Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Semantic-based Path Planning for Autonomous Robot Navigation
The goal of this project is to implement a path planner which uses semantic information in order to perform autonomous navigation and obstacle avoidance for different robots while navigating towards a goal configuration.
In the recent years, the use of robotic platforms has increased exponentially to a wide range of different settings and situations, such as mapping and inspection. In order to enable robots to navigate in complex, unknown and dynamic environments, they must be able to **move with guaranteed safety**, even when facing environments they do not have prior knowledge of. The **full autonomous navigation** still remains a big challenge.
The aim of this project is to develop a **semantic-based planner** that performs trajectory-generation for different platforms, such as Unmanned Aerial Vehicles (UAVs) and ground robots, for **navigation in dynamic environments**. The objective is to utilize the semantic information that can be extracted from a scene in order to steer the robot towards safer areas. The goals are the following:
- Creation of a photorealistic environment for training data collection and for simulation purposes.
- Selection and training of a suitable state-of-the-art semantic segmentation network.
- Design of a path planning framework able to use semantic information.
- Use of the planner for the navigation of a real platform (UAV and/or ground robot).
In the recent years, the use of robotic platforms has increased exponentially to a wide range of different settings and situations, such as mapping and inspection. In order to enable robots to navigate in complex, unknown and dynamic environments, they must be able to **move with guaranteed safety**, even when facing environments they do not have prior knowledge of. The **full autonomous navigation** still remains a big challenge.
The aim of this project is to develop a **semantic-based planner** that performs trajectory-generation for different platforms, such as Unmanned Aerial Vehicles (UAVs) and ground robots, for **navigation in dynamic environments**. The objective is to utilize the semantic information that can be extracted from a scene in order to steer the robot towards safer areas. The goals are the following: - Creation of a photorealistic environment for training data collection and for simulation purposes. - Selection and training of a suitable state-of-the-art semantic segmentation network. - Design of a path planning framework able to use semantic information. - Use of the planner for the navigation of a real platform (UAV and/or ground robot).
- **WP1**: Familiarization with state-of-the-art semantic extraction frameworks and path planning strategies.
- **WP2**: Creation of a photorealistic simulation.
- **WP3**: Selection and training of a suitable state-of-the-art semantic segmentation network.
- **WP4**: Design a path planning strategy for safe navigation using semantic information.
- **WP5**: Design and conduct experiments with multiple mobile platforms (UAV and/or ground robot) to evaluate the selected approach.
- **WP1**: Familiarization with state-of-the-art semantic extraction frameworks and path planning strategies. - **WP2**: Creation of a photorealistic simulation. - **WP3**: Selection and training of a suitable state-of-the-art semantic segmentation network. - **WP4**: Design a path planning strategy for safe navigation using semantic information. - **WP5**: Design and conduct experiments with multiple mobile platforms (UAV and/or ground robot) to evaluate the selected approach.
- Interest in Computer Sciences, Robotics and Autonomous Navigation;
- Python and C++ programming experience;
- Experience in mobile robotics, Linux, ROS is beneficial.
- Interest in Computer Sciences, Robotics and Autonomous Navigation; - Python and C++ programming experience; - Experience in mobile robotics, Linux, ROS is beneficial.
Interested Students please send CV, Bachelor and Master transcripts to Luca Bartolomei (lbartolomei@ethz.ch) and Ruben Mascaro Palliser (rmascaro@ethz.ch). **Do not** send application on Sirop.
Interested Students please send CV, Bachelor and Master transcripts to Luca Bartolomei (lbartolomei@ethz.ch) and Ruben Mascaro Palliser (rmascaro@ethz.ch). **Do not** send application on Sirop.