GPU accelerated iterative physical optics to predict RF propagation in urban environments

Multi-objective path planning during autonomous navigation of rotorcraft is very important to accomplish critical missions in dangerous territories. Besides other objectives, it is desired to know in advance if two-way communications between rotorcraft, and downlink and uplink communications with ground control stations will be maintained during a critical mission. In addition, it is required to predict how communication links change if the scene of the autonomous flight is constantly varying. Hence, it is critical to instantly compute the coverage of antennas on rotorcraft or at ground stations in an urban environment. However, fast computation of antenna coverage becomes very challenging as the scene gets larger compared to a wavelength. In this paper, a technique based on iterative physical optics (IPO) accelerated using graphical processing units (GPU) is presented to predict radio frequency (RF) propagation in urban environments. Results of physical optics (PO) and IPO for RF propagation in an urban environment will be presented along with the speedup achieved using GPUs compared to central processing unit (CPU) based implementations.