Based on many papers, for routing functionality in IoT-WSN networks, we can utilize artificial intelligence algorithms to provide the routing protocol with self-learning and self-adaptive capabilities.
Several studies have attempted to apply reinforcement learning to RPL protocol, with promising results. For example, a study by (https://www.mdpi.com/1424-8220/20/15/4158) presented a Q-learning-based collision probability learning algorithm for LLN networks. The objective of this study was to intelligently utilize the cooperation of multiple communication layers to improve RPL performance. The QL-based mechanism allows sensor nodes to learn to find an optimal forwarding path. They used Contiki/Cooja simulator as a simulation environment.
Can anyone please give guidance about how to apply Reinforcement Learning to RPL protocol using the COOJA simulator?
My main question is, how to execute the idea of applying Reinforcement Learning to RPL in Cooja simulator? What are the steps? Is there an online practical example?
thanks in advance.
………………..