Mobility prediction for directional networking

Directional phase array antenna (PAA) has emerged as one of the most promising technologies to provide a fundamental breakthrough in mobile ad hoc network capacity. Compared to omni-directional antennas, the narrow beams of PAA provide high-gain high-rate links with longer range, low probability of detection (LPD), low probability of interception (LPI) and anti-jamming (AJ) properties. These desirable properties make directional networking technology an ideal candidate to support information superiority in future combat communication systems. However, the directionality of RF beam raises significant network protocol design challenges that rarely occur in networks with omni-directional antennas. A fundamental issue is that, in order to communicate, the sender and receiver have to point their antennas at each other at the moment of communication. This is a challenging task considering the highly dynamic battlefield environment and the absence of a global information database in practical applications. To solve this problem, we propose a mobility prediction scheme that can provide accurate node position information at the time of a directional link communication by closely tracking node movement. Specifically, we suggest a practical mobility model that is suitable for battlefield environment. Based on a node´s mobility model, our mobility prediction algorithm calculates future positions using parameters specified by the mobility model. An error correction filter is further applied to the prediction results to remove system and other consistent errors. The filter is an exponentially-weighted moving average filter continuously trained by measurement update. We validated our mobility prediction algorithm via simulations and analytical calculations. The results showed that our algorithm is capable of tracking node movement in a highly dynamic battlefield environment and thus can serve as a key enabling technology in directional networking