Distance conversion for ir

vert02 · Posted: Tue Mar 24, 2009 2:21 am

Helloo..

i m currently doing my colledge project and i m required to convert the value of sharp's ir sensor into distance value..

i am using PIC16F877A and an ir sensor to determine the movement of motor

i know how to get input from the ir sensor but it is in raw adc value..

i want the adc value to be converted into distance so that i can used it for my motor

required help,thank u

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

I guess, you're using some GP2Dxxx sensor? Unfortunately you didn't tell.

Did you read the sensor data sheet? The sensor has a nonlinear and also type dependant output voltage. The datasheet suggests a method to approximate the sensor function by a linear function of 1/x, the additional parameters possibly need individual calibration. Alternatively, you can use an empirical look-up table to convert analog voltage to distance.

vert02 · Posted: Tue Mar 24, 2009 11:41 pm

To fvm

Thanks for your reply. Yes I'm using that ir sensor.

Where can I find this empirical table?
Is it in the datasheet?
I read before but did not get the method of converting.

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

Empirical in my understanding means, you have to find the values by calibration measrurements with your individual device. This would be the preferred method to achive maximum accuracy. A more suitable way is probably to use the 1/x function suggested by the datasheet. As a general form

vert02 · Posted: Thu Mar 26, 2009 1:30 am

to fvm...

thanks again for your help..

i try use that method and report back to you..

vert02 · Posted: Thu Mar 26, 2009 2:45 am

hi fvm..

i have tried your solution but i m stuck again..

is the a,b and c is a fixed value which i can obtained from any point of the graph?

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

They are arbitray constants (a and c may be possibly negative), that have to be choosen to make the transfer function fit your intended characteristic.

I made some detailed suggestions for calculation methods without further considering your level of mathematical knowledge, cause I'm generally used to similar techniques. I basically tried to elobaorate what's missing in the Sharp datasheet.

My assumption is, that a three point calibration, spread in a meaningful way over the distance range appearing linear in the datasheet, would give one unequivocal set of parameters a,b,c.