MINIMIZING THE CHANNEL SWITCHING EVENTS FOR QOS-BASED ROUTING IN COGNITIVE RADIO AD-HOC NETWORK

摘要

Wireless network connectivity systems have very short capacity to adhere the changes due to spectrum mobility and user interference to maintain the Quality of Service (QoS) parameters during end-to-end routing in Cognitive Radio Ad-Hoc Network (CRAHN). The reconfiguration of the network layer parameters in secondary users is challenging and demanding in case of sudden arrival of primary user on its licensed channel and spectrum mobility. Whenever, secondary user senses the primary user activity called as user interference, secondary user has to switch to any other available channel to continue its transmission. This channel switching increases due to the user interference and spectrum mobility which degrades the average data rate. Hence, it will effect directly on the QoS-based end-to-end routing in CRAHN. The addition of reinforcement learning techniques in network management can reduce the channel switching events and user interference by improving the QoS-based routing. This paper presents an algorithm for channel selection in cross-layer approach to minimize the number of channel switching events for QoS-based routing in CRAHN. The methodology is based on the previous network state observation of the primary user for its channel selection and secondary user will explore it for future routing decisions. It can be implemented using a learning agent in a cross-layer approach and modifying some existing routing parameters of Ad-Hoc On-Demand Distance Vector (AODV) routing protocol. This methodology is also very useful as the existing routing protocol can be modified for Cognitive Radio Ad-Hoc Network (CRAHN). We provide a self-contained learning method based on reinforcement-learning techniques which can be used for developing QoS-based routing protocols for CRAHN. We simulated the proposed algorithm using Cognitive Radio Cognitive Network (CRCN) simulator based on NS-2. The results are evaluated and compared with another routing protocol for CRAHN on the basis of some QoS parameters for the proposed algorithm. The results are evaluated and compared with the existing AODV routing protocol on the basis of some QoS parameters for the proposed algorithm. The proposed methodology can provide the basic use of Artificial Intelligence in routing protocols for CRAHN.