KCs Project Book

Even before the days of cyberpunk and William Gibson’s Necromancer, in which fantasies of virtual reality and wearable (or implantable) computers were popularized, there were visions of a future that included computers worn on the wrist or glasses that could be used to videoconference with others with a single voice command.  These original rough concepts of wearable devices have already come to be possible and some of them have even been marketed to the public. Now that we can technologically accomplish many of these “science fiction” endeavors, we find that the reality of what we had hoped to accomplish is not always what we expected.  There are still many shortcomings in creating useful devices that are an asset to humans interacting with their environments. Programmers, system designers, system architects, and visionaries are now left to conceptualize, analyze, develop and iterate where this type of technology will go.

Although many would consider wearable computers to be a niche product, we find an ever-growing population carrying “always-on” devices that could each be classified, to some degree, as a “wearable”.  By examining some of the common uses of wearable systems and their architecture we can examine some of their strengths and weaknesses and incorporate improvements into basic designs for a consumer, everyday wearable computer.

F. Tufail in “Wearable Wireless Body Area Networks”, describes wearables as “…body worn computer that (that) are always on, ready and always accessible” (Tufail & Islam, 2009). He goes on to describe that two of the most important considerations with these types of devices are that they are always on and ready to use and that they can handle doing multiple processes at one time (Tufail & Islam, 2009).  Although it often goes unspoken, a wearable computer needs to be an asset. In other words, it needs to accomplish some useful function and, as far as user experience is concerned, the positives must outweigh the negatives. This is important to consider because wearable computers are often specialized for different functions and environments and this can greatly affect what will be perceived as an asset and what will be perceived as a liability.  A wearable computer designed for everyday use might vary greatly from one that is designed for intense military use. By examining some of the existing systems, we can glean what some of their pros and cons would be if they were placed into a different scenario.

Two wearable computer systems currently on the market are the GD300, used by the military, and the MX3Plus, used by warehouse employees.   The GD300 runs on Android and uses 600 Mhz ARM A8 Cortex processor. It is equipped with 256M of ram and has a 3.5 inch resistive touch screen display. The device is worn on the wrist or chest and its primary focus is to display GPS related movement data and communications (General Dynamics Itronix Corporation , 2010).  The MX3Plus is focused on inventory and warehouse related functions. It too is worn on the wrist. It runs windows CE on a 400Mhz PXA255 processor with 128M of RAM.  Its screen is 6.5 inches and it has a keyboard for information entry (EMS Technologies, Inc, 2010). Both systems support 802.x wireless, Bluetooth and additional types of communication.

Both systems provide adequate function for their intended environments but lack some basic functions when considered for use as an “everyday use” system. Neither system has enough RAM to run robust operating systems. A robust operating system like Linux would allow a wide range of onboard intelligence that has already been made available to the public to be leveraged.  Another weakness is that these systems power all their peripherals from the main power supplied by the unit. The problem with this is that heavy battery packs make the units less portable and as power is distributed through the entire system, it is often not used as efficiently as it could be.  Lastly both of these systems are awkward and bulky. Walking around daily with a computer strapped to ones arm would likely not serve the needs of most people.  So what would? Although the list would grow as peoples’ needs change, we can deduce some functionality that could be valuable to most who would use such a device. The system should be powerful enough to run software that provides the capability to run existing multi-tasking operating systems and process environmental and input data via software. Additionally, the system must be extendable via hardware (via Bluetooth or usb) and via software (by users), unobtrusive and provide a simple means of interaction.

Recent studies in wearable computers reveal the value in connectivity between sensors, peripherals and the primary unit.   One type architecture is introduced in “Kuka: An architecture for Associating an Augmented Artifact with its User using Wearable sensors” by Kaori Fujinami and Susanna Pirttikangas in 2008. Kuka is software and an RFID architecture that “…incorporates wearable sensors attached to a user’s body to associate the wearer’s identity with an artifact that is utilized at the time” (Fujinami & Pirttikangas, 2008).  This system associates users and peripherals to understand a user’s activities.  Although this paper provides valuable information in what is essentially a model for understanding “activity recognition”, I believe its most valuable contribution has to do with wearable requirements. Systems must be scalable and quick to respond (Fujinami & Pirttikangas, 2008).  In this system all the sensors interact directly with the processing unit and the system does not communicate externally. In contrast, “Information system architecture for wearable cardiac sensors personalization” reviews a system that gathers immediate data from sensors and peripherals but can gather additional information from external sources, as the case requires (Krupaviciute, Fayn, Rubel, Verdier, McAdams, & Nugent, 2009).   An important element to include in the design of a system will be the ability to connect to external sources so that the system can be dynamic and updatable.

Requirements:
System powerful enough to run current expandable, mult-tasking operating system that can leverage community development and growth
Easy to interact with
Wearable for everyday use
Seamless information is provided by system
Can connect to community driven changes? for updates and additional datasets

Prototype:
The most current release of the BeagleBoard-xM prototyping board has the power to accomplish prototype needs of exceeding the base requirements for a daily wearable system. The DM3730 ARM processor runs at 1Ghz and utilizes 512M of onboard DDR memory (BeagleBoard.org, 2010).  It additionally supports microSD cards. An 8Gig SD card will be ample space to provide storage for OS, programs and expandability.  For input to the device, we will utilize the Twiddler-chording keyboard for text input and additionally fabricate a detector that can sense input from finger movements to maximize user interaction which, as Xiubo Liang (et al.) speaks of in Motion-based perceptual user interface when they describe effective gestures should  “mimics human-to-human” interaction (Liang, Zhang, Xiang, & Weidong, 2009). For display, we will fabricate a basic monocular eyeglasses display unit by modifying a pair of wrap920 display by Vuzix. The computer will run Ubuntu 10.10.  The power unit will be a lithium battery system that utilizes CANON LI-IO BP-2L 7.4V 1500mA batteries.  The primary system requires 5V to operate while the additional voltage will power the display unit after a voltage divider circuit.

Conclusion:

So although I’m pretty sure I don’t look cool yet, It is a good start to the overall system.

The basic design of the Beagle-board and accessories is a robust enough platform to allow casual experimentation with wearable technology. The system, after adding a GPRS connection, 802.x wireless and Bluetooth connectivity, has ample capability to support community connectivity. A webcam needs additionally to be added to provide input for augmented information feedback. Additional software development is required to utilize the hand detector to provide simple input to Ubuntu.

Although additional design will need to occur for daily use of this system, it is a solid platform from which to launch a community project in open wearable technology.  Additionally, the weakest link is the display unit. Improvements in this technology should allow viewing of 1024×768 as to provide useful interface for viewing the web (See lab notes for additional information on design).

Works Cited
BeagleBoard.org (Ed.). (2010, Sept 14). BeagleBoard-xM Product Details.
Retrieved 11 25, 2010, from BeagleBoard.org – hardware-xM:

http://beagleboard.org/hardware-xM

EMS Technologies, Inc. (2010). Product Detail. Retrieved 12 1, 2010, from
LXE An EMS Technologies Company:

http://www.lxe.com/solutions/product.aspx?id=145

Fujinami, K., & Pirttikangas, S. (2008). Kuka: An architecture for associating
an augmented artifact with its user using wearable sensors. (pp. 154-161). IEEE.
General Dynamics Itronix Corporation (2010). GD300. Retrieved 11 26,
2010, from Product Information PDF: http://www.gd-itronix.com/upload/specifications/us/GD300_datasheet_080210.pdf
Krupaviciute, A., Fayn, J., Rubel, P., Verdier, C., McAdams, E., & Nugent, C.
(2009). Information system architecture for wearable cardiac sensors
personalization. 2009 14th IEEE International Conference on Engineering of Complex Computer Systems, (pp. 265-272).
Liang, X., Zhang, S., Xiang, Z., & Weidong, G. (2009). Motion-based
perceptual user interface. 2009 Third International Symposium on Intelligent Information Technology Application. 1, pp. 247-251. IEEE.
Tufail, F., & Islam, M. H. (2009). Wearable wireless body area networks.
2009 International Conference on Information Management and Engineering, (pp. 656-660).