9. Final Remarks
We have presented a completely decentralised and parsimonious mechanism for collision free-operation of beacon transmissions based on learning. The solution aims to solve crucial problems in current wireless mesh networks, such as those based on IEEE 802.15.4e and IEEE 802.11s standards, where beacon transmissions may successively collide making it difficult to discover neighbouring nodes and maintain synchronisation. It can also be used to efficiently support broadcast traffic in receiver-initiated WSNs. The proposed algorithm converges almost surely in finite time and the actual time to convergence in the scenarios of interest is low, making it quite practical for mesh networks involving sporadic mobility. We have also defined how to select the learning parameter in order to: i) keep the time to convergence low and ii) maintain the protocol simplicity and low overhead. Finally, we have considered the practical implications of deploying the presented mechanism considering non-ideal channel conditions as well as taking into account its integration in current standardisation efforts. We believe the proposed approach can be readily implemented in IEEE 802.15.4e and IEEE 802.11s, among others, with some changes in the standards.