Double Leveled Unequal Clustering with Considering Energy Efficiency and Load Balancing in Dense IoT Networks

Clustering could be one of the important approaches in saving energy and load balancing in IoT networks with high density. This paper uses these two factors to improve the network performance and network routing scheme. This paper proposes a Double Leveled Unequal Clustering routing protocol in which traffic load is shared among cluster members to reduce the energy consumption and dropping probability in crucial nodes by sending packets independent of cluster head. A new clustering approach is also proposed that optimally creates unequal-sized clusters with the lowest common spaces between clusters to reduce the number of cluster heads. This approach aims to fulfill the following purposes: decrease the number of clusters to reduce the number of cluster heads and thus reduce the load on crucial nodes, reduce the energy consumption with optimizing the created clusters, optimize the channel bandwidth usage, reduce the interference between control packets by avoiding congestion using framing times, control the redundant control packets by descending the number of reclusterings, and decrease the drop timeout (i.e., whether a packet can wait for a time duration to be sent. If it cannot wait, it will be dropped automatically) by simultaneous transmissions as well as drop of interference by utilizing multi-channel approach. In addition, nodes can turn off their radios in the semi-sleep mode for saving their energy within their idle times. In this way, overall lifetime of the network can be increased by enhancing lifetime of the nodes. According to simulation results, under the proposed approach, throughput of the network increases while the energy consumption decreases due to less reclustering.