Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Real-Time Detection of Challenging Obstacles using Deep Learning and a Multi-Camera Setup
The goal of this project is to develop novel deep-learning algorithms for jointly estimating depth and detecting challenging small obstacles using camera data only.
Keywords: Obstacle Detection, Deep learning, Depth Estimation, Computer Vision
Obstacle avoidance is a crucial capability for many mobile robotic applications. When autonomous robots operate in complex places such as warehouses, supermarkets or hospitals, they need to reliably detect all sorts of obstacles in order to safely navigate within the environment.
The goal of this project is to develop novel algorithms which use camera images only to jointly estimate depth and robustly detect challenging objects. These include small obstacles, gaps or hard-to-perceive objects for which traditional stereo matching algorithms would fail. To this purpose, a deep learning model will be designed, trained and deployed on a real robot equipped with multiple cameras in a surround-view setup. The student will be invited to contribute to a common framework and will have the chance to work with close-to-market robot prototypes. Finally, by the end of the project, the student will have developed considerable knowledge in the very interesting topics of computer vision, deep learning and obstacle detection. The hands-on experience with a mobile robotic platform will further augment this experience with a deep understanding of software-hardware interplay.
**What We Offer**
- Contribute to ongoing research in the exciting and highly dynamic field of deep learning based perception.
- Join an active team of developers in a Zurich based robotics startup, and collaborate with the Autonomous Systems Lab: one of the largest robotic research teams in the world.
- Deploy successful algorithms to different robotic platforms.
Obstacle avoidance is a crucial capability for many mobile robotic applications. When autonomous robots operate in complex places such as warehouses, supermarkets or hospitals, they need to reliably detect all sorts of obstacles in order to safely navigate within the environment.
The goal of this project is to develop novel algorithms which use camera images only to jointly estimate depth and robustly detect challenging objects. These include small obstacles, gaps or hard-to-perceive objects for which traditional stereo matching algorithms would fail. To this purpose, a deep learning model will be designed, trained and deployed on a real robot equipped with multiple cameras in a surround-view setup. The student will be invited to contribute to a common framework and will have the chance to work with close-to-market robot prototypes. Finally, by the end of the project, the student will have developed considerable knowledge in the very interesting topics of computer vision, deep learning and obstacle detection. The hands-on experience with a mobile robotic platform will further augment this experience with a deep understanding of software-hardware interplay.
**What We Offer**
- Contribute to ongoing research in the exciting and highly dynamic field of deep learning based perception. - Join an active team of developers in a Zurich based robotics startup, and collaborate with the Autonomous Systems Lab: one of the largest robotic research teams in the world. - Deploy successful algorithms to different robotic platforms.
- Make yourself familiar with our robotic perception framework as well as current state-of-the-art depth estimation and obstacle detection solutions.
- Build upon the state of the art by developing your own ideas and your supervisor's input.
- Design, train and deploy a real-time capable solution for detecting challenging obstacles.
- Design and conduct experiments with a mobile robot to evaluate the selected approach.
- Make yourself familiar with our robotic perception framework as well as current state-of-the-art depth estimation and obstacle detection solutions. - Build upon the state of the art by developing your own ideas and your supervisor's input. - Design, train and deploy a real-time capable solution for detecting challenging obstacles. - Design and conduct experiments with a mobile robot to evaluate the selected approach.
- Strong self-motivation and curiosity for solving challenging robotic problems.
- Previous experience in computer vision, machine learning and/or deep learning.
- Excellent programming skills ideally in Python and C++.
- Experience with Linux, ROS, and typical development tools such as git are advantageous.
- Experience deploying deep learning algorithms on embedded hardware is a plus.
- A very good academic record is desirable but may be compensated by expert knowledge in the areas mentioned above.
- Strong self-motivation and curiosity for solving challenging robotic problems. - Previous experience in computer vision, machine learning and/or deep learning. - Excellent programming skills ideally in Python and C++. - Experience with Linux, ROS, and typical development tools such as git are advantageous. - Experience deploying deep learning algorithms on embedded hardware is a plus. - A very good academic record is desirable but may be compensated by expert knowledge in the areas mentioned above.
If you are interested, please send your transcripts and CV to Fabian Blöchliger (fabian.bloechliger@sevensense.ch), Andrei Cramariuc (andrei.cramariuc@mavt.ethz.ch) and Renaud Dubé (renaud.dube@sevensense.ch).
If you are interested, please send your transcripts and CV to Fabian Blöchliger (fabian.bloechliger@sevensense.ch), Andrei Cramariuc (andrei.cramariuc@mavt.ethz.ch) and Renaud Dubé (renaud.dube@sevensense.ch).