Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Fault Tolerant UAV State Estimation
The goal of this thesis is to develop a fault tolerant UAV state estimation module using Visual, inertial and GPS sensors such that when one sensor fails, data from the other sensors may still be used for UAV navigation.
The goal of this thesis is to develop a fault tolerant UAV state estimation module using Visual, inertial and GPS sensors such that when one sensor fails, data from the other sensors may still be used for UAV navigation. This makes our system tolerant to erroneous measurements from GPS random walks and also compensates for long term odometry drift. The project implements a proposed enhancement of the currently used system for autonomously monitoring agricultural fields on a weekly basis within the Flourish Horizon 2020 project, during which the architecture should be tolerant to communication dropouts (Wi-Fi and RC), GPS dropouts and drift, large illumination changes and moving objects.
The goal of this thesis is to develop a fault tolerant UAV state estimation module using Visual, inertial and GPS sensors such that when one sensor fails, data from the other sensors may still be used for UAV navigation. This makes our system tolerant to erroneous measurements from GPS random walks and also compensates for long term odometry drift. The project implements a proposed enhancement of the currently used system for autonomously monitoring agricultural fields on a weekly basis within the Flourish Horizon 2020 project, during which the architecture should be tolerant to communication dropouts (Wi-Fi and RC), GPS dropouts and drift, large illumination changes and moving objects.
- Literature review and familiarization with ROS, ROVIO and the Modular Sensor Fusion framework.
- Implement a new update module in the MSF framework which fuses GPS as well as ROVIO pose estimates.
- Develop a fault detection strategy to ignore measurements when a sensor fails (GPS dropout or ROVIO divergence).
- Implement the fault detection strategy into a robust and reliable fault detection module within the framework.
- Literature review and familiarization with ROS, ROVIO and the Modular Sensor Fusion framework. - Implement a new update module in the MSF framework which fuses GPS as well as ROVIO pose estimates. - Develop a fault detection strategy to ignore measurements when a sensor fails (GPS dropout or ROVIO divergence). - Implement the fault detection strategy into a robust and reliable fault detection module within the framework.
- Reasonable C++ programming skills (you will get to improve them during the project).
- Familiarity with ROS is beneficial.
- Experience with flying small UAVs is a bonus.
- Reasonable C++ programming skills (you will get to improve them during the project). - Familiarity with ROS is beneficial. - Experience with flying small UAVs is a bonus.
Please send a CV/ web or github profile and transcripts to raghav.khanna@mavt.ethz.ch
Please send a CV/ web or github profile and transcripts to raghav.khanna@mavt.ethz.ch