No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Seamless transition from simulation to experiment in micro-optics assembly process optimization
Abstract
Digital twin of a test assembly setup was developed to simulate the different steps of the alignment sequence of a coupling system. Furthermore, a software framework was created that can control both the assembly setup and the simulation engine. According to the tests, the simulated results correlate well with the experimental data, even under different environmental conditions, such as the presence of background noise. This approach can enhance the development of assembly sequences for different kind of micro-optics modules
Despite major progress in developing brilliant laser sources a huge
potential for cost reductions can be found in simpler setups and
automated assembly processes, especially for large volume applications.
In this presentation, a concept for flexible automation in optics
assembly is presented which is based on standard micro assembly systems
with relatively large workspace and modular micromanipulators to enhance
the system with additional degrees of freedom and a very high motion
resolution. The core component is a compact flexure-based
micromanipulator especially designed for the alignment of micro optical
components which will be described in detail. The manipulator has been
applied in different scenarios to develop and investigate automated
alignment processes. This paper focuses on the automated alignment of
fast axis collimation (FAC) lenses which is a crucial step during the
production of diode lasers. The handling and positioning system, the
measuring arrangement for process feedback during active alignment as
well as the alignment strategy will be described. The fine alignment of
the FAC lens is performed with the micromanipulator under concurrent
analysis of the far and the near field intensity distribution. An
optimization of the image processing chains for the alignment of a FAC
in front of a diode bar led to cycle times of less than 30 seconds. An
outlook on other applications and future work regarding the development
of automated assembly processes as well as new ideas for flexible
assembly systems with desktop robots will close the talk.
One of the fundamental difficulties in engineering design is the multiplicity of local solutions. This has triggered much effort in the development of global search algorithms. Globality, however, often has a prohibitively high numerical cost for real problems. A fixed cost local search, which sequentially becomes global, is developed in this work. Globalization is achieved by probabilistic restarts. A spacial probability of starting a local search is built based on past searches. An improved Nelder–Mead algorithm is the local optimizer. It accounts for variable bounds and nonlinear inequality constraints. It is additionally made more robust by reinitializing degenerated simplexes. The resulting method, called the Globalized Bounded Nelder–Mead (GBNM) algorithm, is particularly adapted to tackling multimodal, discontinuous, constrained optimization problems, for which it is uncertain that a global optimization can be afforded. Numerical experiments are given on two analytical test functions and two composite laminate design problems. The GBNM method compares favorably with an evolutionary algorithm, both in terms of numerical cost and accuracy.
We demonstrate a tiny optical module which integrates an ultra-compact wavelength tunable laser, InP-based transmitter and PLC-based receiver. The optical module exhibits low-insertion loss (<13.6 dB) and high bandwidth (>45 GHz) for ≥400-Gb/s coherent-optical links.
Understanding signal and noise quantities in any practical computational imaging system is critical. Knowledge of the imaging environment, optical parameters, and detector sensitivity determine the signal quantities but often noise quantities are assumed to be independent of the signal and either uniform or Gaussian additive. These simplistic noise models do not accurately model actual detectors. Accurate noise models are needed in order to design optimal systems. We describe a noise model for a modern APS CMOS detector and a number of noise sources that we will be measuring. A method for characterizing the noise sources given a set of dark images and a set of flat field images is outlined. The noise characterization data is then used to simulate dark images and flat field images. The simulated data is a very good match to the real data thus validating the model and characterization procedure.
ZOS-API using Python.NET
- Humphreys