Braided on Demand Multipath RPL in the Mobility Context

Wireless sensor networks (WSNs) are considered as the key technology for IoT (Internet of Things) applications thanks to their robustness and their deployment ease. The IPv6 Routing Protocol for Low power and Lossy Networks (RPL) is placed as the routing standard for multi-hop WSNs. However, RPL poorly adapts to sensor nodes' movement which rapidly alters the network performance. Therefore, we propose in this paper, a Bayesian model to accurately predict the sensor nodes' speed distributions. Then, we introduce the Mobility based Braided Multipath RPL (MBM-RPL) to support mobility over RPL. MBM-RPL establishes a primary path based on a new routing metric that exploits the predicted sensor nodes' speed values. An alternative path is established to prevent links expiration along the primary path. To evaluate the performance of MBM-RPL, we first validate the accuracy of the Bayesian model using the Cooja simulator. Then, we compare the performance of MBM-RPL with other RPL based approaches in terms of packet loss rate (PLR) and average transmission delay (ADT).