Maximum Likelihood Systematic Effect Modeling and Matched Filtering to Detect Trans-Neptunian Objects with TESS

TL;DR

This paper introduces a novel pipeline using maximum likelihood and matched filtering to detect trans-Neptunian objects in TESS data, enabling automated, rapid, and sensitive surveys for distant Solar System objects.

Contribution

The paper presents a new optimization-based background subtraction and matched filtering framework for TNO detection, improving speed and sensitivity over previous methods.

Findings

Successfully detected known TNOs like 90366 SEDNA and 2015 BP519.

Identified two new TNO candidates through visual observation.

Achieved rapid processing of 100 trajectories in 5 minutes on a GPU.

Abstract

We present a pipeline for searching for trans-Neptunian objects (TNOs) using data from the TESS mission, that includes a novel optimization-based framework for subtracting the effects of scattered light and pointing jitter. The background subtraction procedure we adopt, when combined with a moving average, allows one to see TNOs such as 90366 SEDNA, 2015 BP519 with the "naked eye." Moreover, this procedure also enabled us to identify two TNO candidates via direct visual observation (subsequently identified to be 2003 UZ413 and 2005 RR43). To automate the extraction of candidate TNOs, we apply a matched filter that can be tuned to objects at different distances and orbital inclinations. We also demonstrate the performance of the algorithm by recovering signals of three trans-Neptunian objects automatically with a high level of confidence. We further validate the approach via synthetic experiments that test recovery rate as a function of magnitude and distance. We find that there is a trade-off between distance and magnitude; controlling for magnitude, it is easier to detect faster moving objects. Our method can detect objects at distances of 250 AU for magnitudes of +21, and closer objects at fainter magnitudes. On a single contemporary GPU (NVIDIA A100) the method can search for 100 trajectories on 1000 2048 x 2048 frames in 5 minutes, dramatically faster than previous approaches. This method can be used to perform large scale fully automated surveys for TNOs and potential Planet Nine candidates.