Sustainable Multipath Routing for Improving Cross-Layer Performance in MANET Using an Energy Centric Tunicate Swarm Algorithm

Abstract

Generally, battery power is a valuable resource for mobile devices in a Mobile Ad Hoc Network (MANET). Therefore, energy efficiency and network lifetime should be taken into account when developing control strategies. However, designing an energy-efficient routing mechanism necessitates consideration of many nodes from many layers, such as remaining energy, overall traffic load, and channel assumptions. The traditional layered strategy is unsuccessful in dealing with power-related issues that might affect all layers of the stack. In this paper, the Energy Centric Tunicate Swarm Algorithm (ECTSA) is proposed to perform the cross-layer routing over MANET. The fitness metrics considered in the ECTSA to improve the cross-layer routing are residual energy, communication cost, Data Success Rate (DSR), and mobility. Additionally, an Adaptive Competition Window (ACW) adjustment is used for minimizing the energy consumption caused by the contentions. The performance of the proposed ECTSA is analyzed by means of energy consumption, Packet Delivery Ratio (PDR), End-to-End Delay (EED), and routing overhead. Next, the existing techniques, namely, CEELBRP and EECRP-PSO, are used to evaluate the efficiency of the ECTSA method. The energy consumption of the ECTSA is 7.1 joules and prolongs the network lifetime up to 1603 s for 50 nodes, which is better when compared to the existing CEELBRP and EECRP-PSO techniques.

1. Introduction

The MANET is an active and self-structuring network that has numerous mobile nodes, and the communication among the nodes is wireless [1]. Due to the shorter transmission range, the nodes in the MANET use multiple hops to perform the communication. The data packets generated by the mobile nodes are broadcast to the adjacent nodes in a multi-hop fashion. Here, multi-hop packet broadcasting is accomplished without any centralized architecture [2,3]. The MANET’s characteristics are restricted memory, restricted processing resources, limited power sources, and higher mobility. On the other hand, the MANET is highly efficient and improves the internet’s access to a range of stable and mobile nodes [4,5]. The MANET is an infrastructure-less network, which causes an extra overhead for maintaining the routing and changes in the network topology [6]. Moreover, the MANET has developed as a promising application for the Internet of Things (IoT), emergency communication, and disaster management [7]. The MANET is considered a highly promising solution in emergency and disaster management. When telecommunication infrastructure fails due to earthquakes or flooding, the MANET is seen as an option to help with rescue operations. Thus, the MANET’s reliability must be ensured in terms of maximized packet delivery ratios [8]. In emergency scenarios, the MANET has to cover a larger area to provide more efficient communication over the entire region. For larger area coverage, the network utilizes a multi-hop communication environment. The possibility of link failure during transmission over larger areas is high due to the mobile nature of the nodes [9]. When the network has to cover a larger area for data transmission, the network is divided into groups with the help of a routing algorithm [10]. The mobile nature of the nodes makes them change their connection from one mobile node to another mobile node, without depending on any other nodes in the network. Thus, the continuous data transmission through the network is affected [11].

The performance and the lifetime of the network rely on its energy resource. Cross-layer design places an emphasis on the optimization of system performance by allowing different layers of the communication stack to transfer state information or synchronize their activities to maximize network performance [12]. Network routing must be optimized across all of its layers in order to support QoS. The possibility of enhancing the routing is raised by cross-layer design [13]. Cross-layer optimization concentrates on collaborative solutions integrating many protocol layers. This motivates the cross-layer design, as the routing needs to be adaptive for network dynamics, mobility, and tackling the constraints [14]. The cross-layer routing techniques in the MANET enhance network management by localizing the route setup and minimizing the routing table of the remaining nodes [15]. Every node endures the additional workload, which causes primary power depletion; therefore, it eventually partitions the network and reduces the network lifespan [16]. The cross-layer design is different from the conventional network design. In conventional network architecture, the each layer of the protocol stack is independently operated, and data are sent only among the nearby layers [17]. However, this conventional protocol stack is inadequate for the MANET due to its direct dependencies among the physical and the upper layers [18]. However, the data are dynamically swapped to the different protocol layers in the cross-layer design. Hence, the data are directly swapped among any layers of the protocol stack [19]. The transmission medium of wireless networks has many restrictions, such as interference, bandwidth restriction, and dynamic topology of the network [20]. Therefore, the Energy Centric Tunicate Swarm Algorithm (ECTSA) is proposed and simulated. The simulation findings show that the ECTSA solves real-world case studies with complex search spaces and produces better optimized solutions than other competing methods. The major contributions of this research are as follows:

• The cross-layer design (ECTSA) is utilized in the network to enhance the performance of the high-layer protocols which contain a restricted wired network.
• The ECTSA cross layer contains an important merit that is exploited to create interoperability among the layers.
• The link stability and energy level of a node’s battery are prevented by the data-broadcasting processes of the MANET.

The structure of this research is as follows: Section 2 elaborates the literature review of existing work on cross-layer routing; Section 3 describes the process of the ECTSA method and its network model, along with mathematical equations; The performance measures and its comparative analysis are explained in Section 4. Finally, the conclusion is stated in Section 5.

2. Related Work

Chen et al. [21] presented the Topological change Adaptive Ad hoc On-demand Multipath Distance Vector (TA-AOMDV) for the dynamic topology of the MANET. Here, the stable path selection approach was designed by considering the queue length, available bandwidth, and residual energy, along with the probability of link stability. Then, the proposed TA-AOMDV estimation was combined with the routing for adapting to the rapid changes in the network. However, the EED of the TA-AOMDV was not less while broadcasting the data packets.

Rajashanthi and Valarmathi [22] developed the multipath routing in the MANET to minimize energy consumption. Initially, the K-medoid clustering was used to group the nodes into clusters. Next, the Opposition Genetic-based Fish Swarm Optimization (OGFSO) was used to discover restricted-energy utilization-based multipath routing over the MANET. However, the designed OGFSO used the minimization of energy consumption as the main objective rather than the other parameters e.g., mobility and distance.

Kumar and Kukunuru [23] implemented the simulated annealing-based disjoint multipath routing for the MANET. The cost metrics used in the simulated annealing were link availability, network load, and rate of energy drain. According to the derived path cost, the simulated annealing discovers the multiple-node disjoint paths over the network. However, the PDR of the simulated annealing-based routing was decreased when there was an increment in nodes.

Alghamdi [24] presented the Cuckoo Energy Efficient Load Balancing on-demand Multipath Routing Protocol (CEELBRP) for the MANET. An adequate routing path was discovered using the CEELBRP according to the residual energy of the nodes. Next, the failure route was swapped for the unused path which has higher residual energy during the path maintenance phase. The designed energy-based routing was used to minimize the routing overhead. However, the developed CEELBRP was mainly concentrated only on the residual energy during path generation.

Shivakumar and Patil [25] developed the Particle Swarm Optimization (PSO)-based Energy-Efficient Cross-layer Routing Protocol (EECRP) to improve communication. The fitness metrics considered in the PSO are residual energy, mobility, and DSR. Accordingly, the PSO was used to generate an optimal path over the network. Moreover, the contention calculated from the MAC layer and residual energy was used to perform the dynamic adjustment of CW, which was used to minimize the energy utilization. However, the designed EECRP-PSO failed to consider the distance to discover the optimal path. If the transmission distance was high, then the energy consumption of the nodes was high in the network.

Swarm Intelligence has been used to detect Cross-Layer Packet Drop Attacks in the MANET (CLPDM-SI) by Premala Bhande [26]. This protocol used a cluster-based collective swarm intelligence detection approach to locate a rogue node in a real-time data collecting system that was being attacked with packet drops. This protocol (CLPDM-SI) was compared with the Adapting Cross-Layer Approach for Detecting and Segregating Malicious Nodes (ACLDSM) methods, which do not use swarm intelligence. False positives based on memory and CPU use were used to identify malicious nodes using heartbeat signals. Adaptive cluster creation was used to overcome deficiencies such as frequent topological changes, which were examined in various ways. However, it was noted that, as the number of nodes grows, the delay grows as well.

Moulay Hicham Hanin, Mohamed Amnai, and Youssef Fakhri [27] developed a new adaptation approach to improve QoS in mobile ad hoc networks based on cross-layer and TCP over protocols. This research offered a new optimization method that was influenced by various modifications to the cross-layer technique for improving decision making. The goal is to reduce the impact of retransmissions of lost packets and energy usage by analyzing and selecting a suitable TCP routing protocol that can improve QoS in the MANET. Obtained from the simulation results and a QoS investigation, the suggestion dramatically improved TCP performance in MANETs. However, it is unable to distinguish between packet losses caused by MANET-specific characteristics and losses caused by congestion.

For effective void management in the MANET, Bagirathan Kaliyamurthi and Anandhakumar Palanisamy [28] have shown geographic routing with a Hybrid Firefly Algorithm (FF) and Galactic Swarm Optimization (GSO). The GSO was utilized in the suggested hybrid design because it has a quick convergence rate. Numerous exploration and exploitation cycle patterns have been used in this approach. Additionally, FF offers exploitation fine-tuning because it is utilized for local search. The data connection layer’s MAC protocol used an optimized channel access procedure. By transmitting a data packet, this proposed HFFGSO supports the network layer routers in their route search. Although this hybrid routing technique proved effective, the total architecture’s calculation time was found to be lengthy.

The Cross-Layer and Energy-Aware AODV (CLEA-AODV) routing has been demonstrated by Mansour et al. [29] to enhance Flying Network performance. Due to their direct influence on the routing process, the suggested methodology concentrates on the least energy, summation energy, and hop count aspects as cost indicators to discover the best route in the network. The residual energy and threshold values were used to compute the routing path by expending the least amount of energy and cumulative output. Conversely, this architecture varies quickly due to its greater mobility.

From the overall literature, it is clear that the continuous movement and unpredictability of the nodes causes a dynamic change in the topology of the network. Consequently, routing between the nodes is regarded as one of the most significant challenges in MANETs. Attaining optimal routes between the nodes is the primary task of routing protocols. Recently, it has been stated that issues cannot be resolved using the conventional layered design. There has been a drop in the procedures that depend on cross-layer communication across various levels in an attempt to enhance the effectiveness of MANETs. Therefore, a novel cross-layer design called ECTSA is proposed in this article.

3. Proposed Method

Wireless networks are facing various difficulties that can be resolved using a novel cross-layer design. Using another cross-layer architecture, an important property of the wireless medium is exploited. The ECTSA for the MANET is developed in this work to achieve an energy-efficient cross-layer routing [30]. Cross-layer configuration alludes to dividing data between layers for effective utilization of organization assets and accomplishing high adaptivity. In a cross-layer configuration, each layer is portrayed by a couple of key boundaries and control processes. The boundary limits are assigned to different layers to assist them and provide the optimal rules for their control processes while concerning the current organization status. Cross-layer configuration is normally planned as an enhancement issue, with advancement factors and imperatives from various layers. The control process in the layers produces the ideal qualities by resolving the enhancement issue. The development of a cross layer is a protocol design approach that exploits the interaction among the layers; the optimization and joint design have occurred more than once in each protocol layer.

3.1. Preliminaries

In this segment, the network model and communication model of MANETs are presented in detail.

3.1.1. Network Model

In the MANET, the locations of nodes is not fixed; therefore, it needs to frequently update its location information to all other nodes. The node location is tracked by GPS, which is used to determine the extra coverage that a node’s transmission accomplishes. A MANET is made up of a mobile node count, which is referred to as $m$ (i.e., $1<i<m$). The mobile nodes are arranged in a rectangular area of $R\times R uni{t}^2$ size. In the defined location, the mobile nodes are deployed uniformly and randomly. The transmission among the mobile nodes is represented by a wireless link among the nodes. The nodes’ interconnections are limited by the radio range, i.e., $R$. The size m is distributed to each mobile node $\left(X_d,Y_d\right)$. The nodes travel about the region because of their mobility and the direction. In a MANET, the mobile node can serve as both a host and a router. Depending on the required needs, a single node in the MANET acts as the source or destination. Even though the nodes are movable, the reference value of the nodes in the MANET could be used to determine their location. The routing method can be used to forward data between the nodes in the MANET [31,32]. The path selection is tough due to the erratic node mobility. The data transmission from the source to the destination can be done in numerous paths by the mobile nodes. The routing protocol requires the appropriate route from source to destination with QoS support. The information is transmitted from the source to the destination when the appropriate pathways are determined with the QoS commitment. The MANET model is given in Figure 1.

3.1.2. Communication Model

This section discusses the mobile ad hoc network’s communication model. The mobility of the mobile nodes, as well as the energy of the individual nodes and the link lifespan supporting communication, all play a role in their interaction. Here, a discussion about the communication mobility model is held below.

Mobility Model

The MANET’s mobile node moves about unpredictably. Despite the lack of dependability, the extent of node mobility is critical. The mobile node movement pattern, as well as the change in position, acceleration, direction, orientation, and kinetic energy of mobile nodes with time, is displayed. The mobility model is used to represent node mobility in the MANET. The communication between the nodes results in poor network quality whenever the intermediary nodes for packet transmission are not in the specified range [33]. They are a trace-based mobility model that offers movement patterns using a simple method and a synthesized mobility model that offers realistic movements. Let $R$ be the mobile node transmission radius. The mobile node is transferred to a new site due to the mobility of the node $V_m$. The dimension of the new position of the movable node can be described as Equation (1):

$$R_t=m \left(X_d,Y_d\right)$$

(1)

where time is stated as $t$; dimension is referred to as $m\left(X_d,Y_d\right)$. Through the mobility $V_m+1$, the mobile node is stimulated to a fresh position inside the communication limits. In the simulated MANET model, a threshold changes the size of the mobile node at the new destination. The connection distances of the nodes are determined to be inside the limit by adding or subtracting the threshold value from the measurement. It is provided by Equation (2),

$$R_{t+1}=m \left(X_d\pm T,Y_d\pm T\right)$$

(2)

where $T$ is the threshold value that allows the node to create a momentary sequence inside the communication range. Since it uses a battery-powered source, the lifetime of the MANET is extended by properly managing the energy linked to individual nodes. Since the lifespan of a system is specified as the time from the start of a program until the nodes in the MANET ran out of fuel, the MANET’s energy model takes into account all radio operational parameters, including sending, receiving, inactivity, and sleep. The identification of the node-expended energy is feasible at the moment of the routing protocol for data transmission. For packet forwarding, the network routes between the nodes that utilize the least amount of energy. The routing table can be used to store such routes.

3.1.3. Cross-Layer Routing

Figure 2 shows the model representation of cross-layer routing. The development of cross-layer routing is a protocol design approach which exploits the interaction among the layers [30].

The information shared among the various layers is used to create adaptability, which mainly depends on the protocol design method. Next, this design considers the channel features in the higher layers, as well as considering the traffic’s stochastic arrival in the lower layers. The definition of the cross layer is derived from the clear dissimilarity among the cross-layer design and layered protocol design, without considering the layered architecture. Similar to the seven-layer open systems interconnection model, the layered architecture separates the whole networking task into various layers and expresses the service’s hierarchy, which is required for using the individual layers. The development of the protocol for various layers is used to realize the service in various layers. Moreover, the direct interaction among the non-adjacent layers is avoided by the architecture design, and the communication among the adjacent layers is restricted to the process calls and responses. The addition of several headers to the data results in data overhead. The requirement for at least one protocol standard per layer is another drawback. It takes a while to establish and implement the standards, since there are so many layers.

3.2. ECTSA Method

Wireless links pose network difficulties, which are overcome using a cross-layer design. For example, consider a traditional transmitter that misinterprets a wireless fault as a network congestion notification. Using additional cross-layer design techniques, one of the most important properties of the wireless medium is utilized. In this work, an energy-efficient cross-layer routing is accomplished by developing the ECTSA for the MANET. This ECTSA selects the optimal path by considering four distinct objective functions: residual energy, communication cost, DSR, and mobility. Among these, the energy, DSR, and mobility are calculated from the network layer. Additionally, the network contention is calculated from the MAC layer and the CW is adjusted according to the residual energy and contention to minimize the energy consumption of the nodes. The flow chart for the ECTSA-based cross-layer routing is shown in Figure 3.

The steps for the flowchart are given as follows,

• Step 1: Start the process of cross-layer routing.
• Step 2: Initialize the parameters present in the network.
• Step 3: Deploy the nodes in a random manner.
• Step 4: Calculate the fitness function parameters.
• Step 5: Update the fitness evaluation with the help of ECTSA-based cross-layer routing.
• Step 6: After the update, the adaptive congestion window gets adjusted.
• Step 7: Once the adaptive congestion window gets adjusted, data are transmitted.
• Step 8: Check the simulation criteria by evaluating the simulation time.
• Step 9: Display the performance measures.
• Step 10: Stop the process.

A clear explanation of cross-layer routing using ECTSA is described at Section 3.4.

3.3. Initialization of Routing Paths

At first, the populations of the ECTSA are initialized as the possible paths between the source node to the destination node, which contains an ID of each node. Subsequently, these possible paths are given to the ECTSA to accomplish the cross-layer routing over the MANET.

3.4. Cross-Layer Routing Using ECTSA

In general, the TSA is one of the swarm intelligence methods to optimize global optimization [34] and imitates the strange actions of tunicates made in oceans, specifically the jet drive and swarm intelligence of their searching process. The process of routing path generation [35] using the ECTSA mainly has two stages: the iterative process and fitness metric estimation.

3.4.1. Iterative Process of ECTSA

An optimal path over the network is identified using the ECTSA based on the residual energy, communication cost, data success rate, and mobility. Generally, the jet propulsion is established in three conditions: avoiding conflicts among the search agents, following the location of the optimal solution, and keeping close to the optimal solution.

The new search agent location $\overrightarrow{A}$ is calculated for avoiding the conflicts among the search agents. This search agent location identification is calculated based on Equations (3)–(6).

$$\overrightarrow{A}=\frac{\overrightarrow{G}}{\overrightarrow{M}}$$

(3)

$$\overrightarrow{G}=c_2+c_3-\overrightarrow{F}$$

(4)

$$\overrightarrow{F}=2.c_1$$

(5)

$$\overrightarrow{M}=P_{min}+c_1.P_{max}-P_{min}$$

(6)

where the gravity force is denoted as $\overrightarrow{G}$; water flow advection in the deep ocean is denoted as $\overrightarrow{F}$; the random numbers generated in the range of $\left[0, 1\right]$ are represented as $c_1, c_2$ and $c_3$; and $\overrightarrow{M}$ denotes the social forces among the search agents. Moreover, the initial and subordinate speeds among search agents are denoted as $P_{min}$ and $P_{max}$. Here, $P_{min}$ and $P_{max}$ are considered 1 and 4, respectively.

The search agent is moved in the direction of the current best solution, once conflict is avoided among neighbors. This helps to reach the optimal solution. The optimal location of the best agent is identified using Equation (7).

$$\overrightarrow{PD}=\left|X_{best}-r_{rand}.\overrightarrow{P_P} \left(x\right)\right|$$

(7)

where $\overrightarrow{PD}$ indicates the distance among the optimal solution $\left(X_{best}\right)$ and location of tunicate $\overrightarrow{P_P} \left(x\right)$; the current iteration is represented as $x$; and the random number generated in the range of $\left[0,1\right]$ is $r_{rand}$.

The location of the tunicate is calculated as shown in Equation (8) to maintain the tunicate location near to the optimal solutions.

$$\overrightarrow{P_P} \left(x\right)=\left\{\begin{array}{c}{X}_{best}+\overrightarrow{A}.\overrightarrow{PD}, if r_{rand}\ge 0.5\\ X_{best}-\overrightarrow{A}.\overrightarrow{PD}, if r_{rand}<0.5\end{array}\right.$$

(8)

Equation (6) returns the updated locations of the best transmission path according to the optimal solution $X_{best}$.

The location of the current tunicate is updated according to the locations of two tunicates for modelling the swarming action, which is expressed in Equation (9).

$$\overrightarrow{P_P} \left(x+1\right)=\frac{\overrightarrow{P_P} \left(x\right)+\overrightarrow{P_P} \left(x+1\right)}{2+c_1}$$

(9)

In this ECTSA, the optimal solution $X_{best}$ is identified based on the fitness function estimation, which is clearly described in the following section.

3.4.2. Fitness Metric Estimation

In this stage, the fitness metric for the ECTSA is estimated using the residual energy, communication cost, DSR, and mobility. These fitness metrics are used to identify the optimal path to accomplish the cross-layer routing over the MANET.

Residual Energy

The traffic load of each node is required to be considered while calculating the residual energy. The total energy consumption of the node is expressed in Equation (10).

$$T{E}_i\left(t\right)=L\left(t\right)\times \left(E_{tx}+E_{rx}\right)$$

(10)

where the traffic load of the node $N_i$ at time $t$ is denoted as $L\left(t\right)$, i.e., $L\left(t\right)=NP/Q{L}_i$; the number of packets saved in the queue $Q_i$ is denoted as $NP$; and the length of the queue is represented as $Q{L}_i$. Moreover, the utilized energy while transmitting and receiving packets is denoted as $E_{tx}$ and $E_{rx}$, respectively. Hence, the primary fitness metric, i.e., residual energy $\left(f_1\right)$, of the node $N_i$ is calculated using Equation (11).

$$f_1=R{E}_i\left(t\right)-T{E}_i\left(t\right)$$

(11)

where the residual energy of the node $N_i$ after transmitting $n$ number of packets is denoted as $R{E}_i\left(t\right)$.

Communication Cost

The communication cost $\left(f_2\right)$ required to communicate with a neighbor node is calculated based on distance, as shown in Equation (12).

$$f_2=\frac{d_{avg}^2}{d_0^2}$$

(12)

where $d_{avg}^2$ denotes the average distance among the node and adjacent node and $d_0^2$ denotes the node’s radius.

DSR

The DSR $\left(f_3\right)$ of each node is calculated based on the Bit Error Rate (BER) of the received data, which is expressed in Equation (13).

$$f_3=1-BER$$

(13)

Node Mobility

The last fitness metric, i.e., node mobility $\left(f_4\right)$, is denoted by means of changes in the node degree $\left(R_{ND}\right)$. A number of adjacent nodes for node $N_i$ is discovered by the Node Degree (ND). The minimum ND of $N_i$ at a network graph $NG$ is shown in Equation (14).

$$N{D}_{min}\left(NG\right)=\min \left\{ND\left(i\right)\right\}, \forall i ϵ NG$$

(14)

Equation (15) shows an average node degree.

$$N{D}_{avg}\left(NG\right)=\frac{{{\displaystyle \sum }}_i^{n}ND\left(i\right)}{n}$$

(15)

The rate of changes in the ND is computed using Equation (16).

$$f_4=\frac{N{D}_{avg}\left(t\right)-N{D}_{avg}\left(t-1\right)}{t}$$

(16)

The derived multiple fitness metrics are converted into a single objective function using Equation (17).

$$Fitness={\alpha }_1{f}_1+{\alpha }_2{f}_2+{\alpha }_3{f}_3+{\alpha }_4\frac{1}{f_4}$$

(17)

where the ${\alpha }_1, {\alpha }_2, {\alpha }_3$, and ${\alpha }_4$ are weight values assigned to each fitness metric. Since this conversion is done because these fitness metrics conflict with each other, the node with higher residual energy is preferred for avoiding node/link failure while broadcasting the packets, which helps to avoid packet loss. A path with less distance to the destination is required for improving energy efficiency. Next, the estimation of the DSR is used to find the path which causes less loss over the MANET. Furthermore, the lesser ND helps to minimize the energy consumption of the node.

3.5. Adjustment of Adaptive Contention Window

The energy consumption caused by the contentions is minimized by using an ACW adjustment technique in the MAC layer. the path with higher contention is required to wait and the nodes with less residual energy are required to accomplish the immediate data transfer. Hence, this ACW is used to minimize energy consumption and improve data delivery. The information shared among the various layers is used to create adaptability, which mainly depends on the protocol design method. Next, this design considers the channel features in the higher layers and it considers the traffic’s stochastic arrival in the lower layers. The performance part of the cross-layered method stems from the selection of appropriate data that are being used.

3.6. Pseudocode

The following parameters are used in the ECTSA: maximum iteration (1000), search agent (30), lower bound (100), upper bound (300), and dimension (30). The major goal of this study is to use the MANET for an energy-efficient cross-layer routing. Here, the protocols are created in accordance with the guidelines of the reference structure. The services of the lower layer are used by the higher layer protocol, which is designed in a layered architecture. The specifics of packet delivery are not restricted by the developed higher layer protocol. Additionally, the lower layer’s services are independent of the details of the higher layer. Using cross-layer design as a foundation, a novel dynamic routing approach called ECTSA is suggested. The analysis and its comparisons are evaluated in the following sections. The pseudocode for ECTSA is represented in Algorithm 1.

Algorithm 1 Pseudocode for ECTSA.

Input$: \mathrm{Tunicate} \mathrm{population} {\overrightarrow{P}}_p$ Output: Optimal fitness value F $\overrightarrow{S}$ Initialize the parameters $\overrightarrow{A}$, $\overrightarrow{G}$,$\overrightarrow{F}$,$\overrightarrow{M}$, and $Ma{x}_{iterations}$   Set $P_{min}$ $\leftarrow$1   Set $P_{max}$ $\leftarrow$4   Set Swarm $\leftarrow$ 0   While (x < $Ma{x}_{iterations}$) do     for i $\leftarrow$ 1 to 2 do       $\overrightarrow{F}$S $\leftarrow$ ComputeFitness ($\overrightarrow{P_p}$) /* Jet propulsion behavior */       $c_1$,$c_2$,$c_3$,$r_{and}$ $\leftarrow$ Rand ()       $\overrightarrow{M}$ $\leftarrow$ $\left|P_{min}+c_1\times P_{max}-P_{min}\right|$       $\overrightarrow{F}$ $\leftarrow$ 2$\times$ $c_1$       $\overrightarrow{G}\leftarrow c_2$+$c_3$-$\overrightarrow{F}$       $\overrightarrow{A}\leftarrow$ $\overrightarrow{G}$/$\overrightarrow{M}$       $\overrightarrow{PD}$ $\leftarrow$ ABS ($\overrightarrow{FS}$ -$r_{and}\times \overrightarrow{P_p}\left(x\right)$) /*Swarm behavior */       If ($r_{and}\le$ 0.5) then         $Swarm\leftarrow Swarm+\overrightarrow{FS}$ + $\overrightarrow{A}$ $\times \overrightarrow{PD}$       else         $Swarm\leftarrow Swarm+\overrightarrow{FS}$ − $\overrightarrow{A}$ $\times \overrightarrow{PD}$       end if     end for       $\overrightarrow{P_p}\left(x\right)$ $\leftarrow Swarm$/(2 + $c_1$)       $Swarm\leftarrow$0       Update the parameters $\overrightarrow{A},\overrightarrow{G},\overrightarrow{F},$ and $\overrightarrow{M}$       $x \leftarrow x+1$    end while $\mathrm{Return} \overrightarrow{FS}$ end procedure procedure $COMPUTE FITNESS(\overrightarrow{P_p)}$ [RE, CC, DSR, Node Mobility]   for i $\leftarrow$ 1 to n do     $FI{T}_p$[$i$] $\leftarrow$ FitnessFunction ($P_p$($\overrightarrow{i}$, : ))   end for   $FI{T}_{Pbest}\leftarrow$ BEST ($FI{T}_p\left[ \right]$)/*Calculate the best fitness value using equation (17) */ return $FI{T}_{Pbest}$ end procedure procedure BETS ($FI{T}_p$)   $Best$ $\leftarrow$ $FI{T}_p$(0)   for i $\leftarrow$ 1 to n do     if ($FI{T}_p\left[i\right]$ < $Best)$ then $Best\leftarrow FI{T}_p\left[i\right]$     end if   end for return Best   /* Return the best ($Fitness$) value */ end procedure

4. Results and Discussion

The results of the ECTSA method are discussed in this section. The implementation and simulation of this ECTSA-based cross-layer routing are done using a Network Simulator-2.34 (NS-2.34), where the system is operated with an i5 processor with 6 GB of RAM. The routing path issue can be overcome by deploying intermediary nodes in conjunction with routing. The MANET’s ensemble relationships are built on the mobile nodes’ unique motion patterns. The mobility model is divided into two versions based on characteristics such as size, mobility magnitude, unpredictability, regional restrictions, and target alignment. The main objective of this research is to perform an energy-efficient cross-layer routing over the MANET. Here, the protocols are developed according to the reference structure’s rules. The higher layer protocol is designed in the layered architecture, which utilizes the lower layer’s services. The developed higher layer protocol does not have a constraint on the details of service delivery. Moreover, the services offered by the lower layer are not dependent on the higher layer’s details. In this research, an IEEE 802.11b MAC layer protocol is used because it is cost-effective, easy-to-set-up, and extensively maintained by manufacturers. The Random Waypoint (RWP) mobility model is used because it is an elementary model which describes the movement pattern of independent nodes in simple terms. Additionally, this mobility model is a simple and straightforward stochastic model that describes the movement behavior of a mobile network node in a given system area, which results in a higher effective speed. Therefore, the Random Waypoint mobility model is used in this research. The simulation parameters of this research are mentioned in

Table 1. Simulation parameters.

Parameter	Value
Number of Nodes	50, 75, 100, 125 and 150
Area	$1000 \mathrm{m}\times 1000 \mathrm{m}$
Propagation Model	Two-Ray Ground
Antenna Model	Omni-Antenna
Traffic Model	CBR
MAC Protocol	IEEE 802.11b
Mobility model	Random Waypoint
Initial Energy	10 Joules
Transmission Power	0.8 watts
Receiving Power	0.5 watts
Simulation time	500 s

.

The performance of the ECTSA method is analyzed by means of energy consumption, packet delivery ratio, end-to-end delay, and routing overhead. Existing research on techniques such as CEELBRP [24] and EECRP-PSO [25] is used to evaluate the efficiency of the proposed method.

4.1. Quantitative Analysis of ECTSA Method

The quantitative measurement of delay and energy consumption is significantly reduced by the proposed ECTSA method. In current years, there has been a rise in attention on the MANET, and one of the most interesting applications is cross-layer design. The ECTSA, its various key indications, and its characteristics are presented in this work. Using cross-layer design as a foundation, a novel dynamic routing approach is suggested. In terms of Delay and Energy Consumption, the suggested ECTSA is compared with the existing Opposition Genetic-based Fish Swarm Optimization (OGFSO) [22]. The proposed and existing approaches are compared in

Table 2. Evaluation Analysis of Optimization Algorithms.

Node Count	Delay (ms)		Energy Consumption (J)
	OGFSO [22]	Proposed ECTSA	OGFSO [22]	Proposed ECTSA
0	0	0	0	0
100	4	2.1	190	150
200	4.2	2.7	200	178
300	4.8	3.9	300	225
400	5	4.2	140	100
500	7	5	50	42

, which clearly shows that the proposed ECTSA produces less delay (5 ms) and energy consumption (225 J), outperforming the existing OGFSO [22], which attains a delay of 7 ms and energy consumption of 300 J.

4.2. Energy Consumption

Energy consumption of the nodes is defined as the amount of energy utilized while broadcasting and receiving the data packets.

Table 3. Performance of Energy Consumption.

Node Count	Energy Consumption (J)
	CEELBRP [24]	EECRP-PSO [25]	HFF-GSO [28]	Proposed ECTSA
50	11.1	10	9.1	7.1
75	11	8.3	8.5	6.8
100	10.5	7.8	7.7	6.3
125	10.3	7.0	7.1	6.0
150	10.3	7.0	6.3	5.4

shows the performance analysis of energy consumption.

Figure 4 shows the comparison of energy consumption for CEELBRP [24], EECRP-PSO [25], HFF-GSO [28], and the proposed ECTSA. The energy consumption of the ECTSA is varied from 5.4 J to 7.1 J, which is less when compared to the CEELBRP [24], EECRP-PSO [25], and HFF-GSO [28]. The developed ECTSA achieves lower energy consumption due to its cross-layer routing, shortest path generation according to the fitness metrics, and CW adjustment. However, the CEELBRP [24] and EECRP-PSO [25] do not considered distance; therefore, the energy consumption is increased because the transmission path has a higher distance.

4.3. PDR

The PDR is defined as the ratio between the amount of received packets and the number of transmitted packets over the MANET, which is expressed in Equation (18).

Table 4. Performance of PDR.

Node Count	Packet Delivery Ratio
	CEELBRP [24]	EECRP-PSO [25]	HFF-GSO [28]	Proposed ECTSA
50	0.61	0.63	0.93	0.94
75	0.70	0.74	0.94	0.95
100	0.76	0.89	0.96	0.97
125	0.83	0.90	0.98	0.99
150	0.91	0.93	-	0.99

shows the performance analysis of the PDR.

$$PDR=\frac{Amount of received packets }{Amount of transmitted packets}$$

(18)

Figure 5 shows the comparison of the PDR for the CEELBRP [24], EECRP-PSO [25], HFFGSO [28], and ECTSA. The PDR of the ECTSA is varied from 0.94 to 0.99. From the analysis, it is concluded that the ECTSA achieves better data delivery than the CEELBRP [24], EECRP-PSO [25], and HFFGSO [28]. The data delivery of the ECTSA is improved by avoiding node/link failure and selecting the path with a high DSR.

4.4. End-to-End Delay

EED is the average time required to perform successful data transmission over the MANET; EED is expressed in Equation (19).

$$EED=\frac{Sum of time taken to broadcast packet in receiver}{No. of packet received by receiver }$$

(19)

Table 5. Performance of EED.

Node Count	End-to-End Delay
	CEELBRP [24]	EECRP-PSO [25]	Proposed ECTSA
50	8.2	7.0	3.1
75	9.0	7.3	3.7
100	10.1	8.0	4.9
125	10.3	8.6	5.2
150	11.4	9.2	5.5

shows the performance analysis of EED. Figure 6 shows the comparison of EED for the CEELBRP [24], EECRP-PSO [25], and ECTSA. The EED of the ECTSA is varied from 3.1 ms to 5.5 ms, which is less than the CEELBRP [24] and EECRP-PSO [25]. The developed ECTSA achieves less EED because of its transmission path with less distance and fewer control packet requirements during path generation.

4.5. Routing Overhead

Routing overhead denotes the number of control packets generated while generating the transmission path.

Table 6. Performance of Routing Overhead.

Node Count	Routing Overhead
	CEELBRP [24]	EECRP-PSO [25]	Proposed ECTSA
50	6009	5028	3510
75	6125	5120	3160
100	6530	5356	2190
125	6530	4789	2357
150	6781	5010	2561

shows the performance analysis of the routing overhead.

Table 7. Performance of Network Lifetime.

Node Count	Network Lifetime (s)
	CEELBRP [24]	EECRP-PSO [25]	Proposed ECTSA
50	1583	1533	1603
75	1517	1501	1588
100	1476	1438	1550
125	1433	1399	1505
150	1390	1371	1490

shows the performance analysis of network lifetime.

Figure 7 shows the comparison of routing overhead for the CEELBRP [24], EECRP-PSO [25], and ECTSA. The routing overhead of the ECTSA is varied from 2561 packets to 3510 packets. From the analysis, it is concluded that the ECTSA achieves less routing overhead than the CEELBRP [24] and EECRP-PSO [25]. The ECTSA-based cross-layer routing does not require a high amount of control packets, because it uses effective fitness metrics to identify the transmission path.

4.6. Network Lifetime

The evaluation of network lifetime is shown in Figure 8. In comparison to earlier CEELBRP [24] and EECRP-PSO [25], the lifetime of the proposed ECTSA has increased. Additional sensor nodes start guiding packets randomly as the network’s node count increases, and there is a good possibility that the node dies at some point. In the proposed ECTSA technique, only the best node is selected to transport packets, providing longer battery life and network lifetime.

Table 8. Performance of Packet Loss Rate.

Node Count	PLR
	CEELBRP [24]	EECRP-PSO [25]	Proposed ECTSA
50	0.62	0.57	0.32
75	0.42	0.36	0.19
100	0.24	0.11	0.07
125	0.17	0.1	0.06
150	0.09	0.07	0.02

shows the results for the comparison of network lifespan performance.

Figure 8 shows the performance analysis of network lifetime. As a consequence, the suggested ECTSA has a longer network lifetime than traditional protocols. The lifespan of the proposed ECTSA has been extended because of the increasing amount of data packets delivered to the destination point.

4.7. Performance of Packet Loss Rate

The suggested ECTSA and element model of the existing method was utilized to test the efficiency of the PLR. When compared to conventional techniques, the ECTSA yields better results. In comparison to the existing CEELBRP [24] and EECRP-PSO [25], the ECTSA has a lower PLR. The results demonstrates that the suggested ECTSA has a better PLR of 0.02% than conventional approaches, which have PLRs of 0.07% and 0.08%, correspondingly. Figure 9 shows the performance analysis of PLR. The performance of the proposed ECTSA, CEELBRP [24], and EECRP-PSO [25] is depicted in Figure 10.

4.8. Performance of Throughput

Figure 10 depicts the results of the throughput performance of conventional and proposed procedures. The proposed ECTSA provided better throughput results than the CEELBRP [24], EECRP-PSO [25], CLPDM-SI [26], and HFFGSO [28]. The access point obtains additional network packets, since the ECTSA has a lengthy network lifetime.

Table 9. Performance of Throughput.

Node Count	Throughput (Mbps)
	CEELBRP [24]	EECRP-PSO [25]	CLPDM-SI [26]	HFF-GSO [28]	Proposed ECTSA
50	0.79	0.75	0.74	0.87	0.91
75	0.77	0.72	0.77	0.85	0.88
100	0.73	0.67	0.80	0.84	0.87
125	0.70	0.64	0.81	0.82	0.85
150	0.68	0.62	0.76	-	0.77

shows that the suggested ECTSA obtained a high throughput of 0.91 Mbps, compared to the existing CEELBRP [24], EECRP-PSO [25], CLPDM-SI [26], and HFFGSO [28].

4.9. Discussion

In recent trends, there has been an increased interest in MANET research, and cross-layer design is one of the more intriguing applications. In this paper, the ECTSA, its multiple essential signals, and its qualities are presented. A novel dynamic routing strategy is offered that builds on cross-layer design. In order to reduce the energy consumption produced by the contentions, an Adaptive Competition Window (ACW) adjustment is also implemented. The overall analysis clearly shows that the proposed ECTSA provided better results than the existing CEELBRP [24], EECRP-PSO [25], CLPDM-SI [26], and HFFGSO [28] in terms of energy consumption, delay, lifetime, PDR, PLR, routing overhead, and throughput.

5. Conclusions

In this work, an energy-efficient cross-layer routing is accomplished by developing the ECTSA for the MANET. In the ECTSA, the energy, DSR, and mobility are calculated from the network layer. Cross-layer design enables complicated structures which are based on extra data obtained from each state. This ECTSA selects the optimal path by considering four distinct objective functions: residual energy, communication cost, DSR, and mobility. Furthermore, the ECTSA achieved superior performance and provided enhanced decisions over existing routing designs, since it comprises additional data. Additionally, the network contention calculated from the MAC layer and the competition window were adjusted according to the residual energy and contention to minimize the energy consumption of the nodes. From the performance analysis, it is known that the ECTSA achieves a better performance than the CEELBRP and EECRP-PSO. The energy consumption of the ECTSA is 7.1 joules and extends the network lifetime up to 1603 s for 50 nodes, which are fewer when compared to the CEELBRP and EECRP-PSO. In future, this research will be extended by using hybrid optimization techniques to improve the performances in the MANET.