Smart Machines en Computer Vision in Micro/Nano Manufacturing
Bison code: L.28190
Our partners, The Netherlands Space Research Organisation (SRON), Salland Engineering, Phix, the Fraunhofer Project Centre and Bronkhorst work together with the Research Chair NanoPhysics Interfaces @ the Saxion to improve the producibility of MEMS and NEMS devices with a special focus on precision assembly techniques. We look into all aspects of the Smart Industry revolution (3D printing, Computer Vision with Machine Learning, Digital Twinning) for this purpose.
Specifically, in this project we shall: a) implement new and use existing precision motion platforms to perform proof-of-principle experiments in assembly as well as b) develop and use computer vision algorithms (eventually with machine learning) in order to guide precision motion or perform advanced defect inspection of MEMS wafers.
There are several different challenges that we need to tackle:
- Find efficient algorithms and fine-tune the camera systems in order to determine locations of micro (MEMS) objects with very high precision (sub-um)
- Design & implement a 3-axis precision positioning stage for MEMS wafer surface modification (droplet-on-demand) or inspection (prior work available)
- Computer Vision algorithms and camera systems, in combination with Machine Learning, to: i) classify micro-defects on MEMS wafers, ii) detect nanometric motion. iii) detect and classify SMD components on waste PCBs for automatic sorting (work together with the 3S E-Waste group)
The students will work towards solving one or more of these challenges as laid out by our industrial partners. Depending on the first meeting with our Research Chair members as well as the interest or makeup of the group, we shall define exact project tasks.
Project or research goal:
Previous groups already worked on several aspects of this project so the new groups can build on their results.
The goal of this student project is to address the problem statement by either defining a research question or via a design/implementation (V-model) approach. The exact type depends on what aspect a)-c) will be chosen as well as the makeup of the group.
We expect research or design reports, data or – if so agreed at the start – a physical deliverable like a measurement setup or instrument.