NateW Posted July 31, 2004 Report Share Posted July 31, 2004 There are gyros now using piezo and MEMS technology, and they're far better than mechanicaly gyros - they're considered solid-state devices, though strictly speaking there's an itty-bitty oscillator in there. Radio controlled helicopters use them to stabilize the yaw axis, which is ridiculously slippery without electronic stabilization. Prices for commercial products are as low as a hundred bucks so I'm sure the sensors themselves are much cheaper. RC helis used mechanical gyros when I got into the hobby (with little penny-sized flywheels!), but the solid-state devices are far better. It's like cheating. If you put a camera on the snowboard and had software to recognize the horizon and the gates, would that give enough location information? I'd just like to see three-axis accelerometer data from the center of the board. Not because I think that would be useful, just because I'm a geek. Longitudinal acceleration would probably be negligible, but how many G's am I pulling in my turns, and how much weight do I put on the sidewall? Cool! Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted July 31, 2004 Report Share Posted July 31, 2004 NateW: I've done research studies in school here about camera technologies like you are suggesting. Using Neural Networks would by far be the preferred method of going about a software solution for that. However, that is based around a large array of "neurons" or threshold logic units, so it takes them a while to make a decision on whether that thing that flashed by was a red or blue triangle or just a shadow on the snow. Those programs work well in pattern recognition (autonomous vehicles finding the center of a road) or detecting shapes on a static background (IDing people walking through the gates while getting on a subway). But since cameras have no depth perception they dont know what is moving and what is a background once the camera is no longer static. However, I think I've devised a way to gather all the data you could ever want. some sensors are already built, but none of the circuits or interfaces...all would need to be custom. I've got the know how...so I guess Ill just have to start by collecting small parts and running tests to verify my designs and theories. Here are the features I included in my design: -Custom made flex sensor array (Board flex patterns) -3-Axis Accelerometer (G-Forces) -Barometric Altimeter (vertical speed/orientation) -Rotation compensated compass (board angle and rotational orientation) -Data aquisition sent via 802.11 WiFi networking with up to 100 readings per second to main computer for processing -802.11 WiFi video camera on nose and tail (visually monitor binding behavior and boot flex). -Rider EKG (Measure heart activity) -Rider Breathing Monitor (Measure breath patterns) Using the above data, I could then write a program to show a live wire diagram of the board during a run on the server computer. It would also log all data and compute various numbers to represent vital stats of the performance in the board/rider that could be reviewed later. Project price: about $2,750 in material costs + a couple hundred hours in construction, programming, and testing:D Heh...any investors interested? Quote Link to comment Share on other sites More sharing options...
bobble Posted July 31, 2004 Report Share Posted July 31, 2004 Originally posted by Ghostrider any techies out there know where to get a board (circuit...not snow) that can convert analog inputs from sensors into a serial or usb output (digital)? I guess 8 would work, but 16 ports would be best. I looked at the ezio board but I dont know if its what im looking for since i cant find any documentation on it other than art students talking about how to use it for art applications. try dataq.com. Quote Link to comment Share on other sites More sharing options...
jason_watkins Posted August 2, 2004 Report Share Posted August 2, 2004 Originally posted by Ghostrider NateW: I've done research studies in school here about camera technologies like you are suggesting. Using Neural Networks would by far be the preferred method of going about a software solution for that. There's a whole class of algorithms from the computer vision people that work very well on that sort of problem. You can find the math under the topics "image flow analysis" and "motion compensation." It's the same basic technology as mpeg compression for a dvd, or soft-ware assisted slow motion like the bullet time effects in the matrix. I like the camera on a board idea, but my guess is it wouldn't give you any data that more useful than the comments of a good coach just eyeballing a normal video. Quote Link to comment Share on other sites More sharing options...
Aisling Posted August 3, 2004 Report Share Posted August 3, 2004 Originally posted by Ghostrider I have an addiction for excessive amounts of "usless" data.:D that's called 'surveying' right, GR? ;) Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted August 4, 2004 Report Share Posted August 4, 2004 haha..thats right! (as I read this post in my car using one of 12 internet connections aquired by "mobile community wireless microwave surveying":D ) heh...ph34r m3. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.