Enhancing the RPL Protocol Using an Artificial Neural Network for Sustainable IoT Infrastructure

Internet of Things, an ever-evolving communications paradigm, will play an eloquent role in sustainable development in near future. The backbone of IoT is a Low power and Lossy Network involving devices with constrained power, memory and processing capability interconnected over lossy links. The efficiency of the network largely depends on the design of the protocol at the network layer of the communication stack. To cater specific routing needs of such networks, the IETF has designed and standardized the IPv6 routing protocol for LLNs (RPL). RPL has proved efficient in tackling major issues but has certain routing gaps that need to be addressed for optimal performance. For instance, in standard RPL the routing decision is based on a single metric which leads to the selection of inefficient paths and reduces network lifetime. RPL suffers from unbalanced load distribution, slow convergence and inefficient bidirectional communication. Over the years, RPL has attracted many researchers who have contributed to improving this protocol to meet the requirements of energy efficiency, real-time, scalability and reliability. This chapter aims to provide a comprehensive review of the means and methods adopted by the researchers to enhance RPL protocol using soft computing techniques in an effort towards a sustainable IoT infrastructure. Initially, the enhancements done to RPL protocol using fuzzy logic technique are reviewed. In the later part, the role of evolutionary algorithms like Ant Colony Optimization, Genetic Algorithm and Firefly algorithm in RPL improvisation is reviewed. Finally, a novel Artificial Neural Network based method for improving RPL is proposed in this chapter. The proposed ANN based RPL is seen to improve network parameters like Energy efficiency, Packet delivery ratio, Latency and Control Overhead significantly thus directing towards sustainable IoT infrastructure.