In this paper, we investigate the dynamic rate allocation for multipath routing with the node-disjoint paths in cognitive radio ad hoc networks composed of selfish secondary users (SUs) which are unwilling to forward the packets. We model the path stability factor of each path under a certain traffic session by distributing available channel in view of the carrier frequency of the available uplink channel on each path. Then the traffic sensing factor is also devised to characterize the priorities of different traffic sessions along the node-disjoint paths. In addition, we propose the differential game model of dynamic rate allocation for multi-path routing by taking into account both the mobility and random nature of the selfish SUs. A set of the non-cooperative optimal solutions to the dynamic rate allocation model is further obtained by deriving an optimization problem. Moreover, the distributed optimal rate update rule is developed to dynamically regulate the rate of traffic session over a certain path. Numerical results confirm that the effectiveness of the proposed dynamic rate allocation model.

In this article, an analytical model for predictable contact between two cognitive users in the intermittently connected cognitive radio ad hoc networks is proposed. The salient feature of the model is that it establishes a relationship between contact availability and the intermittently connected cognitive radio ad hoc network design parameters including cognitive user transmission range, velocity of cognitive user, direction of cognitive user, velocity of primary user, direction of primary user, and interference radius of primary user. However, a detailed analytical model is not available in the literature up to now. In particular, the relationship between contact availability and interference radius of primary user has not been derived in previous models. To research the contact availability, the concept of contact degree is first proposed to quantify the probability of contact availability between two cognitive users. In addition, the formulation for the contact degree between two cognitive users from different cases is illustrated in detail in this article. Under the contact degree scenario, the continuous effective contact time is derived to examine the communication duration of cognitive users based on the modified random walk mobility model. Finally, the simulation results demonstrate the effectiveness of the proposed analytical model for predictable contact in the intermittently connected cognitive radio ad hoc networks.