Center for Adaptive Optics of Valparaíso (CAOVA) received FONDECYT grant to research a system which can be used to measure turbulence inside observatories

Dr. Darío Pérez, professor and researcher of the atmospheric and statistical optics lab of the Physics Institute and Dr. Esteban Vera, from the optoelectronics lab, both who also are the head directors of the Center for Adaptive Optics of Valparaíso (CAOVA) project, received funding to the development of a project which aim is to design a measurement system for the turbulence that can happen inside the Very Large Telescopes, or VLT, in the north part of Chile.

This initiative, entitled Realtime dome turbulence characterization through image motion and scintillation of passive targets -1211848- is going to be a part of the chilean National Funding of Scientific and Technologic Development -FONDECYT, in Spanish-, one of the main programs that are under the management of the chilean National Agency of Research and Development -ANID-. This project, have four years to be done. About this, professor Pérez shared with us some ideas:

How does this research come to mind? Which is the context for this project?

Darío (D): This is something that I’ve been studying for a while. Is a system that a colleague of mine proposed, and that can be use to recollect info about the turbulence, while watching a LED lights matrix distributed in columns. It’s a simple concept, where you observe each individual point with their relatives to obtain information. Now, the technique is in watch an object that is far, far away.

It happens that this have some defects, for example in the anisotropy detection. It doesn’t detect it clearly. It knows that there is, but it can’t quantify it. Another problem that happens is that it does not consider another type of region that aren’t wide open spaces or long distances between the light source and the observer. Because of this, we did an experiment in the lab two years ago. There, we noticed that the numeric original models can’t be applied in the lab, because the propagation phenomena is very short, so it was necessary to evaluate them. So, with one graduate student, we worked in this situation. In the meantime, another student is watching some fluctuations phenomena on big size objects. Thinking on both problems, due to the fact that they have some similarity, I thought on the development of an hybrid technique that can be able to see both situations at the same time, so you can obtain more data. The idea is, so, that this can be applied in a observatory dome, not only in traditional telescopes, but also thinking in the arrival of the ELT, or extremely large telescopes in the following years. Don’t forget that these spaces can affect the observations…

The idea is to introduce a system that can be working 24/7. If this is applied correctly, what we suggest is instead of using artificial light sources such as LED lights, that you use and watch some whites during the night, considering that a minimun light quantity is available so the camera can detect it. The best scenario is, that the system works without affecting the telescope daily routine or labor. So, having a platform that allows to say which is the state of the turbulence anytime inside the dome. This is particularly ambitious, due to the fact that you dispose very little light source, you need to define the appropiate camera. Another novelty that this update on the original technique is that we would use two cameras for the arrangement, to localize the turbulence in space. This opens the space to do telemetry. What is expected to reach with CAOVA is improve the original technique, watching the same white with two or even more cameras in the arrangement. The engineering problem here is, so, see how we work with the volume of input information that each camera is bringing up, to what we have to generate an AI system that can allow classify the sample data aside of the main input available. We’re talking of images of a very high resolution, and these are generated in a big quantity each second away.

Who is going to be part of this project? How is the work plan by now?

D: Prof. Vera will be the co-author, and we will count on various graduate students that are part of the two Center’s lab as well. One part of this first semester is about see which camera is the best candidate. This project have a total work plan for four years. Also, in the first year our focus is the main theoretical model, where we have to inquire some things, as I commented before, and validate our proposal with the use of one white as well. Some of that is previously done with the CAOVA team. For the second year, the analysis comes and also the AI system integration to process what he’s receiving from the cameras. We believe to the end of that year, to make a campaign, that probably will be fine to do, if the health situation suits better to us and is more calm down. For the third year it will be an experimental work with more than one white for telemetry. With that applied, we will be able to do one or more campaigns in the VLT again. If that is succesfully done, we could think in a fourth year with a functional version that can be able to operate in the VLT during a few weeks. With two campaigns done in the VLT, we will be satisfied with the realization of this initiative, which is something that eventually can open a collaboration with the ESO.

What type of equipment and tools are necessary to do this type of research?

D: This project have the focus on hardware build. To what is consider to be the proof of concept, we have compromised two high speed cameras that are very light sensitives. Also, is considered the lens adquisition, and optomechanic equipment, that is useful for us to make optic mounting. All of this is very expensive. Then, a device called digital mirror device is also considered. An optic fiber also, so we can make a quick data transmission with a high sensibility thermometer. After that, we have microcomputers that will work as control systems, service tools and also storage for our data. A big component of this FONDECYT project is support ourselves in the adquisition of specialized equipment. Don’t forget the available infrastructure, that, in the case of my lab, comes from five FONDECYT projects that have been done previously.

Eventually, if the research is successful: which doors would open to further knowledge and inquiries?

D: There are plenty. These systems, which are remote sensing and are designed to calificate the turbulence state, will work as a climate station inside the dome, 24/7, every day. What you are doing with this system, during the measurement of the turbulence at any time is obtaining information of other phenomena that can affect the optical communication, not necessarily  in astronomy-related environments, as one could imagine be to know the vapour concentration amount of water in a forest or a cultivation. From the environmental impact perspective, it will be able to, in a more cheaper version of course, know, following the previous example, how much water you need to do a cultivation in a optimal way. In that sense, for a system that support the monitoring in agriculture contexts, that can be a very interesting idea to explore, because you could be able to distribute better that resource.

To complete this interview: what link can be established between AI and optic, physics and instrumentation issues? why we still do research in these topics?

D: This is something that is being discussed. I believe that AI, to these types of issues that demand big volumes of data, which we know how to analyze but that require a very prepared individual for a long time, is highly efficient. In spite of the sophistication that a well prepared scientific or even graduate student or researchers can have, when it comes to time, it wouldn’t be efficient. The thing is, that a trained system, can support those tasks that can be stressful or cumbersome. You teach the system the parametres in which a person does an evaluation. Learn of the human, after many training sessions. In everything that is remote sensing, the people from my knowledge area is discussing about the possibilities that AI can do when it comes to data collection, in this case, for optics.

To know more about the Center for Adaptive Optics of Valparaíso and the optoelectronics lab of the PUCV’s School of Electrical Engineering, you can check the official website for each one.