In this part, we have proposed a cognitive MAC protocol to establish an overlaying cognitive ad hoc network with QoS provisioning in the presence of the legacy wireless systems. The proposed mechanisms can supplement the insufficiency of the legacy CSMA/CA MAC protocol to fulfill the goals of the cognitive wireless networks. With respect to the four stages in the cognition cycle, we suggest the following techniques:
• Neighbor list establishment in the observe stage: to help CR users having the knowledge of the spectrum usage by the primary and other CR users;
• Improved contention resolution algorithm in the plan stage: to prevent CR users from interfering the existing legacy system and to allow CR users to efficiently and fairly access the channel in the short spare time of the spectrum usage by primary users.
• Invited reservation procedure in the decide stage: to schedule the trans-missions of delay-sensitive traffic with satisfactory QoS requirements for sec-ondary user without interfering the legacy system and to dynamically allocate the bandwidth for various traffic types to avoid the starvation issue for low priority traffic.
• Distributed frame synchronization in the act stage: to distributively co-ordinate the frame transmissions among CR users.
Through the simulations by NS-2, we demonstrate that even in the environment with hidden nodes, the throughput performance of the proposed MAC protocol is at least 50% better than that of the legacy CSMA/CA MAC protocol. The mean access delay and its maximum standard deviation of the proposed MAC protocol are 5 times
0
1000 1500 2000 2500 3000 3500 4000
Offered load (Kbps)
Normalized throughput
New MAC 950bytes DCF 950bytes
Fig. 4.15: Comparison of throughput performance between the proposed cognitive MAC
protocol and the CSMA/CA MAC protocol in the network topology that sec-ondary users are distributed in three separated clusters.
0
1000 1500 2000 2500 3000 3500 4000
Offered load (Kbps)
Dropping rate
DCF 950bytes New MAC 950bytes
Fig. 4.16: Comparison of dropping rate for secondary users between the proposed cognitive MAC protocol and the CSMA/CA MAC protocol in the network topology that secondary users are distributed in three separated clusters.
less than the CSMA/CA MAC protocol. At last, instead of more than 50 % for the legacy CSMA/CA MAC protocol, the dropping rate of delay-sensitive traffic for the proposed MAC protocol is almost negligible.
Chapter 5
Traffic-aware Cognitive Spectrum Handoff with Preemptive Interruption
In this chapter, we discuss the link maintenance issue for CR device when the primary user accesses on the occupied channel during the period of a secondary user’s transmission.
5.1 System Model
Here, we assume that both primary and secondary users uses the slotted system, in which the user’s transmissions on the channel are partitioned into slots. The primary user is assumed to adopt the connection oriented medium access control (MAC) protocol for their data transmissions to meet the quality-of-service (QoS) requirement, such as GSM and WiMax. The secondary user is capable of advocating the channel in the prior slot of the primary user’s transmission by overhearing the reservation information in the legacy system.
Then, we will consider three scenarios for secondary users to continue their transmissions when the primary user appears on the occupied channel. In the first scenario, the secondary user stays in the original channel and postpones its transmis-sion when the primary user appears on the channel, as shown in Fig. 5.1. The CR user will resume the transmission after the primary user completes its transmission.
This non-spectrum-handoff process will be repeated until the secondary user sends all its data. Apparently, the stalled transmission prolongs the transmission time and thus decreases the effective data rate.
The next considered scenario is that the secondary user switches to another channel to proceed its transmission when the primary user appears in the channel.
PU
pPU
Ch 1 SU SU
PU
pPU
SU
Fig. 5.1: An illustration of the transmission scenario that the secondary user stays in the original channel when the primary user accesses on the channel.
This is the so called “spectrum handoff”. To the secondary user, the key issue for the spectrum handoff is the selection of the target channels to continue the on-going transmission. In this case, we categorize two possible scenarios for the target channel selection. Figure 5.2 illustrates the first channel selection scheme. As shown in the figure, the secondary user prepares the list of the target channels for the spectrum handoff before it establishes the link. The channel list can be sent to the receiver during the period of the link setup. When the primary user accesses in the occupied channel, the secondary user can change its transmission on the first channel in the list without waiting for the primary user’s transmissions. Obviously, the effective data rate for secondary users can be significantly improved due to the decreasing transmission time.
However, the pre-determined channel list spectrum handoff scheme in the last paragraph relies on the accurate traffic model to estimate the usage of the primary user on the target channel. The error decision makes the secondary user continuously change its transmission on different channels. The secondary user wastes time on the handshaking for the spectrum handoff, and thus the effective data rate is also decreased due to the erroneous channel prediction. To reduce the error probability of the channel selection, another intuitive way is proposed to determine the target channel after a sensing mechanism, as shown in Fig. 5.3. In the figure, the CR device will perform a wideband radio sensing once the primary user accesses in the
PU
p
PUCh 1 SU
Ch 2
Ch N SU
SU
prediction error
T
oFig. 5.2: An illustration of the spectrum handoff scenario that the secondary user deter-mines the list of target channels before the link is established.
same channel. The secondary user then chooses and changes to the target channel according to the result obtained from the radio sensing. Clearly, the spectrum handoff with radio sensing scheme can avoid the possibility of the erroneous target channel prediction in the aforementioned method. However, this method still spends time on performing the radio sensing, which also prolongs the transmission time and decreases the effective data rate.
Another issue for the two aforementioned spectrum handoff schemes is the num-ber of spectrum handoff trials that a secondary user has to perform during its entire transmission duration. Intuitively, the more the number of handoffs the higher the probability that a secondary user can maintain the established link, whereas the longer the transmission time that a secondary user requires to finish its transmission.
Thus, it requires a way to come to a compromise between the performances of the link maintenance probability and effective data rate in the three secondary user’s transmission scenarios.