Four thin and conformable haptic elements were attached using an adhesive to an armband made of soft elastic fabric, comfortable to wear. The actuators were driven using a custom made PCB, capable of generating various waveforms and apply voltages between 0 and 95V. The connection between the actuators and the PCB was made using flying cables.
After testing individual actuators at different locations on the arm, we were able to determine the locations that have the highest detection probability, even using low actuation voltages (<100Vpeak). The most sensible locations are situated around the elbow, especially the inner part of the arm seems to respond well to the haptic excitation. After this initial step, the actuators were placed on the armband on the inner part of the arm as shown in the figure below.
Then various signals were sent to the tester (5 persons):
- piezo 1 and 2 alternating rapidly
- piezo 3 and 4 alternating slowly
- piezo 2 and 4 simultaneously
- piezo 1 and 3 simultaneously
- activation order 1-2-3-4 repeated
The actuation frequency was on 250Hz and the voltage 95Vpeak. The tests were carried out in a lab environment, relatively noisy. The users (5, 3 male, average age 23 years) were not isolated during the tests, they did not wear noise cancelling headphones or any other device that reduced the noise level around them. They were seated or standing, with their arm moving occasionally. They were not forced to keep it from moving or to hold it in a prescribed position. They were asked to imagine that they are walking or running and were asked to act based on the interpretation they gave to each of the 5 signals.
All the users were able to detect the vibration from the actuators, giving an acceptable feedback level. The signals were designed to deliver the following messages to the user: 1 – accelerate, 2 – slow down, 3 – turn right, 4 – turn left, 5 – stop. However, we asked the testers to interpret the meaning of the signals freely. Some of them proved to be rather intuitive (signals 1 and 2), others needed input from the lab staff. The outcome of these tests is thus positive, as the actuators enable to deliver messages to users, and it shows that a learning session should be organized to make the messages clear.