search

JorgeFernandes-Git / zau_bot

4 stars 1 forks source link

creation of a robotic system for simulation #1

Closed Kazadhum closed 1 year ago

rarrais commented 2 years ago

@sergiodmlteixeira will commit some files that enable the simulation of the Friday robotic platform in Gazebo.

@JorgeFernandes-Git and @Kazadhum based on these files you should be able to adapt it to the dimensions of the Zau robotic platform (see https://autode.sk/3F0TZOE) and combine it with the differential drive system and robotic manipulator (Universal Robots UR5) that we discussed in the last meeting.

JorgeFernandes-Git commented 2 years ago

We already have the UR5 robot operational with MoveIt configured. After receiving the Friday files we'll adapt them.

I noticed in the CAD files that the Zau has a specific end effector. Did Friday has the same? If don't, perhaps we can use the end effector of the Mmtbot.

rarrais commented 2 years ago

Yes, Friday also has an end effector, however I do not think that it is of value to model those at this stage, as we will be more concerned with the placement of sensors (and not so much actuators, such as grippers) in the platform, for calibration purposes.

Also, @JorgeFernandes-Git , if you already have an operational UR5 robot with MoveIt configured, please commit those files to this repository, as it is meant to server as a log of your progress.

JorgeFernandes-Git commented 2 years ago

Okay. I'll commit the files maybe later today.

JorgeFernandes-Git commented 2 years ago

The files needed to launch the first iteration of zau_bot were committed.

There are some modifications to be implemented relating to:

That's all I can recall for now.

miguelriemoliveira commented 2 years ago

Looks good

On Wed, Oct 26, 2022, 16:59 Jorge Fernandes @.***> wrote:

The files needed to launch the first iteration of zau_bot were committed.

There are some modifications to be implemented relating to:

  • Inertia of the base;
  • Mass of the base;
  • The caster wheels are not continuous joints, but fixed, just to give support to the base;
  • The base needs a mesh file to give a better appearance;
  • The robot has no sensors yet;
  • We did not discuss any external sensor for the calibration procedure;

That's all I can recall for now.

— Reply to this email directly, view it on GitHub https://github.com/JorgeFernandes-Git/zau_bot/issues/1#issuecomment-1292267578, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVSRMOGHEQY6VBXUTKLWFFINFANCNFSM6AAAAAARJFY4ME . You are receiving this because you are subscribed to this thread.Message ID: @.***>

JorgeFernandes-Git commented 2 years ago

Today I made a simple CAD model and added as a mesh to the base.

Here is the result:

Screenshot

miguelriemoliveira commented 2 years ago

That looks great. Congratulations!

miguelriemoliveira commented 2 years ago

Can you simulate it? When you can please post a video of the robot moving around using teleop.

Then a video of planning the arm motion using moveit.

Kazadhum commented 2 years ago

Done! Both videos are available in the Readme.

Meanwhile, I was wondering if inertia is relevant for the simulations. Right now we are using estimated values for the base's mass, but if inertia is important for our purposes, it might be in our best interest to use an accurate value

miguelriemoliveira commented 2 years ago

Hi @Kazadhum ,

The videos look great. Thanks.

About the videos: what are the blue and red spheres moving around close to the robot? Are those lidar measurements?

If so then the lidars are measuring the robot, is that it? You can add a cube object in gazebo to see if you get lidar measurements of the cube.

Kazadhum commented 2 years ago

Hello professor, Those are indeed LiDARs, and adding a cube in gazebo we get it's LiDAR measurements, as you can see in the picture. Should we not be getting the spheres near the robot, though? image

miguelriemoliveira commented 2 years ago

Well thats the lidar measuring parts of the robot.

I suggest you put the robot inside a closed environment, so that you always have measurements.

Then you can assess if the lidar is measuring close to 270 degrees each, as they should.

You can also use gazebo to check this. Watch this

https://youtu.be/m69_Z0YK2gc

Keep up the good work.

On Sat, Oct 29, 2022, 1:13 PM Diogo Vieira @.***> wrote:

Hello professor, Those are indeed LiDARs, and adding a cube in gazebo we get it's LiDAR measurements, as you can see in the picture. Should we not be getting the spheres near the robot, though? [image: image] https://user-images.githubusercontent.com/93128909/198830540-fcaaa96e-35d2-4ae9-8a8b-f3a04ce9d155.png

— Reply to this email directly, view it on GitHub https://github.com/JorgeFernandes-Git/zau_bot/issues/1#issuecomment-1295818925, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVUV6RP76623KL75T7LWFUIHPANCNFSM6AAAAAARJFY4ME . You are receiving this because you commented.Message ID: @.***>

JorgeFernandes-Git commented 2 years ago

Thanks for the videos @Kazadhum. They look great.

Regarding the lasers, there are 2 lasers, one in front and one in the back, publishing on different topics. The angle for the lasers is 180º at the moment, but I can change that. I used this value because it is what the lidars can see without hitting the robot.

I'll be adding the visibility to the lasers too and put the robot in a different world too, maybe tomorrow.

miguelriemoliveira commented 2 years ago

Hi @JorgeFernandes-Git ,

No images ...

JorgeFernandes-Git commented 2 years ago

I change the default launch world of the robot to TTB3 house and added visibility to the lasers.

https://youtu.be/fg23Tl_Sg0k

miguelriemoliveira commented 2 years ago

Hi @JorgeFernandes-Git ,

the video looks great. About the lidar placement, each lidar should measure close to 270 degrees, so that the combined coverage is complete.

The lidars shoud be placed more to the outside of the robot so that its possible to measure also what I annotated in white in the image bellow.

image

JorgeFernandes-Git commented 2 years ago

Maybe something like that @miguelriemoliveira? The axis of the lidars are a bit offset from the robot, but otherwise, they will colide with the robot base.

rviz_lasers

gazebo_lasers

miguelriemoliveira commented 2 years ago

Much better.

On Wed, Nov 2, 2022, 1:23 PM Jorge Fernandes @.***> wrote:

Maybe something like that @miguelriemoliveira https://github.com/miguelriemoliveira? The axis of the lidars are a bit offset from the robot, but otherwise, they will colide with the robot base.

[image: rviz_lasers] https://user-images.githubusercontent.com/80167550/199499562-b602583e-dfc6-4fee-9deb-bf34ae5a5dcb.png

[image: gazebo_lasers] https://user-images.githubusercontent.com/80167550/199499017-38b74481-d09a-465b-a141-a4af8ccae925.png

— Reply to this email directly, view it on GitHub https://github.com/JorgeFernandes-Git/zau_bot/issues/1#issuecomment-1300400564, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVQ337M2ZJITLK7Q5DLWGJTMPANCNFSM6AAAAAARJFY4ME . You are receiving this because you were mentioned.Message ID: @.***>