Go Down

Topic: smoothing pot input (Read 2882 times) previous topic - next topic

Coding Badly

My question is, what is a reasonable value to initialize with?


1. Average of N samples where N is about the same history you get from the EWMA (for ? = 0.9, N = 4 is adequate).

2. Start at zero and toss out values until the EWMA has a complete history (for ? = 0.9 the first six values are be discarded).

3. Or, as @Grumpy_Mike said, one sample. 

It depends on what you are trying to accomplish (and whether or not you're a statistician).

skyjumper


Quote
No, its not. Only static variables are initialized by the compiler. Non-statics will have a random value.
It's a global; it is initialised to zero.
Happily, the float representation of zero is the same as int or long.


You're right, I just checked the spec. My bad.

skyjumper


Quote
Currently I am storing the last 20 samples and averaging them, but I have been looking for a way to achieve this without consuming 80 bytes (20 floats)...


What are you trying to achieve?  Is the goal to filter out noise when reading a stable signal (in which case the "windowed average" you're doing now is probably the right choice)?  Do you need the average to "respond quickly" to signal changes?



I am getting speed readings several times each second from a transducer. Each of these readings tends to vary a bit from the prior one, and I just want to get a stable reading to use to present to the driver.


db2db


This works nicely. Thanks. 

Seems like it can be used for minor potentiometer output jitter. If I initialize smooth to the same value of the pot, it starts the calculation there, which is good.

I now see the relationship between the two side of that formula. One side is fast to react, the other is slow but better at smoothing.  Like the petrol needle in your car, very good at smoothing, but too slow to be used for updating a value based on a pot turning.


db2db

#19
Apr 30, 2012, 02:02 am Last Edit: Apr 30, 2012, 02:04 am by db2db Reason: 1
In case anyone wants to try this easily, here's my test code.
If you change the pot value - which has +5/-5 value jitter added - you'll quickly see the data output of 'smooth' stabilize.


Code: [Select]

float smooth;
int randomized_potval;

void setup()
{
 smooth = analogRead(5);  // grab sensor value to be initial value in calculation.
 Serial.begin(115200);
}


void loop()
{
 float potval = analogRead(5);   // read pot on pin 5
 randomized_potval = random( potval - 5, potval + 5);  // add some jitter

 smooth = (0.99 * smooth) + (0.01 * randomized_potval);  // smooth it out

 serial.print(randomized_potval);              // show jittered version of analog 5 value
 serial.print(" ");
 serial.println(int(smooth));            // OUTPUT -  we are looking for this smoothed number to be stable.

 delay (2);
}








Coding Badly

#20
Apr 30, 2012, 04:01 am Last Edit: Apr 30, 2012, 10:13 pm by Coding Badly Reason: 1
Currently I am storing the last 20 samples and averaging them, but I have been looking for a way to achieve this without consuming 80 bytes (20 floats)...  I am getting speed readings several times each second from a transducer. Each of these readings tends to vary a bit from the prior one, and I just want to get a stable reading to use to present to the driver.


α = 0.25 gives a window that almost decays away at 20 samples and can be implemented without floating-point.  I think this will work...

Code: [Select]

unsigned long history;
unsigned short value;

void setup( void )
{
  Serial.begin( 250000 );
  history = analogRead( 0 ) * 4;
}

void loop( void )
{
  history = analogRead( 0 ) + (((3 * history) + 2) / 4);
  value = (history + 2) / 4;

  Serial.println( value );

  delay( 100 );
}



Edit: added rounding.

skyjumper

Thanks!!!  I'll play with that...

Coding Badly

#22
Apr 30, 2012, 06:32 am Last Edit: Apr 30, 2012, 10:12 pm by Coding Badly Reason: 1
I think dot-8 rather than dot-2 fixed-point may give slightly more accurate results.  Try this one instead...

Code: [Select]
unsigned long history;
unsigned short value;

void setup( void )
{
  Serial.begin( 250000 );
  history = analogRead( 0 ) * 256;
}

void loop( void )
{
  history = (64*analogRead(0)) + (((64*3*history)+128) / 256);
  value = (history + 128) / 256;

  Serial.println( value );

  delay( 100 );
}



Edit: added rounding.

db2db


This looks very interesting, but I don't seem to follow the formula.

Can you explain what it's doing?

Coding Badly

#24
Apr 30, 2012, 08:53 am Last Edit: Apr 30, 2012, 10:15 pm by Coding Badly Reason: 1

The basic formula is...
v1 = (α * analogRead) + ((1 - α) * v0)

α = 0.25 or 1/4 ...
v1 = ((1/4) * analogRead) + ((1 - (1/4)) * v0)
v1 = ((1/4) * analogRead) + ((3/4) * v0)

To make it fixed-point with eight bits for the fraction multiple both sides by 256 (2 to the power of 8)...
256*v1 = 256 * { ((1/4) * analogRead) + ((3/4) * v0) }
256*v1 = ((256/4) * analogRead) + ((3*256/4) * v0)
256*v1 = ((64) * analogRead) + ((3*64) * v0)

The right-side has "v0" not "256*v0" so we have to perform the division when calculating the next value.  The multiplication is performed first to preserve the precision...
256*v1 = (64 * analogRead) + ((3*64*v0) / 256)

Finally, to improve the accuracy we need to include rounding...
256*v1 = (64 * analogRead) + (((3*64*v0)+(256/2)) / 256)

So, history is the "actual" value multiplied by 256.  Another way to look it: history / 256 is the whole number part and history % 256 is the fractional part.

Morris Dovey

There's always a better way!

db2db


Going back to the original, is the big disavantage that you have to use a float?

Is that much slower? How much?

Grumpy_Mike

Float is a lot slower, but I haven't got figures to say exactly how much. At least four times slower.

skyjumper



Going back to the original, is the big disavantage that you have to use a float?

Is that much slower? How much?


Floats are bigger as well. In my case, unsigned short int is more than enough space for 2 bytes.

Coding Badly

#29
Apr 30, 2012, 10:12 pm Last Edit: Apr 30, 2012, 10:16 pm by Coding Badly Reason: 1

Good point about storage.  history in Reply #20 can be unsigned short (half the size of float).  Assuming the "raw" values are between 0 and 1023, up to six fractional bits are possible with an unsigned short (instead of multiplying both sides by 256, multiply both sides by 64).  Which is a very nice compromise: about 1.5 decimals, smoothing, and fast all from just two bytes!  Warning: updating history overflows an unsigned short so the equation will have to be cast to an unsigned long before the right-side multiply.

history in Reply #22 has to remain an unsigned long (same size as float).  


Note: I updated #20 and #22 to make them a bit more accurate.

Go Up