PDA

View Full Version : Inertial Navigation Sensors



Robocrazy
09-13-2007, 08:37 PM
Hey!

I'm working on a machine which runs on two Maxon ECmax Brushless DC motors which give an output rpm of about 1700. Well i need to traverse a distance of about 3m in a perfect straight line at an average speed of 1meter/sec. The weight of the machine is about 10kg. Now, i have no external feedback whatsoever.. i.e no lines, no walls, nothing. So i need to utilize inertial and/or magnetic sensor modules. So what is the best possible sensor combination (considering the high speed of the machine) that can be used for making the machine go straight with negligible or no error. The cost of the sensors is not a constraint.

rao
09-13-2007, 10:52 PM
make a thin line follower.. draw a straight line

Physically speaking, it is almost impossible to get a robot driven by two seperate motors to go straight and in your case as the motor speed is high.. there will be considerable difference in both motors..

Search for dual differential drive mechanism... Theoretically speaking, it is one of the few or maybe the only (correct me if I am wrong please) physically implemented drive systems that offers perfect straight motion.. but then tire friction and gear slip come into play..

thats all I can say on this topic..

Rao

orion
09-14-2007, 12:02 AM
Hello there,
Yeah, what you have is a 'difficult' problem statement, not entirely impossible.
You could use a combination of inertial sensors, encoders to achieve good odometry.
Here are some excellent articles to get you started:
http://www.seattlerobotics.org/encoder/200010/dead_reckoning_article.html
http://www.geology.smu.edu/~dpa-www/robo/jbot/index.html
http://www.geology.smu.edu/~dpa-www/robo/lego/index.html

Hope this helps.
Regards,

orion
09-14-2007, 12:20 AM
or maybe do what the robocon guys do. make the robot drop a long straight aluminum bar in front of it and then simply follow it. quick n dirty!

Robocrazy
09-14-2007, 04:59 PM
Thanks a lot for the suggestions!

I came across Honeywell's Trupoint Compass Module http://www.ssec.honeywell.com/magnetic/datasheets/HMR3500.pdf

Do you think this can help? It says it has accuracy of 0.1degrees. But i've heard that in case of compass the null starts to drift just like the rate gyros..

On another note, how about using Angular Rate Gyro with an accelerometer in a Kalman Filter feedback loop? Are there any integrated modules available with a complete rate gyro accelerometer system which give an integrated output ?

docel
09-19-2007, 08:04 AM
Thats a DREAM :!:
1. Forget the 2 motor drive. No matter WHAT you do with the Electronics, you'll fail due to the surface coeefficients.
2. A single wheel drive with some aerodynamics will be appropriate, and more reliable.
But, 10Kgs?? Thats quite a bit of acceleration. The correction points for the speed and direction will be quite large. This may not be reciprocated by the Motors in time.

Good Luck!!