1999 Conference Proceedings

Go to previous article 
Go to next article 
Return to 1999 Conference Table of Contents


Neil Scott
Judy Jackson
Ing-Marie Jonsson
K. Takeuchi
E. Ito
K. Kondo

Project Archimedes, CSLI
Stanford University
119 Cordura Hall
Stanford, CA 94XXX USA
+1 650 723 3774
kouichi@csli.stanford.edu  ito@csli.stanford.edu 


In this paper, we describe the design and development of a network for universal access to devices called the Total Access System, or TASCloud. The Project Archimedes Team at Stanford had previously developed the Total Access System, or TAS, which was presented to CSUN98. The TASCloud is a network based on the original TAS. While the TAS was built to control computers, the TASCloud controls appliances on a local network. Appliances can be anything from televisions or telephones, as well as computers.


In 1990, the Total Access System was developed by Project Archimedes researchers to address the need for accessible computer equipment for disabled users. The TAS consists of a personal accessor, such as a PC running speech recognition for a mobility-impaired user, and a Total Access Port, or TAP, which provides connectivity to the targeted computer.

Accessors are tailored for a particular person according to their needs. A mobility-impaired person may use both a speech accessor and a head tracker, whereas a person with a progressive disease would use an eye tracker. All of these devices are accessors. The TAP links an accessor to the target computer, examples of which are the Macintosh, PC, Sun, SGI or Hewlett Packard workstations. The TAS has been in use for over 5 years by several hundred people. Existing accessors in use today are:


During user testing of the TAS, several design issues made a new design desirable. The TAS is connected by keyboard and mouse cables to a single computer. A user connects to a machine by plugging and unplugging the cables. However, people with musculoskeletal disorders (MSD's) were sometimes unable to unplug the cables. It became clear that a system that connects wirelessly was more desirable. A technology allowing seamless access to a workroom of devices to control was desired, but was not practical due to expense and lack of clear options in the technology.

Control of domestic appliances has become possible with LONWORKS (see the reference below). LONWORKS is a networking technology created by Echelon, using chips from Motorola. It is a peer-to-peer control network, meaning that a central server is not required to maintain the network, or to have any knowledge of a device before it is attached. It is therefore a quick task to configure and add a new device to the network of adaptive aids and controlled devices. Networks can range in size from two to 32,000 devices.

Physically, LONWORKS is local operating network consisting of intelligent devices, or nodes, that are connected by one or methods, such as Radio Frequency, Infrared, or home power lines. Nodes communicate with one another using a common protocol. Internally, nodes contain several chips for network processing and for interfacing with a device. Using LONWORKS as an interface between appliances, consumer electronics, and a user's personal accessor became a new possibility. The LONWORKS network was gaining ground in comparison with other technologies and in standardization efforts (see the reference below).


Several common design principles are shared between the TAS and the TASCloud.

Portability: The design goal of portability was intended, but has been frustrated until recently by the poor audio performance of smaller systems. Portability became more important when it was found during usability testing for the MSD Study that a smaller, more portable unit was desirable. Three of the seven users wanted to take the accessor portion of the TAS home with them, either to control a home computer, or to catch up on work in their own environment.

Separation of adaptive aid and tasks: The system must keep the adaptive sofware and hardware separate from the target machines, and must not require the target machines to be modified. Adaptive software, particularly speech recognition software, frequently conflicts with the software on a user's computer. The TAS must be a portable self-contained system for controlling a target computer.

Extensibility: The system must be easily extended along several lines. It must be easy to control new devices and applications, and to add new adaptive devices or software as needed. The current TAS allowed users to control new computer applications with the addition of a few scripts written by the user. However, adding new adaptive devices to the system was sometimes not easily done, since these were added directly to the target computer. The design of the TASCloud allows us to add devices to the network more quickly.


The TASCloud design has two phases:

1) Duplication of existing functionality on the LONWORKS network.

2) Addition of new network-driven functionality At the time of this writing, a prototype version of the first node had just been manufactured and was in testing. A Phase 2 prototype is expected in mid-1999. A study on users with musculoskeletal disorders will use the Phase 1 nodes, and another study with the U.S. Bureau of the Census will use the Phase 2 nodes. A usability study is planned for early 1999.


There are many systems to provide accessibility for people with one particular kind of disability. However, the authors are unaware of any existing systems that could provide a fully controllable environment, combined with a portable device that can be tailored for many types of disabilities.

Several networks for controlling home automation for disabled users have been explored (see the reference to the Emerging Technologies for Independent Living conference below) but the conclusions reached were that there has not been a good or affordable technology until recently.


1. Scott, Neil. Universal Speech Access, in Proceedings of Speech Tech/Voice Systems Worldwide, 1992.

2. Jackson, Judith A., A Portable Speech Accessor, CSUN Conference (Spring, 1998) Available at http://www-csli.stanford.edu/csli/~jackson/papers/CSUN.txt

3. Scott, Neil G., The Total Access System, CSUN Conference (Spring, 1998)

4. Emerging Technologies for Independent Living Conference, Baltimore, Maryland, 1994. Sponsored by The Center for Universal Design, School of Design, NC State University, Raleigh, North Carolina.

5. If Your Toaster Had a Brain, article on Echelon/LONWORKS, Wired magazine, Mar 1993.

Go to previous article 
Go to next article 
Return to 1999 Conference Table of Contents 
Return to Table of Proceedings

Reprinted with author(s) permission. Author(s) retain copyright.