duPont Hospital for Children and the University of Delaware
Vol. 2, No. 2 -- Winter/Spring 1997

Auto-Vac Design Team

By Jane Fee

In the last issue of Design Team Update, one consumer talked about some of the difficulties he had while home testing the auto-vac. Since that test, the team has focused on the need for different interface devices for different types of disabilities.

To study the various interface devices a project using virtual prototyping, to customize a human/machine switch interface for disabled consumers, was started. A three dimensional living room environment is graphically modeled on the computer. A remote-controlled vacuum cleaner is modeled in this living room and is controlled by various external switches by consumers with different disabilities. Several members of the auto-vac design team tested various switches by performing target acquisition tasks. At the beginning of the test, the consumers/design team members were asked which switch they thought would work best for them. At the end of the test, 70% of the consumers/design team members were more comfortable with a different switch then was first chosen. The interfase project is still under development.

While the virtual prototyping project works on interfaces, another home test took place. This "home" test was done in the rehabilitation unit's carpeted apartment located at the duPont Hospital for Children.

Since the test lasted about 30 minutes (instead of a few days), many of the questions that are in the logbook for team members to fill out were not applicable in this case. Non-applicable questions included: set up time, approximate size of area cleaned, battery charge reading before starting, battery charge reading after completion, bag changing, etc. Tester and design team member Steve Chambers has athetoid cerebral palsy. In his case, arm movements are very erratic and hard to control. The joystick controls for the autovac were not the right kind of interface for him to use (but the only type we have for this prototype), so steering was a problem. Even though the controls were flush with the surface of the interface unit, Steve still hit the trim controls with his hand/arm movements. Because of the difficulty that he had with the joystick controls, the autovac would continue running while he tried to operate the controls and the autovac would overheat. When the autovac overheated the internal electronic controls would not respond correctly and the autovac would either completely stop or glitch many times and then stop.

After the autovac overheated several times, the test was stopped.

As stated above, this "home" test demostrates the need for proper matching of the right interface to the proper consumer. Stay tuned for further developments.

For more information, check out the Auto-Vac Design Team Home Page.