A new ultra-fast Moving Object Discovery Engine for iPTF, ZTF, and beyond

Frank Masci, Adam Waszczak, Russ Laher, James Bauer, Thomas Prince,
George Helou, and Shrinivas Kulkarni

We have developed an efficient and generic Moving Object Discovery Engine
(MODE) designed to operate on transient-source catalogs extracted from
difference-imaging, or on catalogs where stationary sources have been
filtered-out as best as possible. The intent is to support current and upcoming
large time-domain surveys that will detect tens of thousands of astrophysical
transients per night. The moving-object tracklet-forming algorithm uses a high
level of parallelism optimized for the latest in multicore/vector processor
technologies. With minimal vetting of transients detected by the intermediate 
Palomar Transient Factory's (iPTF) real-time difference-imaging pipeline, MODE
can detect moving-objects at completeness and reliability levels of >~ 90% and
>~ 99% respectively to R_PTF ~ 19.5 mag. MODE's performance on iPTF data
demonstrates its adaptability to future surveys, in particular the Zwicky
Transient Facility (ZTF) currently under development. ZTF will survey the
sky ~10 times faster than iPTF using a 47 square-degree field of view, enabling
moving-object discoveries over larger volumes and velocity ranges. It is 
therefore crucial that the discovery process be efficient, reliable, and fast
enough for the follow-up of objects in near real-time.