In this paper we present empirical results from a study examining the effects of antenna diversity and placement on vehicle-to-vehicle link performance in vehicular ad hoc networks. The experiments use roof- and in-vehicle mounted omni-directional antennas and IEEE 802.11a radios operating in the 5 GHz band, which is of interest for planned inter-vehicular communication standards. Our main findings are two-fold. First, we show that radio reception performance is sensitive to antenna placement in the 5 Ghz band. Second, our results show that, surprisingly, a packet level selection diversity scheme using multiple antennas and radios, multi-radio packet selection (MRPS), improves performance not only in a fading channel but also in line-of-sight conditions. This is due to propagation being affected by car geometry, leading to the highly non-uniform antenna patterns. These patterns are very sensitive to the exact antenna position on the roof, for example at a transmit power of 40 mW the line-of-sight communication range varied between 50 and 250 m depending on the orientation of the cars. These findings have implications for vehicular MAC protocol design. Protocols may have to cope with an increased number of hidden nodes due to the directional antenna patterns. However, car makers can reduce these effects through careful antenna placement and diversity.

/pdf/effect-of-antenna-placement-and-diversity-on-vehicular-2ahzd5r049.pdf

Effect of Antenna Placement and Diversity on Vehicular Network Communications

The Car2Car communication will use a reserved frequency band at 5.9 GHz. One of the key requirements for a Car2Car antenna is that the gain is concentrated in the horizontal plane, which is a problem due to the limited ground plane of a car roof. The influence of roof curvature, roof racks and panorama glass roofs were investigated to quantify them. To the authors knowledge such measurements have not been published yet. Three different antennas were simulated and measured on round ground planes and real car roofs in different configurations. The configuration with a panorama glass roof showed an unexpected high loss of gain in the direction of the glass roof in the range of 15 to 20 dB which result in a drastically reduced communication range to the front direction. This phenomenon is still under investigation.

Influence of car panorama glass roofs on Car2Car communication (poster)

Although the level of realism in vehicular network simulation is constantly increasing, antenna radiation patterns have only rarely entered the picture. In this paper we investigate the impact of these antenna patterns on the outcome of a city-wide simulation with hundreds of cars as well as an isolated intersection collision avoidance scenario. We show that in both cases pronounced differences can be observed: Compared to idealistic isotropic antennas, using realistic antennas changes the distribution of angle of arrival, affecting network topology dynamics. In collision avoidance applications, the antenna radiation pattern can make the difference between a crash and a successful emergency brake, strongly indicating that antenna characteristics should find more consideration in vehicular network simulation.

On the impact of antenna patterns on VANET simulation

IEEE standard methods for measuring electromagnetic field strength of sinusoidal continuous waves, 30 Hz to 30 GHz

The receiving performance of mobile antenna systems that have vertical and horizontal elements at the 920-MHz band has been investigated by measurements in both rural and urban areas in Japan. Three types of inverted-F antennas for a car and three types of antenna elements, a monopole, a bent-slot and an inverted-F antenna, for portable equipment were used in the experiments. The average received power and its probability density function are obtained as the parameter to evaluate the performance of antenna systems. The performances of the bent-slot antenna system for portable equipment and the inverted-F antenna system mounted on a car are compared with those of the monopole or dipole antenna systems used as reference antennas. A notable increase in the average received power has been observed in operation in an urban area where the field distributes randomly and consists of cross-polarized components. >

Experimental results with mobile antennas having cross-polarization components in urban and rural areas

Power gain radiation patterns of mobile antennas mounted in six different locations on a test vehicle were measured with and without typical lights and sirens mounted on the roof. The measurements were performed at frequencies representing the frequency bands of 25 to 50, 150 to 174, 400 to 512, and 806 to 866 MHz. In addition, special antennas consisting of three disguised antennas operating at discrete frequencies of 40.27, 162.475, and 416.975 MHz and one slot antenna operating at 413 MHz were also measured. Plots of power gain radiation patterns are given for the mobile antennas mounted in six different locations on the test vehicle and for the special antennas. Results showing the effects of poor grounding characteristics are also included. Recommended locations for mounting the mobile antennas are given for specific frequency bands.

R L. Jesch

Papers

Measured vehicular antenna performance

Measured Vehicular Antenna Performance