An Approach for the Integration of Smart Antennas in the Design and Simulation of Vehicular Ad-Hoc Networks Steffen Moser, Simon Eckert and Frank Slomka Institute of Embedded Systems/Real-Time Systems Faculty of Engineering and Computer Science Ulm University, 89069 Ulm, Germany E-Mail: {steffen.moser | simon.eckert | frank.slomka}@uni-ulm.de Abstract—The radio channel is a limiting resource in vehicular ad-hoc networks. A lot of possible applications in the field of vehicle-to-vehicle and vehicle-to-infrastructure communication will make use of sending messages to destination nodes ad- dressed by geographical coordinates. Beam forming, which is the transmission of a message’s signal explicitly into a specific spatial area by exploiting a phase-delayed array of antennas, can be seen as an approach to reduce the overall channel load in vehicular ad-hoc networks. Due to the implementation of today’s network simulators, the analysis of the advantages and possible drawbacks of beam forming on vehicular ad-hoc network scenarios is hardly possible. In this paper we show a way in which a simulation framework typically used for research on vehicular ad-hoc networks can be extended by a realistic channel simulation and a physical layer simulation which allows studying the beam forming abilities of a smart antenna system. I. I NTRODUCTION Inter-vehicle communication has been seen as one future approach to reduce the amount as well as the impacts of acci- dents on our roads. Typically possible applications based on an inter-vehicle communication platform are giving vehicles the ability to propagate warning messages like information about dangerous street conditions, for example suddenly appearing ice or oil, to other vehicles. At the same time an extension of a driver’s range of sight would be thinkable. This can be interesting, for example, at badly visible intersections or at hidden ends of traffic jams. One approach to implement vehicle-to-vehicle communica- tion is wireless ad-hoc networks. Applied to vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication, the concept of ad-hoc networks is called vehicular ad-hoc net- works (VANETs). While infrastructural nodes like road-side units can be used to extend the type and quality of applications, such components are not necessary for network organization purposes. Both, the network organization and the multi-hop transport of messages from a sender node to the destination node(s) are managed by the network nodes, i.e. the vehicles, themselves. VANETs are quite a challenging field of research, which is due to several well-known reasons: The physical layer (PHY) is suffering from various physical effects of the radio channel, e.g. slow and fast signal fading and Doppler shifts caused by the nodes’ mobility. The medium access control layer (MAC) has to provide a robust, fair and - depending on the application - even a time-constrained access to the radio channel which acts as a shared medium. At the same time, the MAC layer has to cope with packet collisions which occur from hidden or exposed stations. The network layer (NET), which is beyond other tasks responsible for routing messages through the network, has to find a suitable routing strategy for specific situations and requirements of applications. On the one hand, having too less nodes forwarding the messages that they have received will cause a possible loss of information in the network - vehicles might not reach each other. Configuring too many nodes in a way that they resend all of the received messages can easily lead to a too high channel load. As the radio channel is usually a limited resource in VANETs, a highly loaded channel will increase the probability of packet loss due to collisions or medium access timeouts. A lot of studies have been done to analyze and optimize the behavior of the MAC or the NET layer, but typically, the most parts of research and optimization are limited to one layer only. In this paper, we will present a model and a concept for an implementation of a simulation framework which allows a cross-layer analysis of VANETs that combines the radio channel, the antennas and the physical layer with the network. We will show a way to integrate the concept of smart antennas with dynamic radiation patterns into an accurate simulation environment for VANETs. Accurate means that the VANET simulation is based on a radio channel simulation that considers the environment’s effect on radio signal transmission. The paper is structured as follows: In the next section, we will describe a communication scenario that we would like to take as an example to apply our method to and explain the problem statement which is taken as an example for this paper’s study. After then, related work will be introduced. The next section presents basic information about VANET simulations, geocast routing and smart antennas. This will be