Detecting wheel slip from DC motor on low friction surface

AI Thread Summary
The discussion focuses on methods to detect wheel slip from a DC motor on low-friction surfaces, specifically with a coefficient of friction around 0.05. Key suggestions include monitoring current draw for spikes or troughs, using an optical encoder to track wheel speed, and limiting acceleration to prevent slip. Participants debate the effectiveness of current draw as an indicator of traction loss, with some suggesting that a decrease in current could signal slipping. The conversation also touches on the use of additional sensors like accelerometers and gyros to enhance tracking of speed and orientation. Overall, the goal is to establish a reliable system for detecting and managing wheel slip in robotic applications.
Weird Fishes
Messages
23
Reaction score
0
I am trying to figure out what would be the best way (by best way I mean the cheapest, and simplest effective way) to detect slippage of the wheel. The surface is very low friction, somewhere in the range of 0.05 mu.

Here are my ideas:
  1. Sense current, watch for spikes/troughs
  2. monitor wheel speed with optical encoder and compare to velocity
  3. limit acceleration

I do have a question about the current draw solution: what would happen to the current draw if the wheel were to lose traction? Would that be an effective indicator at all. If it would work, that would be my preferred option.

What else could I use? Which of these would/wouldn't work?

Thanks for your help.
 
Engineering news on Phys.org
"what would happen to the current draw if the wheel were to lose traction?"
If you unload the motor, I would think the current would decrease.
 
dlgoff said:
"what would happen to the current draw if the wheel were to lose traction?"
If you unload the motor, I would think the current would decrease.

I suspected that as well, but I plan on doing some tests the next time I have a chance to work on this. Do you think the drop would be noticeable enough to use as a guideline?
 
I think so. You will probably get the same decrease when you slow down quickly however. So you will have to figure a way to determine the difference.
 
I think comparing it to the pwm signal should do the job. I'll post the results after the tests, which will hopefully get done this weekend.

Thanks for your help.
 
Did you plan to use the electric motor to drive the wheel any way, or is the electric motor something you are adding purely to detect wheel slip? Your situation is not clear in this respect, and it does make a difference.
 
First?
 
Dr.D said:
Did you plan to use the electric motor to drive the wheel any way, or is the electric motor something you are adding purely to detect wheel slip? Your situation is not clear in this respect, and it does make a difference.
It is an electric drive train. The electric motors power drive wheels.


Integral said:
First?
robotics?... yeppers.
 
Weird Fishes said:
Here are my ideas:
  1. Sense current, watch for spikes/troughs
  2. monitor wheel speed with optical encoder and compare to velocity
  3. limit acceleration

#1 Probably not practical unless your surface is very uniform. Otherwise you'll see variations due to variations in the surface.

#2 The voltage to the motor is already roughly proportional to the speed. It seems like the optical encoder would be redundant. How are you determining velocity in order to compare the optical encoder to it? It might make more sense to use the optical encoder on an undriven wheel to compare to the speed of the driven wheel or is that what you were trying to say?

#3 Not a bad idea. The voltage across the DC motor is roughly proportional to the speed and the current is roughly proportional to the torque (They're very close if you take the losses of the motor into consideration). If you limit the current to the motor and thus the torque, slipping should be less of a problem. (I once had a car that had a button on the shifter (automatic transmission) that would keep the transmission out of first gear in order to make starting on ice easier)
 
  • #10
skeptic2 said:
#1 Probably not practical unless your surface is very uniform. Otherwise you'll see variations due to variations in the surface.

#2 The voltage to the motor is already roughly proportional to the speed. It seems like the optical encoder would be redundant. How are you determining velocity in order to compare the optical encoder to it? It might make more sense to use the optical encoder on an undriven wheel to compare to the speed of the driven wheel or is that what you were trying to say?

#3 Not a bad idea. The voltage across the DC motor is roughly proportional to the speed and the current is roughly proportional to the torque (They're very close if you take the losses of the motor into consideration). If you limit the current to the motor and thus the torque, slipping should be less of a problem. (I once had a car that had a button on the shifter (automatic transmission) that would keep the transmission out of first gear in order to make starting on ice easier)
1: The surface is very uniform, otherwise I wouldn't have considered it either.
2: Encoder to determine wheel speed. Accelerometer to back calculate velocity (or possibly just use the acceleration). I have considered your suggestion as well, but don't have room for the undriven wheel.
3: Yes, that could be a better use of a current sensor.
 
  • #11
I am also working on this problem. See the big picture http://www.usfirst.org/community/frc/content.aspx?id=418" (watch the center clip on the right for a description of the game.

There are some rotary encoders in the kit of parts so we should be able to get good info on wheel speed. Also there is an accelerometer and a gyro which should be able to track the exact speed and orientation of the bot.

I am hoping to set a max ramp rate on the wheel speed. We just need to find the max acceleration possible without slip. And do some labview work.

I was dreaming of four wheel drive and steering but the limitation on the number and type of allowed motors is making this is impossible.

Good luck
 
Last edited by a moderator:
  • #12
Integral said:
I am also working on this problem. See the big picture http://www.usfirst.org/community/frc/content.aspx?id=418" (watch the center clip on the right for a description of the game.

There are some rotary encoders in the kit of parts so we should be able to get good info on wheel speed. Also there is an accelerometer and a gyro which should be able to track the exact speed and orientation of the bot.

I am hoping to set a max ramp rate on the wheel speed. We just need to find the max acceleration possible without slip. And do some labview work.

I was dreaming of four wheel drive and steering but the limitation on the number and type of allowed motors is making this is impossible.

Good luck

Yes, the motor limit is ruining many of my plans as well.

Good luck to you as well, what team/regional(s)?
 
Last edited by a moderator:
Back
Top