Book Description

English: It may quite frequently occur in an ad hoc network that the density of users in a given area becomes relatively high during some period of time. At the same time, network resource scheduling is not done regarding these moments with a higher load, but it is done according to average parameters of the network in order to have a trade-off solution that reaches the highest possible quality of service (QoS) while keeping the assigned resources for it as close to be fully exploited as possible. As a result, it may happen that the number of users in a network willing to transmit information is higher than the available resources in that network. On the other hand, in ad hoc networks, the frequency hopping spread spectrum (FHSS) multiple access technique is widely used in the radio interface. In FHSS, the total available bandwidth is partitioned into a certain number of frequency channels having the same bandwidth. In time domain, there is also a division into time intervals. This thesis is focused on the case in which the time intervals have the same duration, called hop period, and specifically on the case where the hop periods for all transmitters are synchronous. Then, in every hop period, every transmitter is been assigned some of those channels to transmit. It means that every transmitter will transmit through some frequency channels, and after the hop period, the transmission will hop into some other channels; and this will happen recursively after every hop period. When, in every hop period, every assigned frequency channel has been assigned to one (and no more) transmitter, then, there is orthogonality among users or, in other words, the hopping sequences or codes, according to which channels are assigned to each user in every hop, are orthogonal. When the hopping sequences are orthogonal, there is no interference among the users of the system because no transmitter coincides with any other in the same frequency channel at the same time (namely, there are no collisions). Orthogonality can always hold when the number of transmitters is less or equal to the number of available frequency channels in the FHSS system. However, when this condition is no longer satisfied, i.e., when there are more transmitters that may simultaneously transmit than available frequency channels, orthogonality is no longer preserved and, so, interference occurs unavoidably. The interference distribution among users depends on the design of the hopping sequences. With conventional approaches for designing the hopping sequences, interference is extremly unbalanced across all users and so, a completely unfair scenario takes place. This work tackles this problem and proposes a method for designing frequency hopping sequences that achieve at the same time a fair interference allocation across all users and an optimal interference reduction. The proposed scheme is optimal in terms of interference level because the overall number of collisions that take place in every hop is the minimum possible. At the same time, this fairness-oriented approach distributes the collisions among nodes in such a way that the variance of the interference power seen by the nodes is considerably reduced. Thereby, this scheme achieves a much more equitable interference balancing across the nodes. The results displayed by the fairness-oriented scheme in terms of fairness are much better than those ones displayed by the two reference studied schemes for both slow and fast frequency hopping.