An investigation of maximizing multiple mobile sinks data collection speed and data dissemination of sensors in wsn

Abstract

In Wireless Sensor Networks (WSNs), Mobile Sinks (MSs) are deployed for minimizing the nodeand#8223;s energy utilization during data transmission. For delay-aware applications, a Weighted Rendezvous Planning (WRP) approach has been proposed to reduce the energy utilization in which the Mobile Sink (MS) only visits Rendezvous Points (RPs) instead of all sensor nodes. In this approach, an optimal tour was found by assigning a weight to each node according to the amount of data packets that it transmits to the neighbouring RP and their hop distance from the tour. This WRP was performed based on the Steiner Minimum Tree (SMT) and Shortest Path Tree (SPT) for single MS case. Conversely, this is not suitable for multiple MS with different delay constraints. Hence, the major objective of this research is enhancing an energy-efficient MS node selection approach in WSNs. This research work consists of four contributions and they are as follows: In the first contribution of this research, a WRP with Q-learning based Adaptive Zone Partition (WRP-QAZP) approach is proposed for multiple MSs in WSN to handle different delay requirements. Initially, QAZP is applied to split an entire network into a number of regions for multiple MSs. In each region, the location and residual energy of each sensor node including MSs are collected by the Mobile Anchor (MA) node. MA node is a centralized node consisting of the highest amount of all available resources than other nodes. Based on this information, each region is adjusted to the neighboring MS. Then, WRP is applied in each region to assign the weight of each node based on their hop distance from the tour and the total amount of data packets sent to the adjacent RP. The nodes with the highest weights are selected as RPs to transmit the data to the MS at different delay requirements. Thus, the number of multihop communication and energy consumption are iv reduced in multiple MS cases. However, it is difficult to maintain each RPs position at each time by each MSs. Also, the collaboration of