Sähkö- Ja Tietotekniikan Osasto Sähkötekniikan Koulutusohjelma Implementation of Frequency Hopping Code Phase Synchronization Method for Ad Hoc Networks

Anti-jamming capability and protection against undesired interception are key priorities in ad hoc networks for secure applications. These features are supported by spread spectrum (SS) techniques such as frequency hopping (FH). In a frequency hopping ad hoc network, the phase of the hopping sequence is typically derived from the local time reference (clock reading) of each node. Therefore, network-wide time synchronization is needed in order to get the nodes to simultaneously switch to the same frequency channel, i.e., hop synchronously. Due to the characteristics of ad hoc networks there is no centralized control, e.g., no unambiguous entity, to define a common time reference. Thus, a distributed decision has to be made between the nodes as to whose local time reference is chosen as the common time reference for other nodes to synchronize to. For that purpose, there has to be a predefined procedure for frequency hopping nodes to exchange synchronization information out-of-phase, e.g., through a control channel. This thesis studies the issues and methods related to synchronizing a frequency hopping ad hoc network. One of the studied synchronization methods is implemented on wireless open-access research platforms (WARPs). The implementation consists of a software-controlled frequency hopping routine on top of an already existing physical layer. A common notion of time is achieved throughout the network with the help of a state machine that is responsible for exchanging time information between the nodes. At the initial stage, one of the local time references is chosen as the common time reference by making a distributed decision based on the identifier (ID) numbers of each node. Thereafter, a discrete network synchronization algorithm is applied to maintain the time synchronization. In the measurements, a static two-node scenario is observed where sufficient time synchronization is preserved for the frequency hopping operations. Furthermore, the system is able to maintain the FH-code phase synchronization as long as the clock errors are less than half the dwell time.