PhD Midterm Evaluator - Mikel Etxeberria Garcia, Mondragon, Spain

Bøgh, S. (Internal examiner)

Activity: ExaminationExternal examination

Description

Thesis title: "Reliable Artificial Vision Techniques for Autonomous Vehicles Applied to Urban Railway"

In this research work, it is analyzed whether train localization and positioning can be derived by applying Deep Learning techniques in Visual Odometry (VO). With this goal, several aspects have to be taken into account. First of all, the system has to be reliable and as accurate as possible. Then, the operation has to be optimized to have a low latency processing time, near real-time response has to be achieved. Finally, as the system has to be implemented in many trains and stations, the installation, commissioning and maintenance cost has to be taken into account.
Period11 Jul 2019
Examinee
Examination held atMondragon University

Keywords

  • Deep Learning
  • Visual Odometry
  • Computer Vision