4 ms·

PythonRobotics: Python sample codes for robotics algorithms

208 points by wei_jok 8y ago

dekhn 8y ago
The visualizations are lovely.
rootusrootus 8y ago
Oh, neat! I'm currently taking the AI for Robotics class through Georgia Tech OMSCS, so this is a topic I am interested in, and in the language I currently use professionally. Thank you for sharing!
- aashu_dwivedi 8y ago
  Oh great! Is Sebastian Thrun still teaching that class? I took it a few years ago in OMSCS and learnt a lot! One of my favorite classes in the program.
  - rootusrootus 8y ago
    He does all the videos, but the instructor running this instance of the class is Jay Summet.
- deskamess 8y ago
  Wow... this is pretty much the core component of that class. I wish I had access to this after the assignments. Not having access to working code after we were graded so that we knew what was wrong was just horrible/painful/depressing. To make matters worse, the next assignment did build on that. Helplessness. I did comment in my end of class review about it, but the forum feedback was pretty much 'it had to be that way' so the next cohorts did not have access to it. I was not the only one pleading for working code. I was happy to get a poor grade as a reflection on work submitted, but I wanted to learn how to do it right and that opportunity was not afforded to me. I still did not get access at the end of the class so I hope this site helps me fill in the knowledge gaps that the course refused to do.
nbeleski 8y ago
I wonder how easy it would be to integrate this to ROS for real robotic simulations. Anyone has experience in the matter?
- jefft255 8y ago
  It highly depends on what exactly you want to integrate on which robot. Stuff like control algorithms can be dependent on the type of robot used. Also, a lot of this stuff is already present in ROS as a downloadable package.
  - nbeleski 8y ago
    I confess I only worked on small projects and all the movement/trajectory/obstacle avoidance functions were implemented manually, so this didn't cross my mind. It really makes sense to rely on the ROS packages, I will look what is available, thank you.
- DoritoChef 8y ago
  I have not used real robotic simulations, but I have used RViz and ROS for testing my own path-finding algorithms. Performance on my Late 2015 MacBook Pro was stuttery (about 15-20 fps), but still good enough to be somewhat usable. Overall, I had more issues with finding good resources on the rospy module than anything else. It seems almost all robotics development with ROS is done using C++ instead of Python (in my little experience), so there is little I could find in terms of code samples that would allow to get past a project like this (https://github.com/gandalf15/CS3027--CS5059--Robotics-University-of-Aberdeen https://github.com/gandalf15/CS3027--CS5059--Robotics-Univer...).
  - nbeleski 8y ago
    I felt the same way during my lectures. I wanted to use Python because I had already implemented the fuzzy functions I needed for obstacle avoidance and it is overall simpler, but most people on class used C++ anyway.
    - 8y ago
DoritoChef 8y ago
This is some really cool information! I implemented a map-parsing algorithm that used a data structure that enabled neighboring cell access in constant time. It's biggest weakness is that very high-resolution occupancy maps take quite some time to render. I'm very inexperienced when it comes to sharing my coding projects with the world, so any feedback on the readability/accessibility of my repository would be greatly appreciated! https://github.com/dwrodri/LQTLD3 https://github.com/dwrodri/LQTLD3

208 points by wei_jok 8y ago

The visualizations are lovely.

Oh, neat! I'm currently taking the AI for Robotics class through Georgia Tech OMSCS, so this is a topic I am interested in, and in the language I currently use professionally. Thank you for sharing!

aashu_dwivedi 8y ago
Oh great! Is Sebastian Thrun still teaching that class? I took it a few years ago in OMSCS and learnt a lot! One of my favorite classes in the program.
- rootusrootus 8y ago
  He does all the videos, but the instructor running this instance of the class is Jay Summet.
deskamess 8y ago
Wow... this is pretty much the core component of that class. I wish I had access to this after the assignments. Not having access to working code after we were graded so that we knew what was wrong was just horrible/painful/depressing. To make matters worse, the next assignment did build on that. Helplessness. I did comment in my end of class review about it, but the forum feedback was pretty much 'it had to be that way' so the next cohorts did not have access to it. I was not the only one pleading for working code. I was happy to get a poor grade as a reflection on work submitted, but I wanted to learn how to do it right and that opportunity was not afforded to me. I still did not get access at the end of the class so I hope this site helps me fill in the knowledge gaps that the course refused to do.

nbeleski 8y ago

I wonder how easy it would be to integrate this to ROS for real robotic simulations. Anyone has experience in the matter?

jefft255 8y ago
It highly depends on what exactly you want to integrate on which robot. Stuff like control algorithms can be dependent on the type of robot used. Also, a lot of this stuff is already present in ROS as a downloadable package.
- nbeleski 8y ago
  I confess I only worked on small projects and all the movement/trajectory/obstacle avoidance functions were implemented manually, so this didn't cross my mind. It really makes sense to rely on the ROS packages, I will look what is available, thank you.
DoritoChef 8y ago
I have not used real robotic simulations, but I have used RViz and ROS for testing my own path-finding algorithms. Performance on my Late 2015 MacBook Pro was stuttery (about 15-20 fps), but still good enough to be somewhat usable. Overall, I had more issues with finding good resources on the rospy module than anything else. It seems almost all robotics development with ROS is done using C++ instead of Python (in my little experience), so there is little I could find in terms of code samples that would allow to get past a project like this (https://github.com/gandalf15/CS3027--CS5059--Robotics-University-of-Aberdeen https://github.com/gandalf15/CS3027--CS5059--Robotics-Univer...).
- nbeleski 8y ago
  I felt the same way during my lectures. I wanted to use Python because I had already implemented the fuzzy functions I needed for obstacle avoidance and it is overall simpler, but most people on class used C++ anyway.
  - 8y ago

DoritoChef 8y ago

This is some really cool information! I implemented a map-parsing algorithm that used a data structure that enabled neighboring cell access in constant time. It's biggest weakness is that very high-resolution occupancy maps take quite some time to render. I'm very inexperienced when it comes to sharing my coding projects with the world, so any feedback on the readability/accessibility of my repository would be greatly appreciated! https://github.com/dwrodri/LQTLD3 https://github.com/dwrodri/LQTLD3