Routing Cognitive Engine


In Cognitive Radio Networks, secondary users are affected by the the activity of primary users on licensed channels. To solve this problem, most routing protocols select the most stable links/channels, or use the minimum delay routes and switch to other channels when primary users are active. However, these approaches can lead to competition between the secondary users on the most preferred links/channels, and this might cause congestion and affect the quality of the transmissions. We propose a Multi-metric Routing protocol with SErvice Differentiation (MRSED). The main idea of MRSED is to use different routing metrics for different traffic types to select routes that satisfy the specific requirements of each traffic type, and to make load balancing across the network. 

MRSED Architecture

MRSED Architecture

The above figure shows the different modules of MRSED protocol.


MRSED is evaluated using NS2 with CRCN patch. We test its performance under heavy loaded network, we study the impact of increasing the number of flows, number of PUs and number of available channels on its performance. Simulation results show that, compared to other related approaches, MRSED provides better end-to-end delay and increases packet delivery ratio . It also makes load balancing across the network and enhances the link utilization.

As an extension to this work, we plan to investigate the use of reinforcement learning techniques to improve the proposed routing protocol and to complete building the Routing Cognitive Engine. Reinforcement Learning techniques have the ability to perceive the surrounding environment so they can be used in controlling the utilization of the diffe rent routing metrics considered in MRSED and assigning different weights for each of them according to the perceived conditions. They can also be used to dynamically extracting the CR QoS rules that will be used to select the routing metric according to current network conditions.



  • MRSED: A Multi-metric Routing Protocol with SErvice Differentiation for Cognitive Radio Ad Hoc Networks    [PDF]


  • Survey about Routing Protocols in Cognitive Radio Networks                  [PDF]
  • Survey about using Reinforcement Learning in Cognitive Radio Networks [PDF]