SONAR Sensor Software Driver

[Qbranch]

People have been PM'ing me about my single-pin ultrasonic driver for the Parallax PING))) ultrasonic sensor module.

I've cut out most of the signal processing, noise filtering, and object tracking code that I wrote, but there is enough here to get you your pulse-timing (raw distances). If you have questions post, please don't PM since many end up asking the same questions.

I realize it isn't the prettiest code, but then again I was never intending on posting it when it was written...

Enjoy... I attached the related files below. *cringes waiting for flood of 'this is bad programming convention' posts*

-q

[Protronie]

Originally Posted by Qbranch

Enjoy... I attached the related files below. *cringes waiting for flood of 'this is bad programming convention' posts*

-q

Your right... it ain't pretty but it works and thats what matters...
thanks for posting!

[Roger]

Hey, it has plenty of good comments -- is that what you mean by "bad programming convention"? Thanks Qbranch!

We did get Mike's sonar code working last night (oh, the simple errors we made) -- the places we can now go!

______________________________
It's kind of fun to do the impossible. -Walt Disney

[gnormhurst]

Back in 2006 I installed Mike's sonar driver and it worked great. I followed his instructions for modifying InterruptHandlerLow () and everything was good.

I want to do it again, but when I look at the current code, InterruptHandlerLow () looks very different!

Can someone have a quick look at this and tell me if I have created any obvious conflicts? My modifications are in color.

Thanks!

Norm

void InterruptHandlerLow ()     
{
	if (PIR1bits.RC1IF && PIE1bits.RC1IE) // rx1 interrupt?
	{
		#ifdef ENABLE_SERIAL_PORT_ONE_RX
		Rx_1_Int_Handler(); // call the rx1 interrupt handler (in serial_ports.c)
		#endif
	}                              
	else if (PIR3bits.RC2IF && PIE3bits.RC2IE) // rx2 interrupt?
	{
		#ifdef ENABLE_SERIAL_PORT_TWO_RX
		Rx_2_Int_Handler(); // call the rx2 interrupt handler (in serial_ports.c)
		#endif
	} 
	else if (PIR1bits.TX1IF && PIE1bits.TX1IE) // tx1 interrupt?
	{
		#ifdef ENABLE_SERIAL_PORT_ONE_TX
		Tx_1_Int_Handler(); // call the tx1 interrupt handler (in serial_ports.c)
		#endif
	}                              
	else if (PIR3bits.TX2IF && PIE3bits.TX2IE) // tx2 interrupt?
	{
		#ifdef ENABLE_SERIAL_PORT_TWO_TX
		Tx_2_Int_Handler(); // call the tx2 interrupt handler (in serial_ports.c)
		#endif
	}

        // the following added to support sonar:
        //
        else if (INTCON3bits.INT2IF && INTCON3bits.INT2IE)/* The INT2 pin is RB2/DIG I/O 1. */
        { 
          Hardware_Interrupt_Sonar(); 
          INTCON3bits.INT2IF = 0;
        }

        else if ( PIR2bits.TMR3IF && PIE2bits.TMR3IE )
        {
            Timer_Interrupt_Sonar();  
        }


}

[Roger]

Okay, now I'm getting a little annoyed. It's time to get this sonar working once and for all!

We got Qbranch's code working with our 2006 robot and code. We haven't tried it with newer code, but the major difference I believe was Kevin Watson's updated PWM routine. (Sorry, I don't have the code in front of me to quote chapter and verse.) It's been a month since I stuck my toe into the wonderful world of interrupts, but now I want to dive in and get it done. I'm going to throw some programmers at this and see how far we get this time.

[Jungle_Ed]

Hey I just got assigned to a 2008 FRC team. I am going to try using Mike 0.3 code then Qbranch's Code. Anyone, please let me know of you get some ok readings for the Sonar on and FRC robot, and I will be sure to do the same.

[gnormhurst]

We got the Vex sonar working last night using Mike's driver. I'm really impressed by the whole thing -- it can see a 3' x 4' board at about 15 feet!

Note: the documentation that came with the Sonar sensor I got (from Radio Shack, back when they sold Vex stuff) had the "input" and "output" functions backwards. See my earlier post in this thread.

The wire marked "input" is an input to the sonar, and it needs to come from a rc_digital_io_NN pin configured as an output. You need to set a couple of macros in sonar.h to tell it which pin you are using for the sonar input.

The output needs to go to a digital_io pin that acts as an interrupt. We used rc_digital_io_01. I think that is the default in Mike's code.

[Qbranch]

Originally Posted by gnormhurst

We got the Vex sonar working last night using Mike's driver. I'm really impressed by the whole thing -- it can see a 3' x 4' board at about 15 feet!

Beware of off-axis error! Depending on what sensor you're using (Parallax PING))) is pretty much your best bet) the further a wall is off axis the worse your measurements will be.

For example, if you were, say, following along a wall and servoing off of the distance from the wall (no idea why you'd want to do that ) you're robot will swerve further and further until it crashes, because, each time the robot veers left or right it alters the distance the ping sensor percieves, even though the distance has barely changed. This escalates until, like I said, you crash.

Instead (and we've used this in the past) use encoders to drive your robot straigh, and use ultrasonics for obstacle detection/avoidance.

Getting a reliable distance to anything at any angle is our current goal at 1024.

-q

[cprogrammer]

I have a quick question Mike Dubreuil. YOu ping the sonar with Ping_Sonar and turn the timer on. Then once it interrupts it goed into Hardware_Interrupt_Sonar an into the switch statement for SONAR_WAIT status and you turn the timer on again. Then it goes into the switch statement for SONAR_TIMING for the next interrupt and get phase width as a measurement. Is the phase_width the time it takes for the actuall sound wave to go through the sonar, not the echo from the wall? If it is the echo value, then why start the timer in SONAR_WAIT?

Thanks Stephen

[seg9585]

I have the code setup as gnornhurst posted in the "SONAR Sensor Software Driver" and i'm using rc_dig_out16 for input and digital IO port 1 for "output", but I'm getting a "Cannot find Sonar" error.
I was wondering if there were any additional code changes you needed to make to get this to work with this year's Kevin Watson code?

Thanks.

[GregK]

Last nite, I set to work adding a second sonar. Because of the limited time, I just made a second copy of Mike's sonar.c and sonar.h (which we are using for the first sonar). I named the second copy sonar2.c & h and changed the global variable names. While checking around to find what else I should change (which interupts and pins I should use), I found out that Mike has written another sonar named jvex_sonar.c routine that handles four sonars (at http://www.koders.com/c/fid735E99BB5...A82C4454.aspx).

What the story with that routine? Would I be better off starting with that code or continue with my approach?

[dadamson]

Originally Posted by seg9585

I was wondering if there were any additional code changes you needed to make to get this to work with this year's Kevin Watson code?

After a fair bit of wrestling yesterday, we got the 0.3 Sonar driver talking to Kevin W's 3.0 code. We ripped out references to his pwm.h, mostly (except for #includes) in ifi_frc.c Both it and the sonar driver want to use timer 3.

[B_Dubbs]

We, about 2 hours ago, have gotten all of our sonars to work with Kevin's encoder code. We did this by using both Mike's and Kevin's code together. Once we get everything perfectly straightened out, we could post it.

[mandrews281]

Originally Posted by dadamson

After a fair bit of wrestling yesterday, we got the 0.3 Sonar driver talking to Kevin W's 3.0 code. We ripped out references to his pwm.h, mostly (except for #includes) in ifi_frc.c Both it and the sonar driver want to use timer 3.

I'm trying the same thing with the Vex sensor and haven't had much luck. I ripped out ALL the pwm references and if I interpret the code correctly (a big if), it's using interrupt2, so I've connected the sensor to dig_in_02. Depending on which output/input I connect to dig_out_16/in_02, I get software malfunction or reversed I get an occasional "device not detected" (usually when I press or release the trigger). Anyway, in all cases the pulse_width (and therefore the distance) is 0.

Any chance of getting more details what you did so I can see what am I doing wrong? Thanks