For modern cars, antennas are required for AM reception, FM and TV diversity reception, weatherband reception (USA), terrestrial digital radio, remote control functions, keyless entry, mobile phone for all worldwide used systems, GPS, and in the future, satellite broadcast radio services. Those services cover the frequency range from 150 kHz up to 2.4 GHz. Such kind of a multiantenna system developed for station wagons is presented in this paper. The obtainable FM and TV diversity effectiveness is discussed for several types of antenna arrangements in detail. This value is the number of fictitious completely decorrelated antenna signals and is obtained by virtual test drives. The characteristic of the respective antennas under test is introduced in the software as antenna pattern, measured or calculated with respect to amplitude and phase. During the computer analysis the car with the antennas is driven virtually through a Rayleigh field scenario with desired and undesired signals.
 The only relevant method for improving radio broadcast reception in cars is by using antenna diversity techniques. This has been reported in the past by, e.g., Lindenmeier et al. . For this reason multiantenna and the applied diversity systems find a worldwide interest and are already in use in a great variety of vehicle models [Lindenmeier et al., 1998]. Under consideration of a reasonable effort, 3 to 4 antennas are regarded as an appropriate number of antennas. With sedan type cars compact antenna systems integrated in the backlite or in the windscreen of the car are used for FM and TV antenna diversity reception.
 Such a complex antenna system is applied to the rear window of the new DaimlerChrysler S-class, consisting of a variety of almost inconspicuous antennas for different communication services (Figure 1). The antenna conductors are partly printed on the glass and partly embedded between the sheets of the compound glass. The heater grid consists of thin embedded wires between the bus bars. Including the mobile phone antenna combined with an active GPS antenna, this rear window contains 12 antennas. The lower region of the backlite is covered by a resistive foil in order to reduce infrared transmission for heat reduction. The space above is used to implement an AM antenna structure. All antennas operate as active antennas, with amplifiers placed on a communication bar, which is mounted underneath the vehicle roof and contains the diversity system for FM reception as well. An additional service for remote control function is also derived from one of the vertical antenna structures.
 Similar compact rear window antenna systems have been designed for other sedan type car models produced by DaimlerChrysler, BMW and Audi. For modern station wagons multiantenna systems for the same variety of services are to be developed the overall performance of which must meet what is known from sedan type cars.
 With sedan type cars in general the rear screen is a large windowpane due to the typical flat mounting situation. For those cars, all antennas for AM, FM, weatherband, TV, keyless entry and remote control can be realized as active rear screen antennas with very good overall performance [Reiter et al., 2000]. With station wagons the rear screen (Figure 2) in general is much steeper and, in consequence, much smaller. Therefore the required multitude of antennas must be spread on rear screen and on one or two of the sidescreens. Otherwise, the diversity effectiveness and therefore the reception quality would be considerably below the standard of today. This will be shown below.
 For all types of cars, it is recommendable to mount the small antennas for mobile phone and the antennas for satellite reception on top of the roof, for example in a small housing, as shown by A8 in Figure 2. With sedan type cars it is also possible to mount those antennas on the upper part of the rear screen (Figure 1).
2. Diversity Effectiveness
 An objective criterion for evaluating the diversity effectiveness has been presented by Lindenmeier et al. . For a better understanding of the diagrams shown below a short review will be given first. The criterion is based upon the improvement factor of the signal quality obtained with the diversity system under test in comparison to the signal quality obtained with only one single antenna. With p representing the likelihood of a signal level not exceeding a certain threshold during the drive the signal quality q is defined as q = 20 log (1/p) with qs referring to ps (single antenna) and qd referring to pd (diversity system) and qs and qd both expressed in dB. The improvement is found to be qd − qs and is also expressed in dB. In a theoretical optimum case of all N antennas being decorrelated and each antenna performing equal likelihood of distortions which results in an equal signal quality qs, the likelihood of distortions in the diversity mode is reduced to pd = psN. Thus the maximum available improvement is qd − qs = (N − 1)qs in dB. The percentage of time with distortions therefore is reduced exponentially by means of antenna diversity.
 In a realistic multiantenna arrangement the antennas are never completely decorrelated. For this reason the ratio pd/ps always is less than the theoretical case of decorrelated antennas. Therefore the diversity effectiveness n is defined as the “number of equivalent decorrelated antennas being effective in a diversity antenna arrangement.” For an evaluated ratio pd/ps or qd/qs respectively n can be obtained by n = qd/qs. The diversity effectiveness n in practice is always less than the number N of antennas applied. The percentage of time with distortions with n decorrelated antennas in a diversity system is therefore:
 The method for evaluating the improvement qd − qs of the antenna system under test is based upon antenna measurements that are taken in the car being placed on a turntable, and on a simulated drive in a Rayleigh-distributed field of electromagnetic waves being incident at the car. Since the antennas on the car are usually sensitive to the angle of polarization of the incident waves in the FM range, measurements are taken for horizontal (“h”), slant (“s” = −45°, “z” = +45° referred to horizontal polarization), and vertical (“v”) polarization as well.
 During the virtual test drive the car is driven in a scenario with wave-scattering and defracting objects near the car (Figure 3), which are assumed to be equiangle distributed over the azimuth for simulation. The desired signal is a superposition of received multipath signals with small delay times arriving at the car in form of many waves of random angle of incidence and amplitude forming a spatial Rayleigh field distribution which is the standard situation with mobile reception in urban areas. For simulation of Rice distributed fields the sources are chosen with a concentration at a certain region of the azimuth. The waves superimpose at the position of the car with the antennas on it and the diversity performance is evaluated. Now a second field is virtually generated by taking a set of waves with independently randomly chosen values for amplitude and phase which relates to the undesired signal which in practice results mainly in cities from multipath propagation with adjacent channel or cochannel interference and in mountainous areas with large delays. With the actually used PC a calculation time of 15 min is required for a four-antenna system if 4 polarizations and 11 spot frequencies spread over the FM band (76–108 MHz) are considered. This corresponds to an evaluated distance of approximately 60 km for each set of antennas.
 With sedan type cars in the FM range, a diversity effectiveness of typically 2.0 to 2.4, depending on frequency and polarization, is obtained. Similar values should be achieved with antenna systems for station wagons as well.
3. FM Diversity Performance of Different Antenna Systems in a Station Wagon
 For FM diversity reception various combinations of the available antennas were analyzed. A1 and A2 are located in the left side screen, A3 to A5 are derived from the rear screen, and A6 and A7 are implemented in the right side screen (Figure 2). In Figure 4, the obtained results for the combination of antennas A3, A4, A5 and a complex sum of A3 and A5 is shown. With horizontal polarization the results are acceptable (average value of 2.4) whereas with slant and vertical polarization the values are not in the desired range (average value with both slant polarizations only 1.8) with individual values at spot frequencies of only 1.6.
 From equation (1) the importance of a high number of n in a multiantenna diversity system can be seen. In order to point this out again we will consider a situation of reception with distortions at ps = 10% of time and a signal quality qs = 20dB respectively with all single antennas. In practice this is a signal that one would listen if the content of the broadcasted program is of interest. With a diversity efficiency of n = 1.6 the resulting signal quality is 1.6*20 dB = 32 dB which is equivalent to a distortion time of 2.5%. With the desired diversity efficiency of n = 2.2 the respective values would be qd = 44 dB and pd = 0.63% which is a reduction factor in distortion time of app. 4. In consequence, further improvements are required especially with respect to slant and vertical polarization.
Figure 5 refers to a system that makes use of all three windows in the rear of the car. Now we have one antenna in both of the side screens (A2 and A7) and two antennas in the rear screen (A3 and A5). The diversity effectiveness is much higher with this arrangement; however, the technical effort is much higher due to the required printing on three screens and a complex wiring with coaxial cables between the antenna amplifiers.
 Therefore in Figure 6 a system is finally considered which makes use of antennas in one side screen and the rear screen (antennas A2, A3, A4 and A5). In spite of a decreased performance mainly with horizontal polarization in comparison to Figure 5, the respective values at slant and vertical polarization in average meet the requirements in combination with a moderate technical effort. Antenna systems that have similar arrangements are in production with a German car producer.
4. TV Diversity Performance
 With increasing frequency and smaller wavelength the diversity efficiency increases considerably. An example of what can be expected is shown in Figure 7 for VHF (175–235 MHz) and UHF (470–870 MHz) frequencies with horizontal polarization. At UHF-frequencies the value n of 4.0 is almost obtained at several spot frequencies. This means that the number n of fictitious completely decorrelated antennas is almost equal to the number of applied antenna signals.
 In contrary to sedan type cars with large rear screens the antennas required for modern multimedia application must be spread over rear screen and at least one side screen with station wagons. Thus the overall performance meets the respective results known from sedan type cars.