like making sense of what I say if I scream in your ear at the top of my lungs or whisper from across the room," says Pence. Careful attention to circuit layout minimized noise generation, so that the transceiver reliably deciphers shouts or whispers.

Firefly uses directed or line-of-sight IR: it must be pointed roughly toward a receiver. "We think the natural evolution of IR will go in two directions," says Pence. "To higher speeds - 10-20 Mbs - and to diffuse technology that can increase range and eliminate line-of-sight restrictions."

While diffuse IR overcomes certain restrictions of directed IR, it, too, is subject to limitations arising from local conditions, such as fluctuations in the ambient light - which produce noise in the receiving device's photodetector - as well as obstructions and reflections. However, a group under the leadership of Fritz Gfeller at the Zurich Research Laboratory has developed the first wireless technique that enables an IR-based system to adapt its transmission rate to changing interference conditions, thereby ensuring reliable connectivity.

Their solution, based on a variable-rate transmission approach, adapts the transmission rate to the distance and noise levels. "We use a form of repetition coding," says Gfeller, "which means we repeat the symbols in the presence of noise, as well as in the critical part of a data frame, namely, the header, to mitigate the hidden terminal problem." That effectively reduces the throughput of information, while maintaining the quality of the transmission and avoiding collisions.

In order to develop a full-fledged solution based on this technique, the groups working on the transceiver and variable-rate coding are also collaborating with two departments at Watson that are developing protocols to support such IR communication. "For example," explains team member Mahmoud Naghshineh, "the protocols choose suitable repetition rates to accommodate a given noise level, or channel quality, and they adapt dynamically to the hostile wireless environment."

"They also handle the discovery and coordination process," says Parviz Kermani, manager, personal communications systems. If one or more IR systems are going to communicate, it is not only necessary that they discover what other devices are present, but also the suitable transmission rates between them. During the discovery stage, an IR modem assumes it's in an inhospitable environment and announces its presence as slowly, and therefore as clearly, as possible. The other devices reply, revealing their type - for example, printer, ThinkPad, PDA, etc. - and at the same time the quality of the channel. The multiple access protocol, explains Naghshineh, is able to take into account the multiple transmission rates concurrently, so that all the devices can be coordinated and only one pair are communicating over the common media - that is, IR light - at any moment.

New Enabling Software

More mercurial middleware The mobile computing environment turns all the conventional assumptions about system software on their head. Disconnection is the rule and brief connections occur over expensive, low bandwidth wireless links or over faster, but still slow, phone line modems. That has caused significant problems in the past. "Many network applications simply didn't work in a mobile environment," says Carl Tait, a software researcher at Watson. "And even the ones that worked were typically hard to use or very slow."

Soon, that will change. Robust software will adjust to the primarily disconnected operation of mobile users, help keep data in sync and in some cases automatically download applications and tools as needed by mobile systems.

Tait, along with Henry Chang and their colleagues, worked with IBM's Networking Software Division on a middleware solution called ARTour(TM) Wireless Web Express. Its aim is to make wireless Internet access more efficient and productive for corporations that use the World Wide Web to distribute information or handle transactions.

Web Express works with industry-standard Web browsers. When a browser skims the Web, it uses Hypertext Transfer Protocol (HTTP), the lingua franca of the Internet. "HTTP is extremely chatty and makes for very expensive sessions over slow wireless links," says Tait. "To speed and smooth communications between mobile clients and Internet servers, we are using proxies - software devices that intercept requests on the client side and responses on the server side. Although client and server think they are talking to each other, they actually are addressing proxies, which act as go-betweens to deal with wireless idiosyncrasies. Proxies, in the middle, don't mince words and use a very terse protocol to speed up wireless sessions."

Web Express, with client and server software now in beta test, includes caching and compression techniques to speed response times and a queuing mechanism that supports off-line and asynchronous form submission and Web request queuing. For even greater effectiveness the Watson team is developing features that watch users' patterns and save frequently-accessed URLs to reduce the frequency and duration of wireless connection. Chang is also leading an effort to create remote database access middleware for mobile users based on Java, an Internet-compatible object-oriented programming language. "Java is platform-independent and network friendly, making it suitable for future hand-held mobile computers," says Chang. "It is also a powerful tool for creating flexible mobile applications."

Form-filling in the field Some of Research's work in mobile computing may have a significant impact on a common business activity: filling out forms. In fact, this activity is a critical - and challenging - one for a tremendous population of workers who need access to enterprise data in the field and the ability to return data for subsequent processing. "To address this problem," says Larry Koved, a researcher at Watson, "we are looking at a different approach to user interface design and application development based on task modeling." Task modeling describes the information that needs to be displayed and captured at the user interface and the form in which it is visually organized.

Koved and his colleagues are working with IBM's Healthcare Industry Solution Unit to help streamline the gathering of patient data. "Taking blood pressure is an example of a task that has several sequential elements," he explains. "We interviewed nursing staff to better understand such tasks, and from their descriptions we are generating user interfaces for mobile form-based applications." The work with the Healthcare ISU will also benefit from the mobile middleware described above.

One of the goals of the project - known as the visiting specialist application - is to develop basic mobile-computing technology, user interfaces and application development technologies that can be applied to other industry sectors. "We are also creating 'digital ink' interfaces with which we can capture, store and display signatures and handwritten notes on tablet computers," says Koved, who believes that more effective multimodal interfaces using pens, speech and audio are also key to the success of future mobile applications.

Mobile Multimedia

MPEG video Odd as it may seem at first, mobile computer users are often the most sophisticated users of multimedia. In fact, the new ThinkPad 760 has audio, video and graphics capabilities that rival those of the best desktop systems. And, when it comes to digital video, ThinkPads actually exceed the best desktops. In the fall of 1995, a Research-designed MPEG-2 decoder chip from the IBM Microelectronics Division was introduced in the award-winning 760CD, providing faster frame rates, more detail and crisper images than is possible with all other computers that use MPEG-1. The ThinkPad is able to present digital video clips and movies using either MPEG-1 or the higher-quality MPEG-2.

Realtime video As mobile users gear up to make their first long-distance calls using Internet phone applications, IBM researchers are already perfecting the next step: an Internet video phone. "Realtime video over the Internet, intranets, LANs and radio links became feasible with powerful processors like the Pentium® and the PowerPC®," says Ephraim Feig, a video researcher at Watson. Key enabling developments include the realtime video compression algorithms developed by Feig's group and audio compression work at IBM's Haifa Research Laboratory. These innovative algorithms, which conform to basic video and audio standards, permit a 4.5 Mbs video data stream to be squeezed down to 20 kbs. That 220 to 1 reduction makes possible color video transmission at 8-12 frames per second over conventional telephone wires and computer modems.

Moving Into The Mainstream

Mobile computing is clearly evolving rapidly on many fronts. "Web Express, mobile solutions like the visiting specialist application, and related work in mobile networking and communications access are critical enabling technologies that are propelling mobile computing into the mainstream," says Hortensius. "By making mobile computing part of the existing networking and information infrastructure, a new level of usability will be reached, further transforming the way we work and communicate and blurring the distinction between mobile computing and any other form of computing."


More Information: The Bearable Lightness of Being Digital

The challenge in designing computers that are easily carried, pocketed or worn by users is to make them rugged, lightweight and reliable. ThinkPads, which feature desktop power shrunk to portfolio size and that weigh in at a tad over four pounds, are today's technology leaders. But they represent only a midway point in the evolution to the lighter, more functional, transparently interconnected mobile computers of the future. Here are examples of Research projects that are aimed at producing even more compact approaches to mobile computing:

Cyberphone

Surfing the net while talking on the phone? Answering e-mail one-handed? Too futuristic? Not really. Mobile users will do it soon. In fact, it's done with the help of mirrors and special user-interface software.

A team of Watson researchers - along with John Karidis, manager of new product exploration at the Consumer Division - is cooking up a prototype of a new kind of cyberphone.

The recipe calls for a cellular phone with data capability, to which is added a small computer. That combination is then married to a 1.4-inch-diagonal projection display that projects onto a specially designed mirror that is built into a flip phone cover. The result is a new mobile platform that provides a VGA-resolution display and computer in a handheld cell phone. While talking on the phone, the display appears to the user as if it were the same size as an 11-inch notebook computer screen at normal reading distance. A thumb-operated TrackPoint, located on the side of the phone, allows the user to readily navigate among the icons and information windows.

The device represents an extension of the voice and personal-information-manager (PIM) capabilities of a "smart phone," enabling connections to the Internet or Lotus Notes network so that users can see e-mail, faxes, voicemail, Web pages and other files. It also has the potential to connect to a speech recognition server by phone, which would allow a user to dictate notes or responses and then see the recognized text as it is returned from the server.

ThinkPad Companion

A related project is underway at the Tokyo Research Laboratory, where the aim is to develop a micro personal digital assistant (PDA). The goal, explains Kazuya Tago, manager of mobile software, is to find a way to implement a variety of functions - including the ability to access a server database and to initiate transactions on the server, as well as to view Web pages and read one's e-mail - with a smaller operating system than would normally be used. The compact device, which has less than 640 kilobytes of memory and a small LCD, can be operated with one hand and is also equipped with IR capability to connect to either a ThinkPad or a desktop computer.

Electronic Wallet

Combining a cellular phone, beeper, personal information manager (PIM), scanner and electronic credit card in a single device is a very challenging problem. But a team at the Almaden Research Center is stepping up to it with its Electronic Wallet project. It was designed, says IBM Fellow Ted Selker, the project's leader, to replace the wallet in one's back pocket or purse. The concept includes a display viewable from the outside that identifies who is beeping the owner. The PIM keeps track of one's finances, calendar, addresses, medical records and auto repairs; it also includes a photo gallery of the family. An onboard scanner records receipts and business cards. Individuals can record voice memos at any time. When open, the trifold device is a cellular phone. A removable electronic credit card turns into your plastic of choice (a 10-minute time limit makes it useless to anyone else should the owner lose it). And, yes, the device still has a pocket for paper currency.




    About IBMPrivacyContact