Multiple Input Multiple Output (MIMO) systems have garnered a lot of attention recently in antenna system research, and they are often used to improve the quality of service and efficiency of wireless communications. The automotive environment is an especially challenging problem for antenna systems, since the large metal chassis of cars can have a very negative impact on their performance. Modern automobiles also must accommodate an ever increasing number of wireless services, that all require cabling inside the car from the antenna to the transceiver. As the number grows, it is increasingly hard to accommodate them all due to the limited space for antennas and cabling. MIMO systems could thus provide a much needed improvement in these situations. The research presented here is based on the Master of Science thesis by Arttu Rasku from Tampere University of Technology, Finland (http://www.tut.fi). The research work was done in Elektrobit, Finland (http://www.elektrobit.fi)....
CST MICROWAVE STUDIO® (CST MWS) is ideal for automotive simulations for a number of reasons: the geometries can be easily imported and modified using the powerful interface; the robust and accurate PERFECT BOUNDARY APPROXIMATION® (PBA) allows efficient meshing, thus reducing the computational requirements; and the memory usage of the transient solver grows linearly as a function of mesh cells allowing very large scale simulations, and also allowing the simulation of broadband AM/FM, GSM and GPS antennas. In this study, special emphasis was on the placement of the antennas on the car body. The parametric modeling and the parameter sweep features of CST MWS allows easy automatization of the antenna placement, thus reducing the time spent on setting up the simulations.
In the study, two car models were used for simulations: A dummy model, i.e. the so-called basic car model (BCM), and an advanced car model (ACM) based on third generation (E34) BMW 5 series 4-door sedan from the early 1990’s. The BCM model was also used for radiation pattern measurements that were done at VTT Technical Research Center of Finland facilities in Espoo, Finland. For the antennas two types of monopole antennas – cylindrical and flat bar – with center frequency of 700 MHz were used in the simulation, but only the flat bar model was used in the measurements. Due to the limited time and resources, only two-antenna arrays were used for the final measurements.
The final antenna placements used in the measurements are shown in the Figure. 4. Two types of linear arrays were used, transverse and longitudinal. In the first case, the distance of both antennas from the roof rear edge varied between 60 mm and 770, and the spacing between antennas varies between 0.5 and 1.25λ. These configurations are shown as black boxes. In the latter configuration, shown as green boxes, both antennas were placed on the center line, and the spacing between them was varied as shown in the picture.
As can be seen from the Figures 6 and 7, the differences in the simulation results between the two car models BCM and ACM are very small. As can be expected, the ACM model results have more ripples, whereas the BCM model results are quite smooth. The direction of the peaks and nulls agree well, as do the general gain levels. Figure 6 shows a comparison of the simulation results for various distance of the transverse antenna array with 0.5λ element spacing. As can be seen, the variation between the results is very small, thus the effect of the distance from the roof edge is not very large. For the array closest to the roof edge the results show some more ripple, which is an indication that the surface currents on the ground below the antennas are disturbed by the roof edge. Otherwise, it seems that the effect of the edge is quite minimal.
Figure 8 shows the comparison between measurement and simulations for the longitudinal antenna array with 0.5λ element spacing for shortest simulated distance from roof edge (in the measurement setup, the antenna were 20 mm further away from the roof edge). As can be seen, the measurement and simulation results agree very well. Similarly, Figure 9 shows the comparison between two 90° phased antenna elements with 60 mm distance from roof edge and 5/4λ element spacing, but for transverse elevation plane. Both these results show that CST MWS can be easily used to accurately simulate this kind of problems.