Skip to main content

Our research team searches for planet-size objects further out than Neptune, the currently identified outermost planet in solar system. Last year, astronomers noticed that the 6 observed dwarf planets outside of Neptune (which are transneptunian objects, also called TNOs), all have elliptic orbits with axes pointing in similar directions. This observation led astronomers to believe there is a huge planet at the opposite direction that shepherded those dwarf planets to their current orientation. We call this hypnotized planet “Planet 9.” In our examination of the TNOs, we are trying to find the Planet 9 as well as learn more about the smaller objects outside of Neptune.

The Dark Energy Survey captures part of the night sky from a group of telescopes in Chile. We use the images taken to identify all the bright points – which could be anything from near-earth asteroid, Kuiper Belt objects, stars or simply noise. Then the major part of work Adi, Tom and I are doing this summer is trying to link these bright points into trajectories, and analyze from the trajectory whether the object is indeed an interesting TNO. The trick in this part is that all images are 2D and we have no idea how far each bright point is from earth, but the linked trajectories would give us hint – farther objects moves less in the sky so distant stars are nearly stationary, while near-earth asteroid moves so fast that their motion is obvious in a few hours.

My job is to simulate TNOs motion in space and sketch their apparent motion that should be seen on our observation site. This exercise helps us link the detections into groups that forms possible trajectories. I also developed machine learning method to filter out good groups from bad ones. The machine learning code is surprisingly efficient in identifying 98% of good candidates in the fake data set we use for testing.

The research offered many challenges to build things that I knew little before, the exploration process built up my problem solving skills. Also, our team had lots of ups and downs as we discovered our various methods work or not. The dead ends and challenges made the research experience more rewarding than I expected. It had been a great time working on a meaningful project with a team of like-minded people.

Our research team searches for planet-size objects further out than Neptune, the currently identified outermost planet in solar system. Last year, astronomers noticed that the 6 observed dwarf planets outside of Neptune (which are transneptunian objects, also called TNOs), all have elliptic orbits with axes pointing in similar directions. This observation led astronomers to believe there is a huge planet at the opposite direction that shepherded those dwarf planets to their current orientation. We call this hypnotized planet “Planet 9.” In our examination of the TNOs, we are trying to find the Planet 9 as well as learn more about the smaller objects outside of Neptune.

The Dark Energy Survey captures part of the night sky from a group of telescopes in Chile. We use the images taken to identify all the bright points – which could be anything from near-earth asteroid, Kuiper Belt objects, stars or simply noise. Then the major part of work Adi, Tom and I are doing this summer is trying to link these bright points into trajectories, and analyze from the trajectory whether the object is indeed an interesting TNO. The trick in this part is that all images are 2D and we have no idea how far each bright point is from earth, but the linked trajectories would give us hint – farther objects moves less in the sky so distant stars are nearly stationary, while near-earth asteroid moves so fast that their motion is obvious in a few hours.

My job is to simulate TNOs motion in space and sketch their apparent motion that should be seen on our observation site. This exercise helps us link the detections into groups that forms possible trajectories. I also developed machine learning method to filter out good groups from bad ones. The machine learning code is surprisingly efficient in identifying 98% of good candidates in the fake data set we use for testing.

The research offered many challenges to build things that I knew little before, the exploration process built up my problem solving skills. Also, our team had lots of ups and downs as we discovered our various methods work or not. The dead ends and challenges made the research experience more rewarding than I expected. It had been a great time working on a meaningful project with a team of like-minded people.