The NXT Meets The Mouse... Datalogging Disney
So what do you do if you're taking a family trip to Disney World, and you're me?
Take your NXT of course! As a physicist, I've always been curious (read: obsessed) with things like the acceleration on roller coasters. So, on a recent trip to Walt Disney World in Florida I took an NXT with a field-ready datalogging program & a Hitechnic 3-axis accelerometer attached. In the picture at left, you can see the NXT with me on Expedition Everest... OK, actually what you see is a picture of me screaming like a little girl with one hand raised, as the other crushes the NXT into the seat beside me. But it's there, trust me.
Previously I had datalogged by just grabbing a reading and storing it in a file as quickly as possible. But for a 5 minute ride the resulting file would be large
- so large that I could only record one or two rides a day. Instead, the program stores the average value of the three axis data, with the sampling interval selected by the user at start-up. There were a few glitches, but overall this method worked great. If anyone wants a copy of the program & instructions on how to use it, let us know, and I'll try to pretty it up and post it (teachers, students - does this look like something fun to do?). Usually, I wrote to the datalog about every 100 ms, giving 10th of a second time resolution.
The result is information on not just the "max G's" that the rider pulls, but a more detailed description of the ride. In these graphs, the thick blue line is the total acceleration on the NXT, derived from all 3 components. The acceleration directly down into the seat is shown in light blue - normally this is your weight (the higher the trace, the more you are pressed into your seat). The acceleration forward or back is shown in red (the higher, the more you are pushed against the back of your seat). Finally the side to side acceleration is shown in yellow (as you are pushed against the side of the car in a sharp turn, the side to side acceleration will become larger in magnitude). So for instance as you are pulled up the 1st hill at a steep angle, you are pushed against the back of your seat, as shown by the red trace increasing to a higher level. As you dive through the bottom of a steep drop, the acceleration smashing you into your seat increases dramatically. It really is amazing the details you can see in the data if you know the ride (I've labeled a few interesting points... yes, perhaps I've ridden some of these 'coasters a few times :) ). I'm not going to give away any spoilers here... but if you know the ride, see what you can identify in that graph. The purple trace in this particular graph is from a different ride, showing how the acceleration varies if you are lucky enough to get the very front seat (worth the wait, actually).

I was actually able to datalog several rides during our stay: at Animal Kingdom, I logged Expedition Everest, Kali River Rapids (missed the big drop however), Primeval Whirl (a "crazy mouse" style ride, with some fun spinning and a few notable drops with low-G sections bracketing them), Triceratops Spin (with my daughters in control). Within the Magic Kingdom, Big Thunder Mountain showed some impressive "negative G's" (really just low-G points: as the data shows, you are never actually thrown up
from the seat), while Splash Mountain actually recorded the peak acceleration of anything I measured at Disney, hitting 2.5 G's at the bottom of the drop, and Space Mountain also revealing it's secrets. Finally in EPCOT, I logged Mission Space multiple times to make sure I got it (wonderfully smooth curve... and proof that even Disney hasn't quite got real "zero-G"), and my son was kind enough to do the brake testing section of Test Track.
If people want to see some of the other graphs, I'll pop them up in a further post. But mostly I just wanted to share with you some of the really cool stuff you can do with the NXT in an educational setting. Yes, that's right, I just claimed playing on roller coasters at theme parks is an educational setting. And I'm sticking by that. They are, when you can put the thrill rides into a context that people can understand, or get them to experience it in a new way - which you can do with the NXT, a simple sensor, and a program that a 10-year-old could write.

Now, I've just got to get back there to do some of the ones I missed. Like Tower of Terror (real "zero-G" on that one), Rockin' Rollercoaster (the initial acceleration should be very interesting), the Spinning Teacups (I love those things), & who knows what else. But for now, I'll leave you with a picture of a tired NXT, that has more than earned the right to a bit of rest.
Dang, it's hard work testing toys. ;)
Addendum: To try to keep things linked, I'm including the links to other datalogging blog entries here:
27 Sep 2007 Datalogging in a racecar
10 Sep 2007 Datalogging in NXT-G
14 Aug 2007 NXTlogger: a BT Datalogger
22 Feb 2007 Datalogging + Robotics
23 Nov 2007 Acceleration in the Community
--
Brian Davis
Previously I had datalogged by just grabbing a reading and storing it in a file as quickly as possible. But for a 5 minute ride the resulting file would be large

The result is information on not just the "max G's" that the rider pulls, but a more detailed description of the ride. In these graphs, the thick blue line is the total acceleration on the NXT, derived from all 3 components. The acceleration directly down into the seat is shown in light blue - normally this is your weight (the higher the trace, the more you are pressed into your seat). The acceleration forward or back is shown in red (the higher, the more you are pushed against the back of your seat). Finally the side to side acceleration is shown in yellow (as you are pushed against the side of the car in a sharp turn, the side to side acceleration will become larger in magnitude). So for instance as you are pulled up the 1st hill at a steep angle, you are pushed against the back of your seat, as shown by the red trace increasing to a higher level. As you dive through the bottom of a steep drop, the acceleration smashing you into your seat increases dramatically. It really is amazing the details you can see in the data if you know the ride (I've labeled a few interesting points... yes, perhaps I've ridden some of these 'coasters a few times :) ). I'm not going to give away any spoilers here... but if you know the ride, see what you can identify in that graph. The purple trace in this particular graph is from a different ride, showing how the acceleration varies if you are lucky enough to get the very front seat (worth the wait, actually).

I was actually able to datalog several rides during our stay: at Animal Kingdom, I logged Expedition Everest, Kali River Rapids (missed the big drop however), Primeval Whirl (a "crazy mouse" style ride, with some fun spinning and a few notable drops with low-G sections bracketing them), Triceratops Spin (with my daughters in control). Within the Magic Kingdom, Big Thunder Mountain showed some impressive "negative G's" (really just low-G points: as the data shows, you are never actually thrown up

If people want to see some of the other graphs, I'll pop them up in a further post. But mostly I just wanted to share with you some of the really cool stuff you can do with the NXT in an educational setting. Yes, that's right, I just claimed playing on roller coasters at theme parks is an educational setting. And I'm sticking by that. They are, when you can put the thrill rides into a context that people can understand, or get them to experience it in a new way - which you can do with the NXT, a simple sensor, and a program that a 10-year-old could write.
Now, I've just got to get back there to do some of the ones I missed. Like Tower of Terror (real "zero-G" on that one), Rockin' Rollercoaster (the initial acceleration should be very interesting), the Spinning Teacups (I love those things), & who knows what else. But for now, I'll leave you with a picture of a tired NXT, that has more than earned the right to a bit of rest.
Dang, it's hard work testing toys. ;)
Addendum: To try to keep things linked, I'm including the links to other datalogging blog entries here:
27 Sep 2007 Datalogging in a racecar
10 Sep 2007 Datalogging in NXT-G
14 Aug 2007 NXTlogger: a BT Datalogger
22 Feb 2007 Datalogging + Robotics
23 Nov 2007 Acceleration in the Community
--
Brian Davis
Comments
You are absolutley crazy - and I mean that in the nicest possible way. But a great job you done in the name of science! And seeing the fear in your face...
Last time I went on one of those things they made me empty all loose items (like camera) - so you done well getting away with carying a NXT in your hand (and return home with the NXT intact (and the logged data)!!
Well done
Tim
Admirable indeed!
Now it's a small step to martyrdom only (and eternal glory)! ;-)
What is: "Don't move a..." ...?
While I was looking over the slides I kept thinking how nice it would be to also see a video of the passenger during the ride while the graphs played along. It would be cool to see that correlation.
Humm...
Thanks Brian!
Chris
It would be nice to have a "video" of the ride to correlate... but you can have the NXT do that as well, to an extent. For instance, Expedition Everest goes in an out of a dark mountain, with could easily be logged by a light sensor set to "ambient" attached to the unit. A touch sensor would be handy as a "signature" in the record as well: push it at particular points of interest, so you know when you see the touch sensor depressed for the first time you're at the top of the first hill, and the second touch sensor press is at the start of the loop, etc. I was going to put on the sound sensor (to record noise level during the ride, to try to pick out the screaming portions)... and realized that the sensor would probably be maxed out the whole time, so didn't bother (this time).
Regarding "Don't move a..."... some people haven't ridden the ride, and I didn't want to give away spoilers unintentionally. So there are significant events in some of those logs that are only hinted at, or not even mentioned. In this case, you could actually figure out a lot by studying the traces around that time, to determine what you would be feeling at least.
--
Brian Davis
I'm sorry to say this, but your data acquisition technique leaves much to be desired. Be sure to notify me before your next trip and I'll come along in an advisory role.
Dean
As to the idea of using GPS to "map" the track, I suspect the biggest problem with that is how fast the GPS unit can update its location. However, by datalogging a compass at the same time as the acceleration sensor, you could probably do a rough estimate of where you are at what times. I think the time interval may be the limit here: the more sensor you try to poll, the longer it will be between logged intervals.
--
Brian Davis
"I was going to put on the sound sensor (to record noise level during the ride, to try to pick out the screaming portions)... and realized that the sensor would probably be maxed out the whole time, so didn't bother (this time)."
Might a noise filter be used? Something to cut down the maximum threshold to within the range of the Sound sensor. A sound-block-o-shroud!
Chris
--
Brian Davis
thanks!