Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Deploying ANYmal in an unknown environment: online mapping for autonomous operations
This project is part of a bigger research effort in the Autonomous Systems Lab (ASL) and Robotic Systems Lab (RSL) aiming to integrate an online SLAM (Simultaneous Localization and Mapping) solution on the quadruped robot ANYmal. ANYmal is deployed in search and rescue scenarios and used for inspect
**We strongly suggest you open this project in SiROP for better handling of embedded links.**
**For more details check out online Google Doc**: https://docs.google.com/document/d/1j2jJANbV_VWu_dYIEE333do5lbDm0WC2nhk9MTQAW5Q/edit?usp=sharing
This project is part of a bigger research effort in the Autonomous Systems Lab (ASL) and Robotic Systems Lab (RSL) aiming to integrate an online SLAM (Simultaneous Localization and Mapping) solution on the quadruped robot ANYmal. ANYmal is deployed in search and rescue scenarios and used for inspection tasks: having an online SLAM solution is essential to completing these tasks. Online SLAM could even be used for other purposes, such as the European Robotics League ( https://eu-robotics.net/robotics_league/erl-emergency/about/index.html ) in Sevillia in 2019, or the Mohammad Bin Zayed International Robotics Challenge ( https://www.mbzirc.com ), which is to be held in Winter 2020 in Abu Dhabi, where both ANYmal and multiple drones will compete to achieve common tasks in unknown environments.
Integrating an online SLAM solution involves testing and evaluating different types of sensors (for Visual Inertial Odometry) with the existing hardware of our legged robot (depth camera, lidar, RTK GPS). The online SLAM solution should allow the robot to navigate in unknown environment, and build an incremental map that can be directly used during the execution of a mission. This project will mainly be based on the usage of the open source framework known as Maplab ( https://github.com/ethz-asl/maplab ).
**We strongly suggest you open this project in SiROP for better handling of embedded links.**
**For more details check out online Google Doc**: https://docs.google.com/document/d/1j2jJANbV_VWu_dYIEE333do5lbDm0WC2nhk9MTQAW5Q/edit?usp=sharing
This project is part of a bigger research effort in the Autonomous Systems Lab (ASL) and Robotic Systems Lab (RSL) aiming to integrate an online SLAM (Simultaneous Localization and Mapping) solution on the quadruped robot ANYmal. ANYmal is deployed in search and rescue scenarios and used for inspection tasks: having an online SLAM solution is essential to completing these tasks. Online SLAM could even be used for other purposes, such as the European Robotics League ( https://eu-robotics.net/robotics_league/erl-emergency/about/index.html ) in Sevillia in 2019, or the Mohammad Bin Zayed International Robotics Challenge ( https://www.mbzirc.com ), which is to be held in Winter 2020 in Abu Dhabi, where both ANYmal and multiple drones will compete to achieve common tasks in unknown environments.
Integrating an online SLAM solution involves testing and evaluating different types of sensors (for Visual Inertial Odometry) with the existing hardware of our legged robot (depth camera, lidar, RTK GPS). The online SLAM solution should allow the robot to navigate in unknown environment, and build an incremental map that can be directly used during the execution of a mission. This project will mainly be based on the usage of the open source framework known as Maplab ( https://github.com/ethz-asl/maplab ).
- Getting familiar with existing online SLAM solutions.
- Getting started with Maplab .
- Getting familiar with the ANYmal software stack.
- Integrate and run online SLAM solutions on-board ANYmal, possibly using even other available sensors, such as Velodyne and RTK GPS (already integrated and available on ANYmal).
- Experimentation and evaluation of the proposed solution against benchmarks.
- Getting familiar with existing online SLAM solutions. - Getting started with Maplab . - Getting familiar with the ANYmal software stack. - Integrate and run online SLAM solutions on-board ANYmal, possibly using even other available sensors, such as Velodyne and RTK GPS (already integrated and available on ANYmal). - Experimentation and evaluation of the proposed solution against benchmarks.
- Solid background in C/C++.
- SLAM and computer vision basics.
- (Optional) Knowledge of ROS is a plus.
- Solid background in C/C++. - SLAM and computer vision basics. - (Optional) Knowledge of ROS is a plus.
Send your CV, and grades to
- Marius Fehr (marius.fehr@mavt.ethz.ch)
- Marco Tranzatto (marcot@ethz.ch)