Fig. 4: Measurements of applied platform acceleration in the presence of laser intensity noise.

Measured phase shifts obtained from sinusoidal fits to interference fringes for (a) Gaussian and (b) error-robust pulses as the applied platform acceleration is varied from −100 to +100 μg. The measurements were repeated for interferometers with interrogation times T = 5 ms and T = 10 ms and for 0 and 20% applied laser intensity noise. Phase shifts are obtained by performing sinusoidal fits with fixed period to interference fringes comprising 33 data points, obtained by varying the DC phase offset of the final interferometry pulse in the interval [0, 2π); the value of the applied acceleration a_applied is taken as the average acceleration measured by the classical accelerometer at 40 kSa/s during all the interferometric measurements of duration 2T that comprise a fringe, with horizontal error bars of magnitude 3 μg drawn for ±1 standard deviation of these samples. Vertical error bars denote ±1 standard error in the phases obtained from sinusoidal fits to each fringe and are of order 10 μrad with no applied noise, growing to order 100 μrad and 1 rad in the presence of applied noise when using error-robust and Gaussian pulses, respectively. In the absence of intensity noise, we observe a negligible change in fringe visibility as a function of applied acceleration (the 1σ variation in fringe visibility is less than 0.0012 for all sequences and interrogation times). The dotted lines plot ϕ = 6ka_appliedT², corresponding to the theoretically expected scale factor of \({{{{{{{\mathcal{S}}}}}}}}=6k{T}^{2}\).