The Transiting Exoplanet Survey Satellite (TESS) is optimized to search nearby M dwarf stars for transiting extrasolar planets, providing new opportunities to study planets orbiting low-mass stars. I seek to take advantage of these opportunities by conducting a survey of the coldest planets in these systems to constrain their occurrence rates and find targets for future mass and atmospheric characterization. To identify these planets in the TESS (and K2) data, I have designed a pipeline to detect both single- and multiply-transiting long-period planets. Based on a simulated planet catalog (Barclay et al., 2018), I anticipate 27 single-transiting and 32 multiply-transiting M dwarf planets with P>20 days from the TESS primary mission. To vet targets, I am creating a convolution neural network using the novel method of converting light curves into Gramian Angular Difference Fields to better recognize transit patterns. As nearly half of the anticipated long-period planets will be detected as single-transits, I look to the TESS extended mission to recover additional transits, which will facilitate finding true periods using ground based follow-up.

{"references": ["Barclay, Thomas et al. (2018). arXiv: 1804.05050", "Morton, Timothy D. et al. (2016). arXiv: 1605.02825", "Dressing, C. & Charbonneau, D. (2015). arXiv: 1501.01623", "Wang, Z. & Oates, T. (2015). arXiv: 1506.00327"]}