Map quality assessment for robotics in warehouse environments
Map quality assessment for robotics in warehouse environments
Accurate maps are a critical piece in systems such as autonomous cars or autonomous warehouse vehicle. Currently maps are evaluated manually by a human which takes a significant amount of time and an fully automated or semi-automated tool that can quickly identify bad regions within a map is needed.
The research question to try to answer in this thesis is can we use deep learning tools to evaluate maps with the goal of not needing any ground truth reference. Some preliminary work and results demonstrates yes it is possible but more can be done.
A mobile robotic platform capable of collecting sensor data is available as well as datasets from mock up warehouse environments.
Annonsuppgifter
Annonsör: Örebro universitet
Ansök senast:
Annonskategori: Examensarbete, praktik, uppsats
Intresseområde: Teknik och matematik
Kontaktperson: Unal Artan (Postdoctoral Researcher) unal.artan@oru.se
Webbsida: https://www.oru.se