Energy Efficient Routing Protocol for Heterogeneous Wireless Sensor Networks

Abstract

Wireless sensor networks (WSNs) are the cutting-edge technologies of the twentyfirst century that link people to the outside world. WSNs are networks of separate sensor nodes dispersed throughout a certain area, to keep track of physical or environmental factors like climate, noise, pressure, etc. This group of sensor nodes sends and receives the signals through a wireless medium to a central region, known as a Sink node or Base Station (BS) within the network. Based on attributes, WSNs are categorized formally into two main categories, Homogeneous WSNs (HoWSNs) and Heterogeneous WSNs (HWSNs). The sensors in HoWSNs have similar sensing, processing, and communication capabilities, as well as similar energy constraints whereas, HWSNs may vary in processing power, memory capacity, sensing capabilities (e.g., temperature, humidity, light, vibration), communication protocols, and energy sources (e.g., battery-powered, energy harvesting). Due to such properties, WSNs have various applications in numerous domains that make them versatile and popular in the fields of sensor technology, like, environmental monitoring, industrial automation, healthcare, smart agriculture, home automation, and surveillance. WSNs have many uses, but they also have several drawbacks and difficulties. Some key issues include energy efficiency, low bandwidth problems, reliability, fault tolerance, and many others. Sensor nodes poses limited processing capabilities and memory. But, among all the major challenges, energy efficiency is the prime area of concern. To balance the energy inside the WSNs, various routing protocols have been designed and discussed in the last few decades. This study explores the details of four distinct energy-efficient routing algorithms for heterogeneous wireless networks, highlighting their importance in optimizing network performance and extending network lifetime. Four HWSN major routing protocols have been proposed in this work, namely, Dynamic Cluster-Based Protocol (DCBP), Deep-Q-Routing Based Inter-Cluster Data Aggregation Technique (DQRBIDAT), Reinforcement Based Energy Aware Protocol (RBEAP), and Multi-Objective Differential-Evolution-Based Clustering Protocol (MDEBC). The proposed protocols are simulated with different versions of MAT LAB software and findings are compared with the existing standard protocols. First routing protocol which is Dynamic Cluster Based Protocol (DCBP), aims to improve the Network Lifetime (NL) and Stability Period (SP) using a dynamic clusterbased approach. Initially, the network is divided into small clusters using adaptive clustering where clusters are managed by the Cluster Heads (CHs). Afterwards, a general variable neighborhood search is used to obtain the energy-efficient paths for inter-cluster data aggregation which helps to communicate with the sink. The performance of the proposed method is compared with competitive energy-efficient routing protocols such as EDAL, OIO, and NMCO, in terms of various factors such as Stable Period, Network Lifetime, Packets sent to Base Station (PBS), and Packets Sent to Cluster Head (PCH). Extensive experiments prove that this protocol improves SP and NL than the existing protocols by at least 1.3765% and 1.7817%, respectively. An optimal path is also the foremost measure to balance energy and maintain the performance parameters in various conditions. The second routing protocol focuses on the optimal path problem to preserve energy efficiency inside the network where on the optimal path problem to preserve energy efficiency inside the network where a Deep Q-routing-Based Inter-Cluster Data Aggregation Technique (DQRBIDAT) is designed to improve the inter-cluster communication in WSNs. In every epoch, the Recurrent Neural Network (RNN) is considered to compute the shortest path between cluster heads and sink. The network is trained in such a way that it considers various features of WSNs and can decide which sensor node will be selected as the next hop to establish the shortest path between elected cluster heads and sink. The non-cluster head nodes may also be considered while selecting the shortest path. The findings are compared with well-established protocols such as GEEC, HPSO, SEED, and FHML. Extensive experimental results show that this protocol outperforms competitive energy-efficient protocols. The third protocol, Reinforcement Based Energy Aware Protocol (RBEAP) is used to improve the network lifetime and stability period was designed using a reinforcementbased technique and implemented in such a way that a State-Action Reward-StateAction (SARSA) is used for learning a Markov decision process. Extensive experiments are considered to evaluate the significant improvement of the protocol. Wellvi established protocols such as GEEC, HPSO, SEED, and FHML are used to perform a comparative study. A comparison analysis reveals that the suggested protocol uses less energy than the other protocols. The average improvement in terms of the last node dead, packets transmitted to the cluster head as well to the base station, and stable period have been analyzed. The fourth protocol presented in this work is Multi-Objective Differential-EvolutionBased Clustering Protocol (MDEBC) protocol in which nodes are distributed among multiple regions where all the nodes are divided into certain clusters. The clusters automatically elect the cluster heads (CHs), which are carriers of data to the sink node. Multi-objective differential evolution is used to select the shortest path among the CHs and the base station. Results have shown that this protocol improves the execution of the network field using heterogeneous nodes. The four different protocols are proposed, implemented, and compared with existing standard protocols. These protocols have their own merits like DCBP uses dynamic threshold to select cluster head. This protocol also uses adaptive clustering which saves the energy of nodes so that NL increases. DQRBIDAT protocol uses RNN to compute the shortest path between CHs and sink. The merit of this protocol is that no-cluster heads may also be considered while selecting the shortest path. RBEA protocol is proposed for wireless sensor media networks to preserve more energy during its working. Finally, the MDEBC protocol is proposed to find the shortest path when multiple criteria become the deciding factor in the WSN. In conclusion, a comparative analysis of these protocols and application domainbased protocols has been studied. The work can be extended and optimized using other meta-heuristic-based approaches.