Detection of collision using optimized deep model and mitigation of collision using dolphin ant lion optimizer in wireless sensor network

摘要

Wireless sensor network (WSN) comprises automatic sensors that are dispersedinto a huge region. WSN is constructed from huge sensors, which isallocated to a particular task and the majority of task involves reporting andmonitoring. However, as the network can be extended to several sensor nodes,there is a high chance of collision. Thus, this paper devises a novel techniquefor performing both collision detection and mitigation in WSN. Initially, thesimulation of WSN is performed, and then the selection of cluster head is doneusing fractional artificial bee colony (FABC). Here, the network-based parameteris extracted that involves received signal strength index (RSSI), prioritylevel, delivery rate, and energy consumed. The deep recurrent neural network(DRNN) is adapted for collision detection. Here, the training of DRNN is doneusing lion crow search optimizer (LCSO). After collision detection, the collisionmitigation is performed with a pre-scheduling algorithm, namely dolphinant lion optimizer (Dolphin ALO). Here, fitness is considered for collision mitigationthat includes energy, sleep index (SI), delivery rate, priority level,E-waste, and E-save. The proposed method outperformed with the smallestenergy consumption of 0.185, highest throughput of 0.815, highest packetdelivery ratio (PDR) of 0.815, and highest collision detection rate of 0.930.