Powered by Blogger

Thursday, January 19, 2006

Project Paddle Odyssey Technology Update

Project Paddle Odyssey is a not for profit based in Florida. Its objective is to design, build and deploy kayaks that can be paddled solo by a blind person. You can learn more about and hopefully make a contribution to PPO at http://www.paddle-odyssey.com. The following is the most recent technology update about the project.

Project Paddle Odyssey started in 2003 with a few ideas of my own and a number of others added by friends of the project. In the two years hence, we have spent a lot of time and energy getting our tax exempt status, getting the web site up and running, fundraising and getting the effort coordinated. As we put the organization together, we had little time to actually work on the technology.

I’m proud to announce that the PPO Proof of Concept v.01 software has been written and tested by a pair of graduate students working under my supervision. This version works quite well on land and has never been tested on water. We can successfully have one unit running in leader mode and a second unit in follower mode. The two units communicate their locations to each other and the leader tells the follower when and where to turn.

The History of PPO Technology

When we first started, the Paddle Odyssey activists envisioned a system that would run on a laptop computer, probably running Windows XP and JAWS, a terrific screen reader from Freedom Scientific. Then, as time passed, we started to work on a design that included a PAC Mate PDA also from Freedom Scientific but we discovered that, although a PAC Mate provided us with a nice form factor, it would be extraordinarily difficult for us to waterproof the device – an essential as the goal is to use the PPO system in a pretty wet environment.

We contemplated a home grown solution but that thought lasted about five minutes before we started looking hard at the iPaq from Hewlett-Packard and the Axim from Dell. These two lines of consumer products provided us with all of the computing power we would need and had wireless networking built in. Selecting a mainstream device also gave us a big economic savings as well as the ability to take advantage of the rapid upgrades that come from HP and Dell. The latest entries from HP even have GPS built into the unit so we can eliminate another part from our list.

So, the development challenge moved from a mostly hardware integration project to one that can be done primarily in software. The Proof of Concept v.01 runs on one iPaq and one Axim and, on solid ground, provides very good tracking information.

How It Works

The idea for the current proof of concept came up in a conversation between University of Florida Professor Sumi Helal and I. The most difficult problem we had at that point was trying to create an algorithm in which the leader/follower concept could work with an arbitrary number of followers. Sumi asked why we didn’t simply set up an ad hoc network among the PDAs – a feature built into Windows Mobile. So we did.

The current leader/follower system works by setting up an ad hoc network with the individual user logging in as either a leader or follower. Once a leader has logged into the system, no others are permitted to join with this status. The leader’s PDA broadcasts its GPS coordinates across the wireless network to the followers whose PDAs calculate the line between the follower’s position and provides them with compass points to tell them which way to turn. Currently, there are a number of talking compasses available for blind people so this system will work for the audio only version of the final system.

The proof of concept sends data directly to the speech synthesizer, a technique known as self voicing. Each follower may hear different information depending upon their location relative to the leader. In principle, the follower should then be able to paddle to the leader’s position. As the leader moves, the follower’s PDA will change trajectory and tell them to make a turn or correction in their movements.

What’s Next?

The people who lead PPO (our directors, advisors and volunteers) all share a common philosophy about assistive technology. Stated simply, use as much mainstream technology as possible and only build what we cannot buy (or have donated). Thus, our first proof of concept fits the model to a point, it uses off-the-shelf PDAs and GPS mouse units but the software remains self voicing and, therefore, far more difficult to maintain than if we used a screen reader. So, the next step will include rewriting the code so it can be read off the screen by Mobile Speak Pocket (MSP), a screen reader for Windows Mobile based PDAs, written and published by Code Factory, a software development company based in Spain.

MSP provides PPO programmers with an interesting feature – it uses Lua, an off-the-shelf programming language that provides programmers with access both to the screen reader as well as the operating system. It is still quite uncertain if PPO will build its final product in Lua but, for now, it looks like it provides us with the power and flexibility we need.

Other immediate steps include making the software aware of marine charts so the system doesn’t try to lead a paddler through an island and, of course, wet testing the software to determine how well it works in its intended environment.

Open Questions

Research has brought PPO to the point described above. More research will be needed to answer questions like, Is the paddler receiving too little or too much information? Can a blind person react quickly enough in a moving kayak to compass points in order to be an effective follower? Would a stereo headset that “pointed” the paddler to the leader be more effective or, conversely, would such a headset take away too many of the ambient sounds that can also provide useful information?

Commentary

Yesterday, Joe Clark, a notable web accessibility expert and my favorite Canadian curmudgeon, posted a comment that said that DVS audio can be found by playing the DVD and then striking “audio” until the DVS comes up. I use a Bose Lifestyles system at home and I don’t know if and/or where the audio button might be. So, if Mr. Clark sees this post, please add a little more description as to what hitting “audio” means.

1 Comments:

Anonymous Will Pearson said...

Nice to see PPO progressing well. As far as I know, it's the first of it's kind, in that it attempts to make a wider range of outdoor recreational activities accessible. Hopefully, where PPO goes more will follow, and I've got an idea for a similar project myself, if only I had the time.

"More research will be needed to answer questions like, Is the paddler receiving too little or too
much information? Can a blind person react quickly enough in a moving kayak to compass points in order to be an effective follower? Would a stereo headset that “pointed” the paddler to the leader be more effective or, conversely, would such a headset take away too many of the ambient sounds that can also provide useful information?"

I'll give you a call next week and pass on some potential starting points, as these are questions I've thought about myself.

4:03 PM  

Post a Comment

<< Home