Title: A search technique to observe precessing compact binary mergers in the advanced detector era

Abstract: Gravitational-wave signals from compact binary coalescences are most efficiently identified through matched filter searches, which match the data against a pre-generated bank of gravitational-wave templates. Although different techniques for performing the matched filter, as well as generating the template bank, exist, currently all modelled gravitational-wave searches use templates that restrict the component spins to be aligned (or anti-aligned) with the orbital angular momentum. This means that current searches are less sensitive to gravitational-wave signals generated from binaries with generic spins (precessing), suggesting that, potentially, a significant fraction of signals may remain undetected. In this work we introduce a matched filter search that is sensitive to signals generated from precessing binaries and can realistically be used during a gravitational-wave observing run. We take advantage of the fact that a gravitational-wave signal from a precessing binary can be decomposed into a power series of five harmonics, to show that a generic-spin template bank, which is only $\sim 3\times$ larger than existing aligned-spin banks, is needed to increase our sensitive volume by $\sim 100\%$ for neutron star black hole binaries with total mass larger than $17.5\, M_{\odot}$ and in-plane spins $>0.67$. In fact, our generic spin search performs as well as existing aligned-spin searches for neutron star black hole signals with insignificant in-plane spins, but improves sensitivity by $\sim60\%$ on average across the full generic spin parameter space. We anticipate that this improved technique will identify significantly more gravitational-wave signals, and, ultimately, help shed light on the unknown spin distribution of binaries in the universe.