Forecasts of cosmological constraints from HI intensity mapping with FAST, BINGO and SKA-I

We forecast the cosmological constraints of the neutral hydrogen (HI) intensity mapping (IM) technique with radio telescopes by assuming 1-year of observational time. The current and future radio telescopes that we consider here are Five-hundred-meter Aperture Spherical radio Telescope (FAST), Baryon acoustic oscillations In Neutral Gas Observations (BINGO), and Square Kilometre Array phase I (SKA-I) single-dish experiments. We also forecast the combined constraints of the three radio telescopes with Planck. We find that the 1σ errors of (w ₀, w_a ) for BINGO, FAST and SKA-I with respect to the fiducial values are respectively, (0.9293, 3.5792), (0.4083, 1.5878) and (0.3158, 0.4622). This is equivalent to (56.04%, 55.64%) and (66.02%, 87.09%) improvements in constraining (w ₀, w_a ) for FAST and SKA-I respectively relative to BINGO. Simulations further show that SKA-I will put more stringent constraints than both FAST and BINGO when each of the experiments is combined with Planck measurements. The 1σ errors for (w ₀, w_a ), BINGO + Planck, FAST + Planck and SKA-I + Planck covariance matrices are respectively (0.0832, 0.3520), (0.0791, 0.3313) and (0.0678, 0.2679) implying there is an improvement in (w ₀, w_a ) constraints of (4.93%, 5.88%) for FAST + Planck relative to BINGO + Planck and an improvement of (18.51%, 23.89%) in constraining (w ₀, w_a ) for SKA-I + Planck relative to BINGO + Planck. We also compared the performance of Planck data plus each single-dish experiment relative to Planck alone, and find that the reduction in (w ₀, w_a ) 1σ errors for each experiment plus Planck, respectively, imply the (w ₀, w_a ) constraints improvement of (22.96%, 8.45%), (26.76%, 13.84%) and (37.22%, 30.33%) for BINGO + Planck, FAST + Planck and SKA-I + Planck relative to Planck alone. For the nine cosmological parameters in consideration, we find that there is a trade-off between SKA-I and FAST in constraining cosmological parameters, with each experiment being more superior in constraining a particular set of parameters.