Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Mapping for Online Path Planning and 3D Reconstruction
The goal of this project is to create a fast and accurate 3D map of an environment from sensor data from a mobile robot.
Keywords: Autonomous Robots, UAV, Mapping, 3D Reconstruction, Signed Distance Fields.
Singed Distance Fields (SDFs) are often used for local path planning on-board Unmanned Aerial Vehicles (UAVs). An SDF maps every location in 3D space to a field value which represents the distance to the closest obstacle/surface in the scene. This information is invaluable for path-planning and path-optimization on an autonomous mobile robot.
Since computational power on mobile robots is generally limited, SDF implementations usually discretize space into a regular grid of voxels. The voxel size greatly impacts the usability of the map: large voxel sizes allow for fast data integration and low memory usage, but the resulting reconstruction lacks detail and robot safety might be compromised. Small voxel sizes, on the other hand, allow for an accurate distance field and a high quality 3D reconstruction of the scene, but might be too slow for real-time use.
The goal of this project is to overcome this compromise and to implement a map which can create a high fidelity 3D model even with large voxel sizes. One possible approach is to embed high quality textures into the SDF to effectively emulate a higher voxel resolution. This voxel map needs to integrate incoming depth data form a sensor and fuse the new data into the existing map in real-time and provide accurate distance values for path-planning.
Singed Distance Fields (SDFs) are often used for local path planning on-board Unmanned Aerial Vehicles (UAVs). An SDF maps every location in 3D space to a field value which represents the distance to the closest obstacle/surface in the scene. This information is invaluable for path-planning and path-optimization on an autonomous mobile robot.
Since computational power on mobile robots is generally limited, SDF implementations usually discretize space into a regular grid of voxels. The voxel size greatly impacts the usability of the map: large voxel sizes allow for fast data integration and low memory usage, but the resulting reconstruction lacks detail and robot safety might be compromised. Small voxel sizes, on the other hand, allow for an accurate distance field and a high quality 3D reconstruction of the scene, but might be too slow for real-time use.
The goal of this project is to overcome this compromise and to implement a map which can create a high fidelity 3D model even with large voxel sizes. One possible approach is to embed high quality textures into the SDF to effectively emulate a higher voxel resolution. This voxel map needs to integrate incoming depth data form a sensor and fuse the new data into the existing map in real-time and provide accurate distance values for path-planning.
- **WP1**: Familiarization with state-of-the-art mapping approaches for path-planning and for 3D re- construction.
- **WP2**: Design a mapping framework which can create high fidelity 3D models from larger voxel sizes.
- **WP3**: Implement the framework and create benchmarks for comparisons with state-of-the-art ap- proaches.
- **WP4**: Design and conduct experiments with a UAV to evaluate the selected approach.
- **WP1**: Familiarization with state-of-the-art mapping approaches for path-planning and for 3D re- construction. - **WP2**: Design a mapping framework which can create high fidelity 3D models from larger voxel sizes. - **WP3**: Implement the framework and create benchmarks for comparisons with state-of-the-art ap- proaches. - **WP4**: Design and conduct experiments with a UAV to evaluate the selected approach.
- Interest in Computer Sciences, Robotics and Autonomous Navigation;
- C++ programming experience;
- Experience in mobile robotics, Linux, ROS is beneficial.
- Interest in Computer Sciences, Robotics and Autonomous Navigation; - C++ programming experience; - Experience in mobile robotics, Linux, ROS is beneficial.
Interested students please send CV, Bachelor and Master transcripts to Yves Kompis (ykompis@ethz.ch) and Luca Bartolomei (lbartolomei@ethz.ch). **Do not** apply on Sirop.
Interested students please send CV, Bachelor and Master transcripts to Yves Kompis (ykompis@ethz.ch) and Luca Bartolomei (lbartolomei@ethz.ch). **Do not** apply on Sirop.