We explore the feasibility of estimating primary cosmic ray composition at high energies from the study of two parameters of Extensive Air Showers (EAS) at ground and underground level with Monte Carlo simulations using the new EPOS and QGSJETII hadr
onic models tuned with LHC data. Namely, the slope and density at a given distance of the muon lateral distribution function are analysed in this work. The power to discriminate primary masses is quantified in terms of merit factor for each parameter. The analysis considers three different primary particles (proton, iron and gamma), four different zenith angles (0$^{circ}$, 15$^{circ}$, 30$^{circ}$ and 45$^{circ}$) and primary energies of $10^{17.25}$ eV, $10^{17.50}$ eV and $10^{17.75}$ eV.
