Drift data logging? - Page 3 - Drifting forum

PhatBob · 21-04-2008, 13:28

Surely the way that you would eliminate the huge number of variables would be to employ fuzzy logic controllers.

Back when I was studying A.I. I read some of the reports on a project which was using doppler and IR sensors to make a system that would allow cars to autonomously brake and accelerate without the surging effects as the cars react to each other. The control logic with fuzzyification was frightening and I only saw a brief overview of it.

If you consider the level of control that the DARPA challenge vehicles have:
Welcome
Then its possible to simulate a huge number of factors in real-time, possibly even drifting, but it would be far from easy, and you'd probably need a passenger seat made of SHARC arrays to get close to the processing requirements.

chip-3door · 21-04-2008, 13:37

Bob, indeed, fuzzy logic would certainly be one way of doing a full on controller, however, I have no intention of doing that anyway, I dont have the resources to do so.

PhatBob · 21-04-2008, 13:57

So you're not wanting to do a controller, I'm guessing that you've got Matlab and want to build a model then?

chip-3door · 21-04-2008, 14:08

Nope, not wanting to do anything at all yet bob, just see if there are any patterns that might lead to someting useful, and any electronics we do is all from scratch on our own hardware, I dont get involved in the hardware side of things though mate so couldnt tell you much about the details of that side of it, im purely a software expert only.
Its just a hobby to me, but to the other guys its commercial, so I find stuff I enjoy playing with, and if they think there is a potential market they'll get involved.

At the moment we are still finishing off a basic launch/traction and full throttle shift unit anyway, and got a unichip style device (except it sits after the ecu not before it) that we are doing after that which is designed to allow easy turboing of a non turbo car (by allowing ignition retard and extra fuelling to be added to the existing setup)

So this stuff really is just purely "out of interest" only for me at this stage.

T.K · 25-04-2008, 16:34

Strange thread.

V-MAX · 26-04-2008, 08:57

You could do it by using a Datron Micr-eep datalogger to record pulses from front and rear wheel encoders and an RT3000 Inertial Navigation system to measure slip angle, lateral acceleration etc or remotely with a V-Box (drift box option).
Datron also produce Measurement steering wheels that can independantly perform slalom (elk) tests that cope with slip angle etc. You can hire all of this kit but it is expensive.

What you need is a friendly automotive engineer to give you some data, I'm sure they would have all you are looking for from testing ESP systems.

Although interesting as a project I agree with most other people that drifting is about having fun with no electronic aids and usually when I get into a car the first thing I do is switch off ESP!

Good luck with it

chip-3door · 30-04-2008, 01:28

Thanks for that V-Max, and im sure you are right about how most drifters would sooner not have electronic aids anymore than they would want hiv aids, lol

PhatBob · 30-04-2008, 08:46

VMax has a good point, it would be simple enough to use the ABS sensors to measure wheel speeds, then three Analog Devices accelerometers one at the front of the car, one midships and one at the rear. Steering input is a known with a rotary encoder or pot.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

21-04-2008, 13:28	#41 (permalink)
PhatBob Mr Sir Join Date: Aug 2004 Location: Newmarket Posts: 1,560	Surely the way that you would eliminate the huge number of variables would be to employ fuzzy logic controllers. Back when I was studying A.I. I read some of the reports on a project which was using doppler and IR sensors to make a system that would allow cars to autonomously brake and accelerate without the surging effects as the cars react to each other. The control logic with fuzzyification was frightening and I only saw a brief overview of it. If you consider the level of control that the DARPA challenge vehicles have: Welcome Then its possible to simulate a huge number of factors in real-time, possibly even drifting, but it would be far from easy, and you'd probably need a passenger seat made of SHARC arrays to get close to the processing requirements.

21-04-2008, 13:37	#42 (permalink)
chip-3door dorifto kingu! Join Date: Sep 2007 Posts: 900	Bob, indeed, fuzzy logic would certainly be one way of doing a full on controller, however, I have no intention of doing that anyway, I dont have the resources to do so.

21-04-2008, 13:57	#43 (permalink)
PhatBob Mr Sir Join Date: Aug 2004 Location: Newmarket Posts: 1,560	So you're not wanting to do a controller, I'm guessing that you've got Matlab and want to build a model then?

21-04-2008, 14:08	#44 (permalink)
chip-3door dorifto kingu! Join Date: Sep 2007 Posts: 900	Nope, not wanting to do anything at all yet bob, just see if there are any patterns that might lead to someting useful, and any electronics we do is all from scratch on our own hardware, I dont get involved in the hardware side of things though mate so couldnt tell you much about the details of that side of it, im purely a software expert only. Its just a hobby to me, but to the other guys its commercial, so I find stuff I enjoy playing with, and if they think there is a potential market they'll get involved. At the moment we are still finishing off a basic launch/traction and full throttle shift unit anyway, and got a unichip style device (except it sits after the ecu not before it) that we are doing after that which is designed to allow easy turboing of a non turbo car (by allowing ignition retard and extra fuelling to be added to the existing setup) So this stuff really is just purely "out of interest" only for me at this stage.

25-04-2008, 16:34	#45 (permalink)
T.K kudo fruiting Join Date: May 2005 Location: London Posts: 1,619	Strange thread.

26-04-2008, 08:57	#46 (permalink)
V-MAX doughnutter Join Date: Aug 2006 Posts: 39	You could do it by using a Datron Micr-eep datalogger to record pulses from front and rear wheel encoders and an RT3000 Inertial Navigation system to measure slip angle, lateral acceleration etc or remotely with a V-Box (drift box option). Datron also produce Measurement steering wheels that can independantly perform slalom (elk) tests that cope with slip angle etc. You can hire all of this kit but it is expensive. What you need is a friendly automotive engineer to give you some data, I'm sure they would have all you are looking for from testing ESP systems. Although interesting as a project I agree with most other people that drifting is about having fun with no electronic aids and usually when I get into a car the first thing I do is switch off ESP! Good luck with it

30-04-2008, 01:28	#47 (permalink)
chip-3door dorifto kingu! Join Date: Sep 2007 Posts: 900	Thanks for that V-Max, and im sure you are right about how most drifters would sooner not have electronic aids anymore than they would want hiv aids, lol

30-04-2008, 08:46	#48 (permalink)
PhatBob Mr Sir Join Date: Aug 2004 Location: Newmarket Posts: 1,560	VMax has a good point, it would be simple enough to use the ABS sensors to measure wheel speeds, then three Analog Devices accelerometers one at the front of the car, one midships and one at the rear. Steering input is a known with a rotary encoder or pot.