Rowing simulator: how to measure or build

October 16, 2010 3:40 PM

Looking for a way to measure movement of simulated rowing. Thoughts/advice + possible engineering involved!

I'm working with a rowing simulator that is using actual water (think of it being that your paddle actually swings through water in a still pool) while on dry land.

1) I'd like to be able to measure how fast or slow I'm going.
2) Based on this, I'd like to gear up a computer to actually play video that moves faster or slower based on my rate.

Google isn't helping. I've looked at the rowing simulators on the market and items like the endless pool. Is there anything on the market that does something like this?

If there isn't, I realize that building this may (or may not) be difficult.) I've been thinking about how to accomplish this. And I'm sure I'd have to hire help - I'd probably put something here on jobs…but also over on make magazine.

I need to be able to measure the change in the way the water moves accurately with some sort of sensor. I'd probably need some general measurements of how the water is moving to determine speed.

Then a rough translation to dictate how many frames (or what speed) video moves forward.

Am I missing anything?

posted by filmgeek to health & fitness (9 answers total)

I thought the whole point of rowing was that the oar stays fixed relative to the water, and you use that to push off against. If you were to simulate that with the boat part not moving, then you have to have the water moving. Otherwise you aren't rowing, you're just creating turbulence in water which is a much different situation -- you can't measure the water velocity if you're making turbulence because it's flowing all over the place in all different directions and it won't correspond at all to actual rowing speed. In that case I would try to simulate measuring the water speed by measuring the angle of the oar and using that to estimate the position of the tip, and from that estimate what the water speed would be if you were actually moving. You'll also have to estimate a drag parameter because you'll need to simulate the boat slowing when the oar is out of the water.
posted by Rhomboid at 4:37 PM on October 16, 2010

Are you using this type of rowing tank, or something else?

I think that all you'll need is some kind of stroke meter (like this one, the same sort you'd find on an erg) and pressure sensors under your feet and on the blade handle. You could probably get away with just having sensors under your feet, but it'd be nice to have them on both your feet and hands -- it'll make it easier to evaluate whether you're using proper form. (Specifically, making sure that you're waiting until the end of the stroke to incorporate arm motion.)

Putting in pressure sensors like this would be a lot easier than gauging water flow rate. The flow rate depends so much on the geometry of the pool that it'd be a total pain to calibrate.
posted by leslietron at 4:37 PM on October 16, 2010

The rowers I knew seemed to measure their workout progress by ergs (that is by units of work). Does that help at all?
posted by maryr at 4:38 PM on October 16, 2010

It sounds like you don't need a well-calibrated measurement of your effort/exertion, you just want the visuals to roughly correspond to speed, right?

One easy way might be a thermal flow sensor in the pool— two thermometers, one of which has a heater attached to it. The closer their temperature readings are, the faster the water is moving. Attach them to ADC inputs on an arduino, do some seat-of-the-pants calibration to translate that to frame rate, done.

A strain gauge attached to the frame on which the oar+seat+footrest are mounted would also work, but would probably be a pain to DIY.
posted by hattifattener at 5:26 PM on October 16, 2010

I think that all you'll need is some kind of stroke meter (like this one, the same sort you'd find on an erg) and pressure sensors under your feet and on the blade handle.

I was going to say the same thing. Modelling the system for real world speed once you have pressure sensors on your hands and feet positions and maybe position sensors on the seat and the oar (maybe a linear pot on the seat and rotary on the oar at the pivot or similar) would do it.

Also, I think it's worth stressing that the water is a distraction - measuring the result of the work put in (as people have said) is easier and possible more accurate (especially through repeatabiity) than a restricted body of water (with odd currents and the like). Some of the static style rowing machines I have seen have graphical displays similar to that which you describe, so working out how they do it and modelling that rather than trying to work out what the water is doing is probably a better path to take.
posted by Brockles at 5:30 PM on October 16, 2010

My suggestion: Measure the forces exerted at the oar pivot. This will be the force moving the shell forward. Do a literature search to find some empirical estimates of C_d for a rowing shell. From this, you will calculate the force resisting forward motion. If you can find a model for aerodynamic drag too, then great, drop it in there, but I wouldn't count on it. Now that you have your forces, work out your equations of motion such that you know your velocity at any given time.

I can't really offer much re: the video display. I know pretty much diddly about free surface problems, much less what the current techniques are for simulating them for graphics (i.e., where we do not care about accurate physics). Good luck with that.
posted by indubitable at 6:56 PM on October 16, 2010

Crap! I forgot to mention, you'd also want to capture the effects of the rower sliding forward to the catch. It looks like some other people have mentioned measuring the pressure exerted through the shoes; I'd go with that.
posted by indubitable at 7:04 PM on October 16, 2010

Argh, I mean the force exerted through the shoes.
posted by indubitable at 7:05 PM on October 16, 2010

leslietron - very, very similar.

This is non fixed oar (no locks - think dragon boat racing) along with non sliding seats. So I can't reliably put anything on the oar itself.

Additionally, it's a 'group' effort - I'm not looking for individual members, rather the whole group rowing together.

hattifattener- I'm looking to measure some level of work - it doesn't have to be 100% accurate (or 70% for that matter) - merely accurate to itself; like scales that are accurate to a kg rather than to a gram. It can vary from reality - as long as it's self consistent.

The movement of video, is actually less difficult than the measurement of the water. Based on x (where X is speed per second), I can have QuickTime adjust a frame rate of prerecorded video. Then it just becomes having someone translate the value from the rowers to QuickTime's SDK to a preset video.
posted by filmgeek at 9:36 AM on October 17, 2010