Nearly a century after the discovery that we live in an expanding Universe, and two decades after the discovery of accelerating cosmic expansion, there remains no direct detection of this acceleration via redshift drift - a change in the cosmological expansion velocity versus time. Because cosmological redshift drift directly determines the Hubble parameter H(z), it is arguably the cleanest possible measurement of the expansion history, and has the potential to constrain dark energy models (e.g. Kim et al. 2015). The challenge is that the signal is small - the best observational constraint presently has an uncertainty several orders of magnitude larger than the expected signal (Darling 2012). Nonetheless, direct detection of redshift drift is becoming feasible, with upcoming facilities such as the ESO-ELT and SKA projecting possible detection within two to three decades. This timescale is uncomfortably long given the potential of this cosmological test. With dedicated experiments it should be possible to rapidly accelerate progress and detect redshift drift with only a five-year observational baseline. Such a facility would also be ideal for precision radial velocity measurements of exoplanets, which could be obtained as a byproduct of the ongoing calibration measurements for the experiment.

We propose an experiment, the Cosmic Accelerometer, designed to yield velocity precision of $\leq 1$ cm/s with measurement stability over years to decades. The first-phase Cosmic Accelerometer, which is at the scale of the Astro2020 Small programs, will be ideal for precision radial velocity measurements of terrestrial exoplanets in the Habitable Zone of Sun-like stars. At the same time, this experiment will serve as the technical pathfinder and facility core for a second-phase larger facility at the Medium scale, which can provide a significant detection of cosmological redshift drift on a 6-year timescale. This larger facility will naturally provide further detection/study of Earth twin planet systems as part of its external calibration process. This experiment is fundamentally enabled by a novel low-cost telescope technology called PolyOculus, which harnesses recent advances in commercial off the shelf equipment (telescopes, CCD cameras, and control computers) combined with a novel optical architecture to produce telescope collecting areas equivalent to standard telescopes with large mirror diameters. Combining a PolyOculus array with an actively-stabilized high-precision radial velocity spectrograph provides a unique facility with novel calibration features to achieve the performance requirements for the Cosmic Accelerometer.