SCARA robot arm

Yes that is the one. 

@bldrgy 

Unfortunately I don't have a Huskylens so I don't have any experience using it.

Perhaps @oracid or someone else who has used the Huskylens will chime in?

It appears you can teach it to recognize an object (shape, template) by placing a rectangle around it.

If on a conveyor belt its position relative to the arm would be known given its position reported by Huskylens.

Will it return the orientation (rotation around its center) of the shape as well as position, anyone know?

Why the L shaped brackets?  Just an experiment?

@robotbuilder Thanks for input. I’m trying to make a mini product line that would stack the brackets in a certain number. 
Other issues I foresee, is the robot fast enough to either catch it with grippers or try using a suction cup/magnet and try to catch it in the middle of the bracket. 

@bldrgy 

Let me just say if you have actually built this robot arm then I am very impressed.

https://howtomechatronics.com/projects/scara-robot-how-to-build-your-own-arduino-based-robot/

Catching a bracket while the conveyor belt is moving might be tricky. Can you control the conveyor belt speed such as stopping it while the object is lifted off? The arm is not all that fast so I assume you have some control over the conveyor belt speed.

What is the purpose of stacking the brackets? Is this just an experiment or something practical?

How do they get on the conveyor belt in the first place?

An electromagnet may be a way to grab metal brackets off the belt as they pass. Then the rotational orientation of the bracket on the arm might be determined for stacking.

By the way it is possible to attach an image if you want to give some visual idea of your setup by clicking the Attach Files option at the bottom of the post with the paper clip image.

Posted by: @robotbuilder

@bldrgy 

Let me just say if you have actually built this robot arm then I am very impressed.

Thanks!

What is the purpose of stacking the brackets? Is this just an experiment or something practical?

It’s kinda both, and experiment that hopefully leads to something practical. The brackets need to be stacked so they can easily be handled to be placed in a package.

How do they get on the conveyor belt in the first place?

That’s a whole neither issue! I guess I’ll figure that out later if I can make the pick and place process work out. I’ve also pondered if the Husky lens could pick them from a random pile. 

An electromagnet may be a way to grab metal brackets off the belt as they pass. Then the rotational orientation of the bracket on the arm might be determined for stacking.

That was my though also. 

 

Here is a picture.

@bldrgy 

A camera could be pointed down ahead of the pick up arm so it would have a clear view. The arm could move into place and grab it at just the right time. It would also be useful if the computer can control the speed of the conveyor belt.

You never indicated the type of L bracket? The easiest to recognize would be the 2d flat L shaped bracket. The other two might be easier for the gripper to pick up (unless a solenoid magnet is used but then how do you orientate it).

To enlarge an image, right click the image and choose Open link in new window.

It may be possible to align the brackets by pouring them onto a sloping table with an edge on the bottom before dropping them onto the conveyor belt.

Of course there is no competition here between the slow robot arm and a human doing the same job 🙂

Posted by: @robotbuilder

@bldrgy 

A camera could be pointed down ahead of the pick up arm so it would have a clear view. The arm could move into place and grab it at just the right time. It would also be useful if the computer can control the speed of the conveyor belt.

Conveyor would be variable speed. I suppose it wouldn’t even need to stop. Missed pickups could be recycled back into the belt. Unless you are thinking to stop the belt each time the camera sees a brackets?

You never indicated the type of L bracket? The easiest to recognize would be the 2d flat L shaped bracket. The other two might be easier for the gripper to pick up (unless a solenoid magnet is used but then how do you orientate it).

Brackets are like the first image but 1.5x1.5x2.5 inches in width. So there could be enough surface to grab it. 

To enlarge an image, right click the image and choose Open link in new window.

It may be possible to align the brackets by pouring them onto a sloping table with an edge on the bottom before dropping them onto the conveyor belt.

I’ve tried two conveyors on a 90 and a downward slope from one to the other. That was fairly successful. 

Of course there is no competition here between the slow robot arm and a human doing the same job 🙂

Not yet:) I’ve also looked at Delta robots, they are much quicker but the range of motion appears to be a little limited. Thanks for you thoughts, much appreciated. 

 

I think my first step should be to get the lens wired to the Arduino and CNC shield. The Husky lens has 3 wires- power and communication. Would the communication wires go the pins 0 and 1 of the Arduino? 

If I know it’s wired properly and can establish communication between the lens and Arduino, then I can approach trying to code it. I’m going to assume that if communication is established and sent to the serial monitor, that’s how to code the robot to somehow go to that position?

@bldrgy I think you should first start with the Husky Lens tutorial, then think about incorporating that into the arm software. Aren't pins 0 and 1 reserved?

Hi

To day, since I waked up, I challenge this. Hope you appreciate.
I refer to this link, HuskyLens/HUSKYLENSArduino (github.com) and I write a less than 30 lignes code. I hope I have wrote it as simple as possible.

With id=0 you can recognize any things that are already in the HuskyLens memory. If you want to recognize your own thing, you have to learn it before.
The goal here is simply to print the parameters of an OBJECT_RECOGNITION on the Arduino IDE Monitor.

In your Internet Browser, let say that you look for cats. So you can scan hundreds cats pictures with the HuskyLens. It will recognize them very easily. The same for dogs, horses or cows, etc.
Do not hesitate to ask.

// Oracid-HuskyLens-OBJET_RECOGNITION - 18/08/2022 - ref.  https://github.com/HuskyLens/HUSKYLENSArduino 
#include <Wire.h>
#include <HUSKYLENS.h>
HUSKYLENS huskylens;
        
void  setup() {
  Serial.begin(9600);
  Wire.begin();
  while (!huskylens.begin(Wire)) {Serial.print("\n\t Check I2C"); delay(100);}
  huskylens.writeAlgorithm(ALGORITHM_OBJECT_RECOGNITION);
}

void loop() {
  int id=0;
  huskylens.request(id); 
  HUSKYLENSResult result = huskylens.read();
  if (result.command == COMMAND_RETURN_BLOCK){
    Serial.print("\n");
    Serial.print("\t xCenter=");  Serial.print(result.xCenter);
    Serial.print("\t yCenter=");  Serial.print(result.yCenter);
    Serial.print("\t width=");    Serial.print(result.width);
    Serial.print("\t height=");   Serial.print(result.height);
    Serial.print("\t ID=");       Serial.print(result.ID);
  }
  else{
    Serial.print("\n\t Object unknown!");
  }
  delay(1000);
}

@oracid That github library is invalid. 

@zander 

I don't remember exactly, but I guess I followed the instructions on this page, https://dronebotworkshop.com/huskylens/ which refers to this link, https://github.com/HuskyLens/HUSKYLENSArduino but I'm not sure.

Can you tell me which IDE do you use and how do you get this message.

@oracid I am using the normal IDE 1.8.20 When I installed the github download I get the error message. It's simple to fix, just move the code up one level.

Obviously you either did not use the Arduino IDE or you encountered the error and fixed it as well. I have reported it to the maintainers but don't know how long it will take them to respond. For now, just move the source code up one level. See attached pic of correct organization.