Overview
Topic status: We're looking for students to study this topic.
Project Summary
Most localization and mapping algorithms are based on appearance. In dynamic environments, it is difficult to model temporal changes. We propose to build maps that contain semantic information. For example, knowing that an object is a person or a door would be useful. A person is a mobile object and a robot should not be surprised if this object is not present next time it visits the place, whereas the disappearance of a door is a rarer event. Knowing that an object is a door also gives information about its possible appearance (closed, opened).
The objective of this project is to build a prototype localization and mapping system that exploits semantic information about the objects in the environment.
Expected outcomes, applications and/or benefits
Expected outcome: A protype system on one of the robots of the CyPhy lab.
Applications and/or benefits: Localization and mapping in dynamic environments is one of the most important problem to solve in robotics.
Required student skills/experience
Basic maths, good programming skills and some enthusiasm.
- Study level
- Vacation research experience scholarship
- Supervisors
- QUT
- Organisational unit
Science and Engineering Faculty
- Research area
- Keywords
- robot, semantic, mapping, localisation
- Contact
- Contact the supervisor for more information
Dr Frederic Maire
