MIRA: a data management and education platform connecting students to robotic telescopes

Imagine the night sky as a giant, cosmic library where every star and planet is a book waiting to be read. For decades, astronomers have been the librarians, using massive, expensive telescopes to check out these books. But what if you wanted to read a book yourself? In the past, that was nearly impossible for a student without a PhD or a mountain of money. You'd need to learn how to drive a giant mechanical eye, wait for perfect weather, and spend hours staring at raw data that looked more like static on an old TV than a picture of a planet.

This is where the world of "robotic telescopes" comes in. Think of these as the library's automated delivery drones. They are telescopes that can move themselves, focus on targets, and take pictures without a human needing to stand next to them in the cold. They are the key to letting regular people, like high schoolers or college students, become real astronomers. However, just having a drone isn't enough; you need a way to ask for the right book, get permission to read it, and then understand what you're seeing. That's the challenge this paper tackles: how to build a system that doesn't just let students point a telescope, but lets them go through the entire process of being a scientist, from writing a request to analyzing the final data, all while the computer handles the tricky parts of moving the telescope and cleaning up the pictures.

---

The MIRA Project: Your Personal Astronomical Internship

Meet MIRA (Mentored Investigations using Robotic Astronomy). Think of MIRA not just as a website, but as a magical bridge connecting a student's bedroom in Switzerland to a real, working telescope in the sky. The goal of this paper is to show how MIRA was built and to test if its "brain"—the part that decides which stars to look at and when—is smart enough to handle a busy night of observing.

The Journey of a Star-Hunter
In the old days, if a student wanted to study a star, they might just get a few minutes of time on a telescope and hope for the best. MIRA changes the game by simulating the real, professional life of an astronomer. Here is how the adventure works:

1. The Proposal (The Request): A student logs in and acts like a scientist. They pick a target (like an exoplanet or a distant star) and fill out a "plan." They have to explain why they want to look at it and how they will do it. It's like writing a grant proposal to a boss.
2. The Review (The Boss's Nod): A teacher or supervisor looks at the plan. They can say "Yes, go for it!" or "No, try again," giving feedback just like a real research committee.
3. The Scheduling (The Puzzle): This is the paper's big star. Once a plan is approved, MIRA has to fit it into a schedule. The telescope can't be in two places at once, and it can't look at a star if it's behind a cloud or if the star is too low in the sky. The system has to solve a giant, moving puzzle: "If I look at Star A now, can I still catch the transit of Planet B later, and will the weather hold up?"
4. The Execution (The Drone Takes Off): The telescope moves itself, takes the pictures (called FITS frames), and sends them back.
5. The Cleanup (The Magic Filter): Raw telescope data is often messy. MIRA automatically cleans it up, calibrates the colors, and measures the brightness of the stars. It then puts the finished pictures in a public archive where the student can download them and use Python tutorials to analyze the results, just like a pro.

The Brain of the Operation: The Scheduler
The most complex part of MIRA is its scheduler. The authors built a "brain" using a method called Bellman dynamic programming. Imagine you are a chef trying to cook a dinner for 10 people, but you only have one stove, the ingredients arrive at random times, and the power might go out for 30 minutes. You have to decide the order of cooking to get the most delicious meal possible without burning anything.

The MIRA scheduler does this for the telescope. It looks at every possible second of the night (down to one-minute intervals) and calculates the best order to observe stars. It weighs different factors:

• Priority: Is this a time-sensitive event, like a planet passing in front of a star (a transit)?
• Air Quality: Is the star high in the sky (less atmosphere to look through) or low down (more atmosphere)?
• Weather: Is it going to rain?

The paper tested this "Bellman planner" against three other common scheduling methods (called astroplan priority, astroplan sequential, and SCOPES) using 100 simulated nights. The results were clear:

• The Winner: The Bellman planner was the most efficient. It managed to schedule an average of 7.0 observations per night, compared to 6.5 for the priority scheduler and only 4.0 for the SCOPES planner.
• The Transits: When it came to catching those tricky, time-sensitive planet transits, the Bellman planner caught 79% of them. The priority scheduler caught 77%, but the others fell far behind (37% and 16%).
• The Quality: Even though it did more work, the Bellman planner didn't sacrifice quality. It kept the "airmass" (how much atmosphere the telescope looks through) very close to the best possible levels, ensuring the pictures were sharp.

The other schedulers had flaws. The "priority" one was greedy; it took the easiest jobs first but left gaps in the schedule, sometimes waiting until the weather turned bad or the star was too low to finish a job. The "SCOPES" planner was too rigid; it packed jobs tightly together but couldn't adapt when a "weather block" (a storm) hit, often leaving the schedule empty or missing the most important transits.

The Architecture: A Team of Specialists
The paper also explains how MIRA is built. Instead of one giant, messy program, it's like a team of three specialists working together:

1. The Web App: The front door where students write proposals and look at data.
2. The Workflow Engine: The project manager that organizes the schedule and the data cleaning.
3. The Telescope Control: The hands and feet that actually move the telescope.

These three parts talk to each other through a shared digital space, so if one part is busy, the others can keep working. This design means the system is robust; if the telescope simulator breaks, the rest of the system doesn't crash.

What's Next?
Right now, MIRA has been tested with a computer simulator that creates fake images of stars. The authors are confident that the scheduling logic works well for the small groups of students they expect to start with. The next step is to take MIRA out of the computer lab and put it on a real 0.5 m telescope at ETH Zurich. They want to see how it handles real-world chaos: actual weather changes, hardware glitches, and the excitement of real students.

The paper concludes that MIRA successfully bridges the gap between classroom learning and professional research. It gives students the authentic experience of proposing, reviewing, and analyzing data, all while the robotic telescope does the heavy lifting. It's not just about looking at stars; it's about teaching the next generation how to think like the scientists who discovered them.