Guest Ghostrider Posted June 5, 2004 Report Share Posted June 5, 2004 I was curious, I'm nearing the end of my college career here at Western Michigan University, and I was wondering if anybody out there knows if there is a tech company involved in the snowboard industry. Basically, how do I turn play into work?! :) In a few months, I will be graduating with a bachelors of science and engineering degree in Computer Science, with a minor in mathematics. Then in about 2 years from now, I will be graduating with a masters in Computer Science Theory and Analysis. If that confuses you, its mostly programming and software engineering. The three fields I enjoy the most are Computer Network/Internet Programming, Artificial Intelligence & Neural Networks, and Nano-technology. I can see where there would be a large need for engineers and CAD people in the industry (and I know there are alot of you on the forum), but what about us computer nerds?! If anybody has any connections to an independant tech company that works with a board maker or even a large board manufacturer that has a big tech department, I would love to hear about it! -Bill Quote Link to comment Share on other sites More sharing options...
knoch Posted June 6, 2004 Report Share Posted June 6, 2004 AH, just shut up and go build me a snowboard :D. I'd be happy to tell you if you're any good at it, then you can design yourself a web page all about how you did it, and how you could obviously do things way better the next time, and you can make the site all animated with cool pictures and graphs and crap, that auta keep you busy or somethin :p Quote Link to comment Share on other sites More sharing options...
lonerider Posted June 6, 2004 Report Share Posted June 6, 2004 Originally posted by Ghostrider I was curious, I'm nearing the end of my college career here at Western Michigan University, and I was wondering if anybody out there knows if there is a tech company involved in the snowboard industry. Basically, how do I turn play into work?! :) -Bill I was writing a post on how it's basically impossible and you should just be like Mike T and just live NEAR a mountain, earn a programmer's salary (which is good money), and just snowboard in the mornings and weekends. Remember that working in the snowboarding industry doesn't necessarily mean you get to snowboard a lot. Anyways, I just thought of something... the first is maybe one could design a system to track a snowboarder's speed, direction, and edge angle... this would allow racers to review their race runs more closely and analyze their runs more easily. My thought would be to place a device with a built in gyroscope on the board, that would you to measure edge angle and board orientation, couple that with a precise altimeter and/or GPS system (I'm not sure if currently technology is precise enough) and you would be able use the raw data to recompute the racer's path through the course, which would allow for very nice comparison between runs. Maybe if you could have extremely *precise* mapping of the race course, you might be able to do a more basic analysis using video data only (they already can superimpose racers over each other to compare runs) the trick is if you can figure out exactly what each rider is doing, you can see which "race line" is the best... you might be able to see more subtle things like when people are hitting turns and the pressure "roll" of their edge to see how they power through the turn... Eh... maybe it wouldn't be that useful... Quote Link to comment Share on other sites More sharing options...
jason_watkins Posted June 7, 2004 Report Share Posted June 7, 2004 It might be useful. In auto racing even a simple intertial data record can show all kinds of bad habits. You'd be surprised what a track day with a g cube can show you. If you could build a such a system for ski racers, you might have worthwhile market. Especially if you could synchronise an intertial record with some video analysis. Quote Link to comment Share on other sites More sharing options...
Neil Gendzwill Posted June 7, 2004 Report Share Posted June 7, 2004 Runners have devices which use accelerometers to measure speed and distance. They're about 97% accurate I think. 'twould be a neat project to apply that tech to skiing/riding. Runners also use GPS units for similar purposes but find they don't work well for instantaneous data. Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted June 7, 2004 Report Share Posted June 7, 2004 I smell a masters project a-brewin'... I have a background in robotics and basic Electrical Engineering, so I could probably rig up the sensors to collect all that data. And the programming knowledge to come up with a neat GUI to apply it. Only roadblock would be hill position..gps data would not be accurate enough...a 6 foot resolution is currently possible, however since the rider is going downhill and not on a flat surface, that margin of error is increased slightly...but still, hardly accurate. However, there are a few WLAN technologies available that could provide the accurate resolution, but they are all in the experimental stages...sounds like a nice challenge to me. and the testing parts sound like fun :-D Quote Link to comment Share on other sites More sharing options...
mirror70 Posted June 7, 2004 Report Share Posted June 7, 2004 Use differential GPS. Local accuracy of 3cm can be achieved that way - it's how large modern ships are docked now. The hard part I see if that you're measuring everything in a 6 DoF system with an ever-changing ground plane. How about a system to monitor board flex? I don't mean at a single point, I mean across and along the whole board. The next step in board design is to have an active dampening system, and the first step towards that is to figure out exactly wtf is happening with the board as you ride it. That's what this system would do. Quote Link to comment Share on other sites More sharing options...
Neil Gendzwill Posted June 7, 2004 Report Share Posted June 7, 2004 Originally posted by mirror70 Use differential GPS. The problem with that is getting updates fast enough to get detailed data. Thus the market in accelerometer-based running computers. They actually cost more than some of the GPS ones but because they do a better job with instantaneous readings and such many people prefer them. Integrating the accelerometer output properly to get an accurate speed/distance output with the device tied to a shoe is non-trivial - the work was done at the University of Calgary and is now being applied by several companies (including Nike I believe). Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted June 7, 2004 Report Share Posted June 7, 2004 JG had similar thoughts some years ago. Something in the board that would trigger a sensor going by the gate....... Thats kind of what I was thinking of using. With WiFi tech, you can do a local area GPS type of location by measuring signal strength rather than time offset. By placing a few access points so the antenna is at the stubbie, you could triangulate and get an accurate location. And since you would have a wireless network, you would eliminate the need for an onrider computer to collect data...just a PDA to relay the signals back to the main server at the bottom/top of the hill. Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted June 7, 2004 Report Share Posted June 7, 2004 The problem with that is getting updates fast enough to get detailed data. Again, using WiFi, your updates would be limited to the speed of the network protocol you chose to use. But most networks could support about realtime (about 10 readings per second) of a couple accelerometers, a location program, and about 2 onboard video streams (a helmet cam and a ankle cam...to see how the body positions are related). If you wanted to isolate data collection to reading one set of inputs, you could probably recieve at 1,000 readings per second or however fast the data could be collected. Quote Link to comment Share on other sites More sharing options...
jason_watkins Posted June 7, 2004 Report Share Posted June 7, 2004 gps is 3d btw... if you think about it, it *has* to be, since we're standing on the surface of a sphere and it's comparing 3 or more distances to locate a point (4+ in general for R3, but we can be pretty sure you won't be OUTSIDE the gps satalite array). The stuff the FAA has put in place for automating landings (WAAS) might be good enough to get real world positions for a snowboard or ski racer... but I wouldn't bet on it being a practicle product. There was a Phd thesis on snowboard physics that used differencial gps. Don't have the link here at work, but I seem to recall it was a chick at stanford. WiFi like 802.11 would be the wrong protocol... and I don't think the hardware is going to give you an accurrate and low latency read on signal strength... so it's probibly impossible to triangulate off it. I still think it wouldn't be all that critical to use real world locations. Even just accellometer info will give you a huge amount you can't see. Looking at video data along with accelleration, you get feedback for what impact each little movement had on your cg. If you _really_ wanted real world location, I think radio or ultrasound locators as used by robotics and motion capture would be the most reliable. But not cheap, and time consuming to set up. An accellometer rig could be as small as a pda tucked into a belt to track the rider's cg. The only other option I can think of is infrared reflectors with multiple cameras... as is also used in motion capture.... but you'd need plenty of cameras to cover a decent sized race course. Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted June 7, 2004 Report Share Posted June 7, 2004 I was thinking a location would be nice because you could mathematically calculate the fastest line through the course, then compare the riders with the calculated.. WAAS is essentialy dgps, although a real dgps i think is way overkill for the problem. It would be nice, but a bit excessive. It would be hard to get a correcting signal on a mountian unless you provided your own correction station within line of sight of the gps reciever on the hill at all times ($$$$$). Also, the 3cm dgps is very heavy on post processing, far from live data. Good for detecting if your cruise ship is drifting, but not for tracking a racer weaving through gates. a WAAS reading from a unit is less accurate (3-5m) but you only get an update about once per second. Again...trying to shave hundredths off of a racer, data once per second is not very helpful as that oculd be 2 turns in a slalom course and the reciever wouldnt even have detected any movement other than 12 meters forward. I'd like to see that dgps study if you could find it.. As for WiFi, agreed that 802.11x wouldnt work. a 100% signal can be achieved from a distance. Can't do a 'good' location (although it is possible, but has been implied in a building or on a campus...larger scale applications). However, bluetooth has a 10 meter distance, put one in each stubbie, and a signal could be calculated quickly and somewhat accurately. Final suggestion, automated video tracking. I did a neural network study and was able to calculate the x and y velocity as well as rotate a camera to a locked on target using only a cheap webcam and a servo controller. With an accuracy that was able to detect my breathing pattern by my chest movement. With a short calibration setup and proper camera location, it would also be able to detect location. Anybody see any problems with that one? Other than a brief loss of accuracy if the rider rode straight at the camera. But that could be solved by picking a good camera location. Quote Link to comment Share on other sites More sharing options...
jason_watkins Posted June 7, 2004 Report Share Posted June 7, 2004 Man, we're totally geeking out here. As part of a funny co-incendence, I bumped into this paper while browsing at work this morning: http://www.cs.unc.edu/~bennett/Proscenium/Proscenium.pdf Most of it is typical professional researcher fluff... but one nice consequence of relating video frames spatially is that establishing tracking is pretty simple... a lot less data entry. And it doesn't depend on image flow analysis or anything else that tends to fall down and go 'boom'. You only have to click enough corrispondence points to build a good spline through the 3d volume. Video tracking would definatley be the cheapest and simplist approach... but you'd be limited to analysis of pretty short course sections where the rider was visable from a single location. Hrmm... it might not be that hard to impliment. I'll keep it in the back of my head as a good project if I decide to learn how the video codec api works on windows. Edit: I'm not sure you *can* calculate the fastest line through gates. Even if you reduced the rider to just the cg over the board, that's 4d + the board/snow interface which is probibly horridly complex. Running an optimiation through that could be ugly. And in the real world, angulation and the rest of body positioning looks to be key, so you'll end up with many more degres of freedom. *Way* beyond math skills I can ever imagine myself having. Someone in that field like Hodgins might be able to do it. Quote Link to comment Share on other sites More sharing options...
mirror70 Posted June 7, 2004 Report Share Posted June 7, 2004 For positioning data, you could also use magnets under the snow/at the gates and an inductor on the board, or vice versa. With only two bases + the board involved, you could narrow down the location of the board to two points, and since you know which side of the gate the rider is on, you could eliminate one of them. This is a similar system to the one Cadillac developed for driver-less cars, and also to one I worked on for landing pilot-less helicopters. Quote Link to comment Share on other sites More sharing options...
lonerider Posted June 7, 2004 Report Share Posted June 7, 2004 Originally posted by Ghostrider I was thinking a location would be nice because you could mathematically calculate the fastest line through the course, then compare the riders with the calculated.. I'd like to see that dgps study if you could find it.. As for WiFi, agreed that 802.11x wouldnt work. a 100% signal can be achieved from a distance. Can't do a 'good' location (although it is possible, but has been implied in a building or on a campus...larger scale applications). However, bluetooth has a 10 meter distance, put one in each stubbie, and a signal could be calculated quickly and somewhat accurately. Final suggestion, automated video tracking. I did a neural network study and was able to calculate the x and y velocity as well as rotate a camera to a locked on target using only a cheap webcam and a servo controller. With an accuracy that was able to detect my breathing pattern by my chest movement. With a short calibration setup and proper camera location, it would also be able to detect location. Anybody see any problems with that one? Other than a brief loss of accuracy if the rider rode straight at the camera. But that could be solved by picking a good camera location. Hey, everyone is running away with my idea! :D I "thought" of the idea based on an article I read about the guy who designed the systems used by the NHL (infamous beach ball hockey tracker) and NFL (yellow first down line). He adapted the system from a missile tracking system developed by the military. The problem I see with video tracking is that for the previous two versions, the "courses" were measured to extremely high accuracy (less conveninent on a race course that changes everyday) and the cameras used had highly accurate placement, pan/tilt angle, and zoom information transmitted to the processing computer... of course we technically don't need to do this real time at first... but I think the setup required would be too expensive for non-spectator sports like ski/snowboard racing (except for the Olympics perhaps). It would just require way too many customized cameras setup with camera men for a normal race I think. If you aren't going to actually use WiFI/Bluetooth to transmit data, I would say use a plain jane radio signal instead. Broadcast it and have several receivers and have all those receivers dump the information into a central server for post-processing (then you just need to get updates at a sufficient speed). Same thing, much cheaper than WiFi and Bluetooth, with greater range, and probably higher sampling speeds. That will give you 3D position and velocity (more or less). Then you just need a device attached to the board with a mini-gyroscope to measure board direction and angle. I don't think you can compute a "race line" through the gates... the rider's height/weight, board size, sidecut, weather conditions, all come into play. Not to mentioned rutts in the snow and the "groove" created by other runs. I think the key is catching sloppy turns and other under-"pushed" turns. That is learnign to go smoother in tight sections and knowing when to open up and bomb on the sections that can hold turns better. Quote Link to comment Share on other sites More sharing options...
F9 Driver Posted June 8, 2004 Report Share Posted June 8, 2004 I have decided that I am no longer intelligent enough for this forum....... I would like to thank each and every one of you for hurting my brain....... Must...not....look...at.....screen.......big words scare Mongo...... Quote Link to comment Share on other sites More sharing options...
Jack M Posted June 9, 2004 Report Share Posted June 9, 2004 I would think racers would also like to track this. You could do it with gyro's. I've been thinking about such a device for years. The problem is, your market is limited to national teams, collegiate teams, and a few prep-school teams, and only those who can spend some serious cash to gain a little speed. And you'd probably have to travel to every one of them to help get them set up. Bye-bye profit. It would be fun though. Quote Link to comment Share on other sites More sharing options...
mirror70 Posted June 9, 2004 Report Share Posted June 9, 2004 Well, what if a mountain were to purchase and install the system, and then rent it the same way they rent hill time to teams? Many teams could then use the same system and also have a consistent set up of that system. Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted June 9, 2004 Report Share Posted June 9, 2004 I'm not a physicist, but you couldnt have a gyro without an accelerometer because you would need to simultaneously subtract the instantaneous g-force vectors at the point of the gyro in order to read the angle of the gyro, otherwise if you pull enough g's, the edge angle read will be zero (close to it). For super accurate measurements, that 'instantaneous' and 'simultaneous' are not 2 words that computer scientists will use very often because computers can only technically do one thing at a time...just a thought. But bottom line to this whole idea is that you arent going to get anything 100% accurate, but you could make a pretty good guess and try to compensate for error in processing. And it would still be a neat system...I'd buy one...I have an addiction for excessive amounts of "usless" data.:D Quote Link to comment Share on other sites More sharing options...
Jack M Posted June 9, 2004 Report Share Posted June 9, 2004 that might work, but you'd have to demand a piece of the pie each time the system is rented. and you'd still have to travel to every installation. which would be fun, but it won't give you an early retirement! Quote Link to comment Share on other sites More sharing options...
Jack M Posted June 9, 2004 Report Share Posted June 9, 2004 I'm not so sure g-forces affect the attitude of a gyro, but I don't know. What do they use in fighter jets? I'm sure you could buy something from an avionics mfr that would work. Quote Link to comment Share on other sites More sharing options...
lonerider Posted June 9, 2004 Report Share Posted June 9, 2004 Originally posted by Ghostrider I'm not a physicist, but you couldnt have a gyro without an accelerometer because you would need to simultaneously subtract the instantaneous g-force vectors at the point of the gyro in order to read the angle of the gyro, otherwise if you pull enough g's, the edge angle read will be zero (close to it). I agree with Jack in that I don't think G-forces will affect a gyroscope in the way that you are implying. From my understanding you simply point the gyroscope at say North and start it spinning, if you get it 2 degrees of freedom via low friction gimbals (sp?), it will always point "North" no matter how you move/rotate the object, even if you jar it. The housing rotates around the gyroscope and the difference in orientations between the two give you your direction. How are you envisioning the gyroscope as working? My guess is that an aircraft gyroscope will be too big and expensive. How about something like this? Update: Oops, the unit I listed is over $800... there probably is a cheaper version somewhere. Quote Link to comment Share on other sites More sharing options...
Jack M Posted June 10, 2004 Report Share Posted June 10, 2004 Originally posted by lonerider How about something like this? Cool! And it works even at 4 G's! I'm not surprised by the price tag. I would imagine the total package for a snowboard/ski race tracker would cost many thousands. Quote Link to comment Share on other sites More sharing options...
northcoast Posted July 30, 2004 Report Share Posted July 30, 2004 What about using 4 mini force plates for measuring heel & toe inputs? Then all you would have to carry is a meter & scientific calculator. ________ Body Science Quote Link to comment Share on other sites More sharing options...
Guest Ghostrider Posted July 31, 2004 Report Share Posted July 31, 2004 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. 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.