Fixed indexes for Target Info VI

[Pirate programe]

We have the Rectangular Target Processing VI imported into the code, but here's the problem: The Target info array is built in the order that they're seen. This means that, if something obstructs the camera on the field, the Target Info VI could potentially be reordered, making it hard to pick a certain target to aim at reliably. So, we'd like to be able to make sure that a target is assigned to a specific index depending on its information, instead.

I have some idea of how to accomplish this: Since the targets are cross-shaped, you can expect them to have close X and Y values; the top and bottom targets will have close X values, and the left/right ones will have similar Ys. From there you could compare certain values to see if they're one of the groups you want, and place them in an array accordingly.

I think the logic is sound (feel free to correct me if I'm wrong), but I'm having trouble with the nuts-and-bolts implementation of it, in LabVIEW code.

[CRLS97]

Originally Posted by Pirate programe

We have the Rectangular Target Processing VI imported into the code, but here's the problem: The Target info array is built in the order that they're seen. This means that, if something obstructs the camera on the field, the Target Info VI could potentially be reordered, making it hard to pick a certain target to aim at reliably. So, we'd like to be able to make sure that a target is assigned to a specific index depending on its information, instead.

I have some idea of how to accomplish this: Since the targets are cross-shaped, you can expect them to have close X and Y values; the top and bottom targets will have close X values, and the left/right ones will have similar Ys. From there you could compare certain values to see if they're one of the groups you want, and place them in an array accordingly.

I think the logic is sound (feel free to correct me if I'm wrong), but I'm having trouble with the nuts-and-bolts implementation of it, in LabVIEW code.

From what I recall, and someone correct me if I'm wrong, the targets are in order by the calculated distance. It doesn't matter which target the camera sees first, but whichever target is closest according to the calculations done in the Vision VI.

Also you could go about doing what you just mentioned but also note that since it's in a cross shape and the x returns a value from -1 to 1 and y from -1 to 1, you could also find the maximum and minimum values of x for the horizontal ones and max/min of y for the vertical and use the index # returned and use the information you need from there..

[Alan Anderson]

Our target selection code loops through the target info array looking for the maximum or minimum x or y, depending on which hoop the gunner wants it to track. The result is a single target cluster of x, y, and distance. The x value gets used to rotate the turret; y and distance control shot angle and strength.

[Pirate programe]

Our target selection code loops through the target info array looking for the maximum or minimum x or y, depending on which hoop the gunner wants it to track. The result is a single target cluster of x, y, and distance. The x value gets used to rotate the turret; y and distance control shot angle and strength.

Yeah, come to think of it, looping through the array to select a target would be what I'd have done. I'm not sure how to loop through an array in LabVIEW, though...

[Alan Anderson]

Originally Posted by Pirate programe

I'm not sure how to loop through an array in LabVIEW, though...

Use a For loop and wire an array into it. I think the tunnel will be set to auto-index by default. What gets passed into the loop is the individual elements of the array, one at a time for each iteration of the loop.

[Pirate programe]

Originally Posted by Alan Anderson

Use a For loop and wire an array into it. I think the tunnel will be set to auto-index by default. What gets passed into the loop is the individual elements of the array, one at a time for each iteration of the loop.

Ah, I see. So then we compare the max/min x or y values to the Position cluster in the Target Info array?

I tried using the Search 1D Array function for this, but it seems that it requires a full cluster of inputs, instead of just one value...

[Greg McKaskle]

Yes, Search 1D doesn't have a wildcard ability, so what you do instead is to loop through and compare just the elements you compare about. It is really pretty easy once you do it.

Greg McKaskle

[Pirate programe]

Actually, do I need the For loop at all? I already have the Max and Min values, so I think I can just send those to the turret depending on the button being pressed.

...no, i can't because then I'd be moving a turret with a Y value *facepalm*

[Greg McKaskle]

Well, you are correct, with enough cascading code you don't need a loop. I can't see the rest of the diagram, but I assume you have a four element index, and you have written code that sorts four elements. When you have time, you may want to play with looping to generalize it since adding a few more elements quickly adds code.

Greg McKaskle

[Pirate programe]

The rest of the diagram is merely gathering all the values I need for the core logic (the Index Array function opening up the Target Info arra, the Joystick Get retrieving joystick buttons).

Yeah, four-element array is correct, since at the most we'll only really be seeing four targets at one time, right?

EDIT: The For loop doesn't seem to be iterating over the array.

I suppose that might be because of the Boolean checking- it's supposed to iterate over the array until it finds a value for which the checking is true.

[Greg McKaskle]

If you breakpoint and turn on the execution hilighting when you get to the loop, I think I'll see that it really is iterating and that it is likely a logic issue.

For what you are doing, the flat sort is fine. I was just encouraging you to work on the general one sometime. It is also trivial to do this in a new VI that doesn't even run on the robot. Then you type numbers into the array and run your subVI until you KNOW that it works. Then you replace the flat code with the subVI. This is called refactoring and is a great technique to practice.

Greg McKaskle

[Pirate programe]

Here's the basic program.

I probed the cluster value that's coming out of the auto-indexed tunnel, and it doens't seem to be changing from the 0th index, namely the first target.

EDIT: actually, no, I used a breakpoint, and it does seem to be changing...

[propionate]

Sorry to resurrect a (possibly?) dead thread, but I couldn't figure out the code that was posted here. I am trying to isolate the X value of the target with the highest Y value (aka the 3rd level hoop). I've attached the code that I made, could someone let me know if this will work?

EDIT: The input to the case structure determines which X value is output'ed