MASTER ASSIGNMENT
Automotive lighting systems such as headlights and side-projectors are becoming more intelligent. For example, adaptive beam shaping and high definition has been made possible by technological advances regarding resolution, speed, definition, sharpness, contrast, color of the illumination systems.
This increasing complexity leads to new demands for both manufacturing and validation. Both internal product specifications and external regulations demand stringent and extensive functional analysis. Light projection quality analysis is one of the most important aspects therein. Projection quality (PQ) analysis system specifications include total analysis time, compactness, projection angles range, intensity range and resolution, near-field or far-field analysis. At IMS we are developing for the best in-line assessment and alignment system for projection quality.
When light projection analysis is integrated within the assembly process, typically it is part of an active alignment station. Here the analysis output is looped back into the measurement system to optimize the measurement and/or the assembly process to optimize the manipulation and fixation of components, in that forming a control system. This way your minimizing faulty end products and reducing waste.
In this assignment, some new technical setup will be built, tested and explored. The work involves design, testing and building of the optical setup, vision system and optomechanics. Next to that a software algorithm will be put in place for active alignment that uses image processing, (mathematical) optimization and maybe machine learning. Colleagues of IMS are experienced in the process and will guide and help you through the matter.
The assignment is about system identification, controlling of the physical system and using conventional controls and machine learning solutions. The setup is up and running and the first steps into the domain of artificial intelligence have been made. With reinforcement learning we are able to align upon the projection into 3 degree of freedom within 5-10 iterations. Nevertheless, we would like to continue this quest into new territory. For instance, more degrees of freedomand/or farfield validation.
A supervisor from IMS will be attached with the project to provide experience with the setup and support with the thesis. A daily supervisor will also be provided by Fraunhofer Project Center at the University of Twente to support with the theoretical and technical aspects of the thesis.
The master thesis will be part of the PRISMA project, a consortium project around Computer Vision applications, to which more information can be provided on request.
Location: IMS Almelo and/or Perron038 Zwolle
Team: IMS Research & Development
Contact Person
