6 ms·

Building a semi-autonomous drone with Python

59 points by elisebreda 10y ago

actionwords 10y ago
anyone have more like this?
drauh 10y ago
Seems like an awesome addition to the Aerial Robotics Coursera (taught by Vijay Kumar at UPenn). https://www.coursera.org/learn/robotics-flight/home/welcome https://www.coursera.org/learn/robotics-flight/home/welcome
anysz 10y ago
This should be called "Building a semi-autonomous drone with ScienceOps" Maybe I'm overreacting, but I always get a bitter taste in my mouth when I start a tutorial whose title claims to teach something but then gaps a huge problem using a product. This is a form of growth hacking that screws over people who come to learn. So many companies do this, it's very frustrating. They could turn this into such a positive by introducing the technical way first... They'd then have it all laid out to describe their product and how much code it saves.
jjwiseman 10y ago
Two or three years ago it seemed that there was a lot of hacking happening on the AR.Drone: We had the Drone Games (née Drone Olympics), the ar-drone node module was under active development, and there was even an effort to write a replacement firmware. Now, http://dronegames.co/ http://dronegames.co/ is still touting the 2013 Games. Parrot has released many new drones since the AR.Drone, but from what I can tell it's not that the hackers are now working on the Bebop or Rolling Spider; I'm not sure where they went. I'm a committer on the ar-drone node library, but lately I've mostly been working on turboshrimp, a clojure library for AR.Drone control: https://github.com/wiseman/turboshrimp https://github.com/wiseman/turboshrimp (API docs at http://wiseman.github.io/turboshrimp/com.lemondronor.turboshrimp.html http://wiseman.github.io/turboshrimp/com.lemondronor.turbosh...). I've also got an Android app (written in Clojure)[1] and a desktop app that uses the CamShift tracking algorithm with OpenCV, which lets you select a target of interest and then will continue to track that target[3]. The AR.Drone is still a really nice, relatively hackable robotics platform. Running Linux and having USB lets you do things like connect an RTL-SDR dongle[4] and display an overlay of nearby aircraft on the drone video stream[5]. [1] https://github.com/wiseman/shrimpdroid https://github.com/wiseman/shrimpdroid [2] https://github.com/wiseman/turboshrimp-tracker/ https://github.com/wiseman/turboshrimp-tracker/ [3] https://www.youtube.com/watch?v=In6KFGakGGw https://www.youtube.com/watch?v=In6KFGakGGw [4] http://lemondronor.com/blog/indexphp/2013/4/cheap-ads-b-on-amateur-drones http://lemondronor.com/blog/indexphp/2013/4/cheap-ads-b-on-a... [5] http://lemondronor.com/blog/indexphp/2013/5/augmented-reality-display-of-air-traffic-for-drones http://lemondronor.com/blog/indexphp/2013/5/augmented-realit...
- robotresearcher 10y ago
  Well-featured ROS drivers are available for the Parrot drones: https://github.com/AutonomyLab/bebop_autonomy https://github.com/AutonomyLab/bebop_autonomy https://github.com/AutonomyLab/ardrone_autonomy https://github.com/AutonomyLab/ardrone_autonomy these are quite well-used in the research community.

59 points by elisebreda 10y ago

anyone have more like this?

drauh 10y ago

Seems like an awesome addition to the Aerial Robotics Coursera (taught by Vijay Kumar at UPenn). https://www.coursera.org/learn/robotics-flight/home/welcome https://www.coursera.org/learn/robotics-flight/home/welcome

anysz 10y ago

This should be called "Building a semi-autonomous drone with ScienceOps" Maybe I'm overreacting, but I always get a bitter taste in my mouth when I start a tutorial whose title claims to teach something but then gaps a huge problem using a product. This is a form of growth hacking that screws over people who come to learn. So many companies do this, it's very frustrating. They could turn this into such a positive by introducing the technical way first... They'd then have it all laid out to describe their product and how much code it saves.

jjwiseman 10y ago

Two or three years ago it seemed that there was a lot of hacking happening on the AR.Drone: We had the Drone Games (née Drone Olympics), the ar-drone node module was under active development, and there was even an effort to write a replacement firmware. Now, http://dronegames.co/ http://dronegames.co/ is still touting the 2013 Games. Parrot has released many new drones since the AR.Drone, but from what I can tell it's not that the hackers are now working on the Bebop or Rolling Spider; I'm not sure where they went. I'm a committer on the ar-drone node library, but lately I've mostly been working on turboshrimp, a clojure library for AR.Drone control: https://github.com/wiseman/turboshrimp https://github.com/wiseman/turboshrimp (API docs at http://wiseman.github.io/turboshrimp/com.lemondronor.turboshrimp.html http://wiseman.github.io/turboshrimp/com.lemondronor.turbosh...). I've also got an Android app (written in Clojure)[1] and a desktop app that uses the CamShift tracking algorithm with OpenCV, which lets you select a target of interest and then will continue to track that target[3]. The AR.Drone is still a really nice, relatively hackable robotics platform. Running Linux and having USB lets you do things like connect an RTL-SDR dongle[4] and display an overlay of nearby aircraft on the drone video stream[5]. [1] https://github.com/wiseman/shrimpdroid https://github.com/wiseman/shrimpdroid [2] https://github.com/wiseman/turboshrimp-tracker/ https://github.com/wiseman/turboshrimp-tracker/ [3] https://www.youtube.com/watch?v=In6KFGakGGw https://www.youtube.com/watch?v=In6KFGakGGw [4] http://lemondronor.com/blog/indexphp/2013/4/cheap-ads-b-on-amateur-drones http://lemondronor.com/blog/indexphp/2013/4/cheap-ads-b-on-a... [5] http://lemondronor.com/blog/indexphp/2013/5/augmented-reality-display-of-air-traffic-for-drones http://lemondronor.com/blog/indexphp/2013/5/augmented-realit...

robotresearcher 10y ago
Well-featured ROS drivers are available for the Parrot drones: https://github.com/AutonomyLab/bebop_autonomy https://github.com/AutonomyLab/bebop_autonomy https://github.com/AutonomyLab/ardrone_autonomy https://github.com/AutonomyLab/ardrone_autonomy these are quite well-used in the research community.