[robotics-worldwide] [software] Open-Source Release: Build your own visual-inertial odometry aided cost-effective open-source autonomous drone
It is our great pleasure to announce the release of our work "Build your own visual-inertial odometry aided cost-effective open-source autonomous drone" that makes use of the off-the-shelf visual-inertial sensor, an onboard computer, and a quadrotor platform (DJI M100) to allow researchers to build high-performance autonomous drone given a fairly reasonable budget.
To make a long story short, you can find the detail from the link below:
Summary of key features
* Demonstration of several use cases utilizing our approach to show its repeatability.
* The total cost is 3169 USD to build your vi-drone.
* Providing intrinsic/extrinsic calibration sample dataset with tutorial documentation.
* Indoor (Vicon motion capture)/outdoor (Leica laser tracker) ground truth comparisons for control and state estimation performance.
* Tracking RMS error 0.036m while hover (more RMS can be found from the paper above).
* 0.82% drift error for 180m outdoor flight test in an agricultural scenario.
* Integrating our open-source packages (Rovio, MPC, MSF, Trajectory generator, and Kalibr)
We greatly acknowledge our support EU projects: Flourish and Aeroworks.
All feedback is very welcome!
ASL team: Inkyu Sa, Mina Kamel, Michael Burri, Michael Bloesch, Raghav Khanna, Marija Popovic, Juan Nieto, and Roland Siegwart