Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Robust Inertial-aided LiDAR Odometry for Search and Rescue in Degraded Environments
Odometry estimation, constitutes a major component for every autonomous mobile robot, as it is a fundamental requirement for several robotic applications. As LiDAR sensors are mainly unaffected by smoke as well as dust and provide a high resolution, they would be a potentially powerful approach.
Keywords: Laser Perception, Motion Tracking, Multi-State Constraint Kalman Filter, SLAM
Odometry, or state estimation, constitutes a major component for every autonomous mobile robot, as it is a fundamental requirement for several robotic applications such as navigation, planning and map building. As LiDAR sensors, are mainly unaffected by smoke as well as dust and provide a high resolution, they would be a potentially powerful approach for the odometry estimation. The nature of a rotating-LiDAR is closely related to rolling shutter cameras for which one of the state-of-the-art approaches is the so called Multi-State Constraint Kalman Filter (MSCKF).
In this project, we seek to get a greater understanding of a laser-based odometry solution for the use in robotics. Specifically, we would like to investigate/research the use of a MSCKF approach for the task of point cloud and feature-based SLAM. Depending on the student’s interest and skills, different representations of the laser scans can be incorporated in the pipeline as well.
Further information: https://docs.google.com/document/d/1QYOF-mzMrUzkEKS0r9M7Lha39oFxzZi9xftKauR5uAM/edit?usp=sharing
Odometry, or state estimation, constitutes a major component for every autonomous mobile robot, as it is a fundamental requirement for several robotic applications such as navigation, planning and map building. As LiDAR sensors, are mainly unaffected by smoke as well as dust and provide a high resolution, they would be a potentially powerful approach for the odometry estimation. The nature of a rotating-LiDAR is closely related to rolling shutter cameras for which one of the state-of-the-art approaches is the so called Multi-State Constraint Kalman Filter (MSCKF). In this project, we seek to get a greater understanding of a laser-based odometry solution for the use in robotics. Specifically, we would like to investigate/research the use of a MSCKF approach for the task of point cloud and feature-based SLAM. Depending on the student’s interest and skills, different representations of the laser scans can be incorporated in the pipeline as well.
Further information: https://docs.google.com/document/d/1QYOF-mzMrUzkEKS0r9M7Lha39oFxzZi9xftKauR5uAM/edit?usp=sharing
- Literature review on MSCKF, rolling shutter and laser-based odometry
- Implementation of laser-based odometry estimation
- Evaluation of the proposed framework
- Literature review on MSCKF, rolling shutter and laser-based odometry - Implementation of laser-based odometry estimation - Evaluation of the proposed framework
- Highly motivated and independent student
- Knowledge in C++
- Strong interest/background in state estimation
- Experience with ROS is beneficial
- Highly motivated and independent student - Knowledge in C++ - Strong interest/background in state estimation - Experience with ROS is beneficial
If you are interested, please send your transcripts and CV to Lukas Bernreiter (lukas.bernreiter@mavt.ethz.ch) and Florian Tschopp (florian.tschopp@mavt.ethz.ch).
If you are interested, please send your transcripts and CV to Lukas Bernreiter (lukas.bernreiter@mavt.ethz.ch) and Florian Tschopp (florian.tschopp@mavt.ethz.ch).