GPU Acceleration and Game Engines for Wireless Channel Estimation in Millimeter Waves

Abstract

In this article, the intended purpose is to show an innovative technique for estimating the MIMO channel at millimeter wave bands, candidates for mobile 5G technology, by using hardware acceleration, game engines and heuristic algorithms applied to optical ray launching techniques. To verify the performance of the ray launching tool, the normalized Power Delay Profile (PDP) was simulated. The channel was analyzed using the mean square delay error (RMS), the average value of the delay (MD) and the basic propagation loss (PL). The results obtained in computational precision and time were compared with those of a traditional ray tracing tool simulation programmed in MATLAB and with the measurements made in the 57 to 66 GHz range in a specialized laboratory. The results show that the presented technique becomes efficiently profitable from a small number of simulated events (reflections, diffractions).