Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 08:07
    jlblancoc edited #954
  • 08:07
    jlblancoc opened #954
  • 08:06

    jlblancoc on lint-clean-ups

    tons of clang-tidy fixes (progr… (compare)

  • 06:21
    jlblancoc edited #854
  • 06:02

    jlblancoc on master

    fix FTBFS in big endian systems (compare)

  • 05:45
    jlblancoc commented #953
  • 05:45
    jlblancoc closed #953
  • 05:43
    jlblancoc opened #953
  • Aug 22 18:51
    jolting commented on 3316834
  • Aug 22 15:30
    jlblancoc commented on 3316834
  • Aug 22 08:13

    jlblancoc on hokuyo-better

    (compare)

  • Aug 22 08:13

    jlblancoc on hokuyo-better

    (compare)

  • Aug 22 08:12
    codecov[bot] commented #941
  • Aug 22 08:12
    jlblancoc synchronize #941
  • Aug 22 08:12

    jlblancoc on add-rotating-scan-obs

    also add a kitti raw test file add VLP16 velodyne example data first attempt at a generic rota… (compare)

  • Aug 22 08:11

    jlblancoc on spin-off-from-add-rotating-scan-obs

    (compare)

  • Aug 22 08:11

    jlblancoc on ros-more-pointcloud-types

    (compare)

  • Aug 22 08:11

    jlblancoc on master

    Add more ROS point cloud conver… remove try/catch Merge branch 'ros-more-pointclo… (compare)

  • Aug 22 08:02
    jlblancoc closed #951
  • Aug 22 08:02

    jlblancoc on master

    BUGFIX: Missing emplace() in op… (compare)

Sanyam Vikram
@sanyam803
My name is Sanyam and I am a final year student of department of Electronics and communication of NIT JALANDHAR (NITJ). I am currently pursuing a bachelor's degree in ECE.
I keep a good knowledge in computer vision ,c++ and python. I would like to ask any available metors,if possible ,for guidance .
I am willing to participate in GSOC 2019
Hunter Laux
@jolting
We're glad you're already thinking about gsoc 2019. Just as students have to be accepted, our participation is not certain. Many organizations favor prior participation in open source communities, so please feel free to contribute since it helps you stand out as a student. If you have a project in mind with feel free to ask questions.
flyingraijin9
@flyingraijin9
@jolting Sir, can you list our some links for current projects of MRPT so that we can start learning and contributing in the project as it seems a bit confusing about what projects are ongoing currently and what issues are active
Siddharth Saha
@bear-my-wrath
Hello, I'm new to open source but I'd really like to learn and help to contribute. I'd be grateful if someone guided me where-from I can get started
Jeetu Kumar
@i-am-g2
Hello, I want to contribute on this project, but I am new to Open Source. Where should I start?
Hunter Laux
@jolting
Cool. Looks like we have some interests for contributing. Sorry for the late reply, but gitter is usually not this active. All issues on open issues on github can be worked on and suggestions are welcome.
joker2904
@joker2904
hello guys
is MRPT a mentoring organization for GoSc this time ? 2019 ?
Hunter Laux
@jolting
Not sure yet. Stay tuned for announcements on our mailing list.
Divya Tiwari
@BackDor3_twitter
Hi can i install the library on raspbian (i've tried the ppa method it's not working). Anyone who has successfully compiled the test.cpp code on raspbian os?
Nikos Koukis
@bergercookie
You can always compile the project from source: https://github.com/MRPT/mrpt#32-build-from-sources :)
And if that doesn't work try opening a new issue https://github.com/MRPT/mrpt/issues
Hunter Laux
@jolting
I don't know if raspbian will work with the ppa. I think you will need an ubuntu based distro.
Ubuntu MATE should run on raspi 2 and 3. On a side note, the jetson nano was recently release. Might be a better option for robotics considering the performance and GPU(although mrpt doesn't have cuda code).
Hunter Laux
@jolting
@jlblancoc can you turn arm64 on in the PPA?
Jose Luis Blanco-Claraco
@jlblancoc
Hi!
It is already enabled ;-)
I have used ubuntu mate in rbpi3 with this ppa successfully, it's the fastest option
DJuego
@DJuego
Hi to all!
I am a absolute newbie in this field. ;-)
I am working in a robotic-based project of Indoor Positioning System.
For this purpose, we are building a robot similar to this: https://www.youtube.com/watch?v=uv2faO7GhXc
First Question: This 4 joints/4 wheels design (or architecture)...
What is its 'formal' denomination?
DJuego
@DJuego
I would like to look for papers or similar that describe or model its kinematics mathematically...
Anyway, we have a ultrasonic indoor "GPS" than promise a +/2cm accuracy.
The challenge is to develop a autonomous drive in order to reach certain coordinates
in the XY plane. Or follow a path as tightly as possible...
What approach can I follow? What family of algorithms solve this kind of problem for mobile robots?
Can MRPT help me?
matrixbot
@matrixbot

bergercookie hey DJuego (Gitter) ,

I'm not sure of the name of this kinematic model. They look very similar to the omni wheels though.
Notice that the wheels themselves are not that special. it's rather the motors above them that rotate them

DJuego
@DJuego
Thanks!
matrixbot
@matrixbot

bergercookie > You could take a look at the mrpt-navigation component

I thinkyou are looking for is the holonomic model

DJuego
@DJuego
@matrixbot Yes. It is holonomic for sure. :-)
matrixbot
@matrixbot

bergercookie https://en.wikipedia.org/wiki/Holonomic_(robotics)

Anyway, there are navigation algorithms in MRPT that can help you navigate from A -> B, given some sensor input. I guess the most popular one would be the reactive navigator: https://www.mrpt.org/tag/reactive-navigation/

bergercookie If you 're using ROS, MRPT also offers ROS wrappers: https://github.com/mrpt-ros-pkg/mrpt_navigation
bergercookie As far as I know though, either the reactive nav2d or the localisation nodes operate using 2D laser scans
DJuego
@DJuego
"reactive navigator". Ok.
Yes. I want to track in real-time. I get the position through a system of ultrasonic beacons.
The promised precision has an estimated +/-2 cm error. The idea is to stay as close as possible to the planned path.
Especially at the destination point.
matrixbot
@matrixbot
bergercookie Well, what's the representation of your planned path though?
DJuego
@DJuego
At the moment I try to avoid ROS. I guess I could change my mind later
matrixbot
@matrixbot
bergercookie is it a black line on a 2D surface?
bergercookie Is it a set of given (x,y) coords?
DJuego
@DJuego
A set of given (x,y) coords sounds good to me.
I want flexibility and versatility.
Actually, the path could have "curves", like the one that can grant a mathematical function (exponential, logarithm, ...).
matrixbot
@matrixbot

bergercookie OK, but how do you reach that (x,y) representation to feed to your algoirthm?

If you have that (x,y) representation and your estimated position from your GPS-like system then you can implement simple PID controllers for the x,y (and theta?) of your robot to follow those waypoints

bergercookie Otherwise you have to "sense" where the line is relative to the robot and have some sort of feedback control loop to follow that line.
Look up path planning algorithms or line following more specifically
DJuego
@DJuego
I understand you. Yes. I hope to have a (X,Y) representation (my desired position) and my "real" position from the sensorial GPS-Like system. I want to minimize the error making the necessary corrections through orders to the 8 actuators (wheels and joints).
DJuego
@DJuego
Well. (X,Y, theta) :-)