Analysis of Propagation Channel in Urban Street Microcell Environment

In station site selection, i.e., the process of determining locations for installing base stations in mobile communication systems, it is highly advantageous to be able to estimate the propagation loss in a manner that allows accurate modeling of interference from surrounding geographical features and objects on the ground. Moreover, active application of time-space signal processing, represented by Adaptive Antenna Array (AAA) and Multiple Input Multiple Output (MIMO) technologies, has recently been investigated in order to achieve large-capacity and high-speed transmission. In the future, estimation of propagation delays and angles of emission/arrival of radio waves is considered to become important [1]-[6]. Ray tracing is a method for unitary estimation of various propagation characteristics. The ray tracing method regards radio waves emitted from a transmission point as individual rays and traces each ray geometrically as it propagates, going through repeated reflection, diffraction and transmission by surrounding structures until it finally arrives at the reception point. Various propagation characteristics at the reception point can be obtained based on the propagation distance, angles of arrival/emission and electric field (complex amplitude) of the traced ray [7]. Since the processing demands tend to be very heavy compared to the computing power available, it has so far been considered difficult to use the ray tracing method for estimation of mobile propagation environments in urban areas (macrocell environments, in particular). However, with the current exponential growth in processing power of modern computers along with recent advances in studies on high-speed algorithms for ray tracing, it has lately become possible to perform accurate large-scale ray tracing-based simulations of mobile propagation environments in urban areas [8]-[12]. For example, the authors compared the simulation results generated by our previously developed “Urban Macrocell Area Prediction (UMAP)” system with a series of actual measurements. The prediction values matched relatively well with the actual measurements; the errors (cumulative 50% value) were found to be 6 dB for received power, 0.2 μs for delay spread and 3 degrees for angular spread (within a horizontal plane on the base station side) [10]. We also evaluated Cross Polarization Discrimination (XPD), which is one of the significant polarized wave characteristics in urban areas [11] [12]. The value predicted by the simulation was 13 dB for the cumulative 50% value, which is approximately 6 dB larger than the actual measurement result. It should be noted that, in ray tracing targeted at urban areas, the only structures that need to be taken into consideration under normal circumstances are walls of buildings. Signboards, traffic lights, signs and similar objects that exist in large quantities in urban areas are not taken into consideration because the scattering caused by such objects is mainly non-normal scattering (in other words, not geometrical-optical reflection and diffraction) and thus considered not to have significant impact on the propagation characteristics. It has not been fully investigated yet whether this assumption actually holds for real-life applications, however. Consequently, if such non-normal scattered waves are taken into consideration in the ray tracing simulation, there is a possibility that the prediction accuracy can be further improved [13] [14]. This article examines the impact of non-normal scattered waves on the propagation characteristics experimentally, as a