Register now After registration you will be able to apply for this opportunity online.
Supporter State Estimation System
Implementation of a localization and state estimation system for a special ground robot using IMU and 2D LiDAR. The ground robot is part of a Gravity offload system for testing of spacecraft solar arrays (which are designed for microgravity) on earth.
Keywords: State Estimation
Localization
**Background Information**
Today, on-ground deployment testing of Spacecraft Solar Arrays happens via passive Gravity Offloading Jigs
(GOJ). These jigs are rail- or balloon-based systems and are heavy, hard-to-transport, and inflexible. This
project is part of a European Space Agency (ESA) funded research project, which will develop, manufacture,
and test an actively controlled robotic Gravity Offloading Jig (GOJ) based on a multi-robot system. This system
aims to be smaller, more accurate, more flexible, and cheaper than traditional passive GOJs. The project is led
by Beyond Gravity Switzerland (BGC, former RUAG Space) and the Autonomous Systems Lab provides
technical support.
**Project Description**
A previous master thesis developed a reactive navigation solution that steers robots using only on-board state
estimation in a manner to support the appendages and cancelling the effects of gravity. During this semester
project, the core state estimation components will be selected and evaluated using data recorded from the
actual robotic systems used in the project. The required components include a pose estimation system fusing
odometry and IMU data, a localization system using 2D LiDAR sensors to map and estimate the robot’s
position, and an object detection system using the 2D LiDARs to detect other robots and obstacles.
**Background Information** Today, on-ground deployment testing of Spacecraft Solar Arrays happens via passive Gravity Offloading Jigs (GOJ). These jigs are rail- or balloon-based systems and are heavy, hard-to-transport, and inflexible. This project is part of a European Space Agency (ESA) funded research project, which will develop, manufacture, and test an actively controlled robotic Gravity Offloading Jig (GOJ) based on a multi-robot system. This system aims to be smaller, more accurate, more flexible, and cheaper than traditional passive GOJs. The project is led by Beyond Gravity Switzerland (BGC, former RUAG Space) and the Autonomous Systems Lab provides technical support.
**Project Description** A previous master thesis developed a reactive navigation solution that steers robots using only on-board state estimation in a manner to support the appendages and cancelling the effects of gravity. During this semester project, the core state estimation components will be selected and evaluated using data recorded from the actual robotic systems used in the project. The required components include a pose estimation system fusing odometry and IMU data, a localization system using 2D LiDAR sensors to map and estimate the robot’s position, and an object detection system using the 2D LiDARs to detect other robots and obstacles.
- Design a perception and state estimation architecture for a ground robot
- Select and implement the architecture
- Evaluate the architecture’s performance on recorded data
- Design a perception and state estimation architecture for a ground robot - Select and implement the architecture - Evaluate the architecture’s performance on recorded data
- Motivated and independent student wanting to work on a real-world robotic project
- Solid programming skills in C++ are mandatory
- Experience with Robot Operating System (ROS
- Motivated and independent student wanting to work on a real-world robotic project - Solid programming skills in C++ are mandatory - Experience with Robot Operating System (ROS
Please apply via Sirop. If there are any questions, you can reach us via e-mail.
Michael Pantic (michael.pantic@mavt.ethz.ch)
Lionel Ott (lionel.ott@mavt.ethz.ch)
Please apply via Sirop. If there are any questions, you can reach us via e-mail. Michael Pantic (michael.pantic@mavt.ethz.ch) Lionel Ott (lionel.ott@mavt.ethz.ch)