Tuesday, February 24, 2009

2/23/09 meeting at Touch Graphic

Steve, Zach, Annette, Ellen, Rob (by phone), and Gene met for a status and update session at the new and improved Touch Graphic’s office space on West 38 Street for approximately 2 hours. Steve announced that we now have a 30 foot area in the office to conduct tests on a WiiCane course. Zach reported that we should have a “working system” with fairly accurate output within a month.

The following details the highlights of the meeting:
* Zach demonstrated the limited audio capacity of the Wii controller. Speech and sounds emitting from the controller are both low in dB and in quality, and speech was unintelligible. Three possible solutions were proposed: wireless headphones that could deliver high-quality speech output, directional sounds, and possibly (low frequency, higher volume) sound that could be received as vibration on the head by deafblind users; a belt-pack-system that could emit speech and sounds and vibratory (rumbler) output when prompted by the controller; a collar-neck pillow that could house an audio-vibratory device(s).and possibly ear buds. Auditory occlusion, positives and negatives for the training device, was discussed. No final selection made. Gene will investigate possible headphone devices.
* Zach demonstrated the limited but detectable vibratory output capacity of the Wii controller. The vibrations were perceptible when the controller was attached to a cane with tape. If this is used, the output would be presented when the cane is not touching a surface.
* Zach demonstrated the infrared sensing capacity of the controller. The controller camera can locate “blobs” of light sources (up to four) and display/plot them on the laptop screen. We discussed the use of a single central strip of light-emitting material on the course, versus a set of two strips defining left-right boundaries. These would be used to detect veering, and as a possible source of data to calculate other variables.
* Testing in Kalamazoo to verify the capabilities of the WiiCane system will probably occur in early June 2009. Arc width detection and veering detection will be the two variables verified. Rob is handling the IRB at WMU; the in-house IRB process is being handled by Annette and this will be delayed until after the WMU tests.

There will be an advisory board meeting soon, probably at the end of March, before an April 2009 report is due to the funder.

If attendees have corrections or additions, they would be appreciated. Likewise, questions and comments are welcomes.



sl said...

Gene, excellent and concise summary. My only comment is that I am not sure what you mean when you say that the vibration stimulus will only be presented when the cane is not touching a surface. In the current demonstration, Zach programmed the wiiMote to vibrate on observing a very significant acceleration event, such as when the cane tip strikes a surface. But this threshold value for this can be easily adjusted, and determining the best way to set up the system will be determined experimentally.

GB said...

The discussion we had indicated that the vibration from the controller could never compete with the vibration caused by a cane striking the surface. Zach indicated that the vibration would be used when the cane was off the ground and not at a lateral apex. This was clear. For further info I suggest we inquire with Zach.

zeveland said...
This comment has been removed by the author.
zeveland said...

A little clarification on the vibratory feedback: the transmitted shock of the cane striking a surface will dwarf any sensation from the remote. Leaving aside the question of what we'll be able to communicate to the user with vibratory feedback and what form that feedback may take, vibration from the remote is only usable when the cane tip is in the air.

The demonstration last night was a simple example of threshholding and the response was pretty much the exact opposite of how we'll ultimately be giving tactile feedback to users.

Also, a correction to the original post: while we should have a working system within a month, that early system won't be particularly accurate. Accuracy is going to take extensive testing and tweaking of the system.

GB said...

egarding the concept of using the headphones for output/feedback for deafblind subjects . . . I discussed this with Bapin. He likes that the feedback could be directional. He thinks it will work, but has not tried a similar system. He agrees that loud/low frequency is most likely to work. He said, "I am sure it will serve the purpose you described without much difficulty."

I'd like to test this out and will be looking for appropriate equipment.

sl said...

Gene, let's begin looking around for some appropriate wireless headphones. Can you let me know the make and model of the ones that you have tried out and like? We should probably buy several different ones and test them for things like range, quality of audio, occlusion of ambient sounds, etc. We have money in the budget to buy three or four. The quality the user experience with these headphones will be very important to users' overall satisfaction with the system, so we should find the best one for our purposes.