Optimal Relay Selection Strategy for Efficient and Reliable Cluster-Based Cooperative Multi-Hop Transmission in Vehicular Communication

Abstract Routing is the key issue for Vehicular Networks since each node in the network has mobility. The dynamic nature of vehicular networks increases with mobility and the consequence is reflected in controlling traffic overhead. As a result, the routing technique and the type of node mobility information are completely dependent on building dependable end-to-end communication links. To address these challenges, a Reliable Cluster-based Multi-Hop Cooperative Routing (RCCR) strategy was proposed by using velocity, distance, and link quality parameters. This algorithm obtains the tradeoff between Quality of Service (QoS) and mobility constraints over link parameters. It improves routing scalability by electing cluster heads and selecting Multi-Point Relays (MPRs) while considering mobility constraints and QoS needs. The proposed technique determines the link quality for every pair of nodes based on values of signal strength and distance parameters. The relay vehicles are chosen based on the highest possible QoS value, which is calculated to assure route stability, reliability, and durability. The heuristic limitations of the multi-point relay selection strategy are handled by considering the link quality, distance from the source vehicle, and cluster-head coverage area to enhance the performance of the multi-hop network with tolerable End-to-End transmission delay. Further, we obtained an optimal number of cooperative vehicles in every hop. Finally, simulation results demonstrate the effectiveness of the proposed algorithm compared to other state of art algorithms.


Introduction
Future intelligent road transportation systems (ITS) involve reliable communication between vehicles, roadside units (RSU), and infrastructure. Integrated with Inter of Things (IoT) in vehicular communication has steered to many user-friendly applications like parking, accident response and traffic congestion (Qureshi and Abdullah 2013) Based on the aforementioned uses, the increase of present system capacity and incremental expansion of data rates are vitally significant, and are being investigated by third Generation Partnership Project (3GPP) as part of the fifth-generation (5 G) standardization effort. Vehicle cooperative wireless (Campolo et al. 2017) is one of the main topics to investigate under 5 G, intending to reduce the effects of multipath fading and signal deprivation. Similarly, the integration of cooperative relaying in vehicular networks (CVN) improves overall performance.
The development and expansion of ITS for future-generation vehicular communications have brought a few critical challenges to them. First, due to the signal degradation parameters and the key challenging properties of their channels transmission power increases with propagation distance due to randomly time-varying nature, path-loss, signal fading, delay spread, and angular spread (Wang, Matolak, and Ai 2018;He et al. 2014). Second, during heavy traffic congestion and accidents latency plays a critical role in sending warning signals (Yang et al. 2004). Advance vehicular communication requires high data rate transfers with minimal latency to provide better service to the end users Furthermore, to enhance the coverage area, the collaboration mechanism between infrastructure and vehicles may be strengthened, resulting in greater road safety and network connection (Silva et al. 2017). Implementing full-duplex relaying to minimize end-to-end latency and also doubling spectral efficiency by eliminating the self-interference; is one of the primary answers to this demand (Kim, Lee, and Hong 2015). However, owing to self-interference at the intermediate vehicles caused by simultaneous transmission and reception over the same channel, cannot be completely prevented (Duarte and Sabharwal 2010;Jain et al. 2011). Meanwhile, because the intermediate vehicles send and receive in distinct time slots and over different frequency bands in half duplex relaying suffer a spectral efficiency loss. The self-interference mitigation in full-duplex cooperative vehicular communication is addressed in (Potula, Ijjada, and Santhamgari 2022).
Most established vehicular communication relay selection and routing methods are based on MANETs which offer the traffic information through device-to-device (D2D), vehicle-to-vehicle (V2V), and vehicle-to-infrastructure (V2I) wireless communications (Englund et al. 2015;Zhou et al. 2018;Devulapalli, Pokkunuri, and Babu 2021;Kumar and Saradhi 2021). The goal of Dedicated Short Range Communications (DSRC) (Kenney 2011) and multi-hop routing protocols (Eze et al. 2016) is to develop an effective transmission method for vehicles to exchange information with one another. However, due to the frequent changes in network topology, packet transmission is challenging. During the routing process, most MANET routing protocols are unable to ensure network topology. For network designers, the increased processing cost of vehicles is essential which caused owing to the control message size utilized for path establishment. The highly dynamic scenario may result in incorrect route selection, shortening the network lifetime and causing connection failures (Wahab, Otrok, and Mourad 2013;Balakrishnan and Karthigha 2017).
The clustering approach is the most appealing strategy presented to overcome the scalability problem which examines a tradeoff between mobility and quality of service constraints to improve network stability (Liu et al. 2018;Li, Boukhatem, and Wu 2016). Optimized Link State Routing (OLSR) (Clausen and Jacquet 2003) is a well-known proactive routing technique that employs a Multi-Point Relays (MPRs) (Rivoirard et al. 2017) strategy to reduce the number of relay nodes by reducing duplicate transmissions within the same zone. The primary idea behind OLSR is to elect a cluster head for each set of neighbor vehicular nodes by beaconing control messages. However, in a high-dynamic environment like Vehicular Ad hoc Network (VANETS), this protocol fails to account for node mobility limits, resulting in repeated disconnections, network overhead, and a considerable reduction in network lifespan (Idoudi, Abderrahim, and Mabrouk 2016;Darabkh et al. 2018). The added control overhead causes a network collision and hence the network resources are degenerative (Maccari and Cigno 2012;Benabbou et al. 2015). To address this issue, various researchers (Stephan and Joseph 2016;Al-Kharasani et al. 2018;Joshua and Varadarajan 2021) focused on QoS restrictions as a critical component to improve the capabilities of routing strategies to reduce the effect of the VANET's high dynamic situation. To meet vehicular communication application needs, network resource information should also be evaluated (Li, Yu, and Deng 2008;Amraoui, Habbani, and Hajami 2016). Incorporating the capability of clustering in the multi-point relay selection method to preserve the network connection and pick swift alternative paths in situations of link failures is a crucial challenge, rather than choosing the neighbor with a high link reachability degree.
OLSR protocol is developed especially for MANETs. It's a refinement of traditional protocols based on link quality to address the needs of wireless mobile users. The vehicles chosen by Multi-Point Relay (MPR) method generate link state information, which is a potential approach for reducing the control packet size. Aside from its easy operation, Optimized Link State Routing (OLSR) functionality responds well to constant changes in topology. It creates a reasonably good latency for ad hoc networks with a highly dynamic environment. This protocol can be readily implemented into Vehicular Ad hoc Network systems due to OLSR characteristics (Al-Kharasani et al. 2020;Toutouh, Garc ıa-Nieto, and Alba 2012). Conversely, the vehicle's mobility and road impediments devise a significant influence on the Optimized Link State Routing operation's efficiency, causing repeated link failures and a large control message overhead required for maintaining routes correctly. Due to their special form of neighbor location knowledge, the vehicular nodes are unable to swiftly calculate the next hops for data transfers. These constraints restrict the reliability of message delivery by lowering the information about mobility and route selection mechanisms. The goal of this project is to solve the route selection process to decrease needless broadcast overhead.
A novel Reliable Cluster-based Cooperative Routing Algorithm (RCCR) for the vehicular network is projected in this paper. To improve the energy consumption of the network, we optimize the number of cooperative vehicles. In this method, the most effective parameters are evaluated to obtain a tradeoff between mobility factor and reliable communication. The proposed algorithm improves the scalability of Optimized Link State Routing by considering capacity, link quality, distance, and mobility metrics. This algorithm determines the link quality between each pair of nodes based on distance and signal strength. Cluster head and intermediate vehicles are chosen based on the highest possible link quality, which is calculated to assure route stability, dependability, and durability. The heuristic constraints of the multi-point relay selection strategy are handled by concentrating on the link quality, distance from the source, and cluster-head coverage area to enhance the multi-hop PDR. Further, we optimize and attain an optimal number of cooperative vehicles in each hop.
The rest of the paper is organized as follows. The review of the associated literature work is presented in Section "Related Work." Section "System Model" provides the description of the system model for the vehicular network and our reliable cooperative routing algorithm is presented in Section "Reliable Cluster-Based Cooperative Routing Algorithm." The optimization of energy consumption is described in Section "Evaluation of the Optimal Number of Relays." The simulation results and analysis are presented in Section "Simulation Results" and finally, we concluded our work in Section "Conclusion."

Related Work
To deal with MANET networks, multi-hop relay selection procedures were frequently implemented. However, owing to the unique properties of a highly dynamic network, conventional methods of communication in MANETs cannot be straightforwardly applied to vehicular communication.
The key issues are network overhead, PDR, and end-to-end latency.
Clustering is one of the methods suggested for the operational exploitation of network resources for routing problems in vehicular communication. It is one of the strategies proposed for handling the issue of scalability and quality of service. Numerous relay selection approaches have been reviewed in Table 1.
The authors of (Clausen and Jacquet 2003) have proposed a Multi-Point Relay (MPR) for Optimized Link State Routing to improve routing scalability by lowering the overhead of control topology. The basic notion of MPR operations is to elect cluster-head (CH) which separates each set of neighbor vehicles into clusters, which is based on the premise of an exploratory selection process. Each node generates and maintains a collection of its neighbors based on connection reachability metrics in every hop set in response to incoming control messages. These CHs then choose a group of MPR relay nodes, which are specialized relay nodes. By reducing duplicate transmissions, this technique decreases the overhead of regulating communications within the same zone. When dealing with a high-mobility environment, this approach suffers from instability selection.
Authors in (Yamada et al. 2010) reduced the number of intermediate vehicles locally, only after all second-hop neighbor vehicles were covered to tackle the challenge of decreasing the number of innate clusters of Multi-Point Relay set. This technique's performance is only visible in networks with high density. It also results in resource waste owing to poor selection. Accordingly, authors (Li, Yu, and Deng 2008) devised a Necessity First Algorithm (NFA) for handling the relay selection problem, which enhances the Multi-Point Relay selection approach to a degree and introduces high performance. The calculation of the Multi-Point Relay set may take more time and greatly increase overhead. As a result, (Benabbou et al. 2015) presented the New Cooperative Algorithm (NCA) to decrease the overhead of control topology by lowering the number of Multi-Point Relay vehicles. This strategy has reduced the number of CHs in the local area by considering the degree of collaboration and connection reachability. To get the smallest set, it separates the nodes into master/slave roles. The Cooperative Communication, NFA, and NCA algorithms were built for MANETs and only provided mediocre performance in VANETs.
In (Badis and Al Agha 2005) to choose the best MPRs authors have been assigned weights to individual links. The average latency and bandwidth parameters were taken into consideration while selecting the best MPRs. With minimum control overhead, the performance of OLSR increases exponentially with QoS. This protocol, on the other hand, was created for MANET. In (Jain and Kashyap 2019), the authors improved routing decisions based on QoS restrictions by proposing Link Defined OLSR (OLSR-LD), which incorporates link quality while selecting MPR sets. Despite showing better performance than the standard one, this metric failed in minimizing link failures and packet transmissions. In (Dahmane and Lorenz 2016) authors have described a strategy for lowering network overhead. To improve the relay selection mechanism authors have considered the link quality, link stability, and vehicle mobility level, which improved routing scalability. The routes that have been identified take the advantage of most crucial information that is exchanged between nodes. The network performance was improved in terms of PDR When selecting relay vehicles, however, the QoS measure was ignored.
Gravitational Search Algorithm-Particle Swarm Optimization (GSA-PSO) was used to offer the capability of detecting signaling mechanisms in (Usha and Ramakrishnan 2019) to a specified set of nodes as appropriate member nodes. This methodology was used on the MPR-OLSR to decrease control topology overhead and make better use of available bandwidth. In terms of latency, packet losses, channel usage, PDR, and throughput, this method has improved routing performance. However, the impact of vehicle mobility was not taken into consideration in their research. For crossroads in VANETs, a Cluster Head Electing in Advance Mechanism (CHEAM) was devised in (Huo et al. 2016). To evaluate and maintain which vehicle is ideal for a CH, the cluster metric's capabilities were strengthened by considering the mobility and transmission power loss. The link quality was improved, resulting in a steady cluster with minimal overhead, particularly when the number of remote vehicles was reduced.
The authors of (Poursajadi and Madani 2021) introduced a Generalized Optimum Relay Selection (GORS) method for selecting the best relay while keeping the broadcast and cooperation phases secure. Then, using an incremental process, they offer an Adaptive Optimum Relay Selection (AORS) method that delivers and adaptively retains security. Due to their special form of neighbor location knowledge, the vehicles (nodes) are unable to swiftly calculate the following hops for data transfers. These constraints restrict the reliability of message delivery by lowering the route selection mechanisms and mobility information.
To preserve network stability during communication, the authors suggested Quality of Service Optimized Link State Routing (QoS -OLSR) in (Wahab, Otrok, and Mourad 2013). To avoid link failure, they evaluated QoS and mobility limitations. This approach keeps the network stable by lowering the transmission overhead and end-to-end delay. However, they did not take the complexity of routing required to maintain the other route into consideration. The authors of (Rivoirard, Wahl, and Sondi 2020) proposed the Chain -Branch -Leaf (CBL) clustering strategy for constructing a virtual backbone in a VANET. By restricting packet retransmission according to a preset approach, they were able to reduce the size of packet flooding. Over numerous scenarios, Simulation of Urban Mobility (SUMO) developed realistic traffic road layouts that assessed both multi-point relay and Chain -Branch -Leaf. The Chain -Branch -Leaf can operate based on location and velocity data without taking into account any possible conjunctions at the CH, which is associated with conventional members. The control burden related to the proactive method, especially in VANET circumstances, is the key disadvantage.
Authors in (Song et al. 2017) proposed an unsupervised-learning-based method for the selection of relay nodes to help broadcast. The base station needs to multicast the data to the relays and data packets would be distributed to the whole network through D2D communication between vehicles. Authors in (Dang et al. 2021) suggested an enabling method for multi-carrier relay selection by sensing fusion and supervised ML through cascade Artificial Neural Networks to enable quick deployment and effective exploitation of multi-carrier multi-relay selection in intelligent vehicular communication systems. The repetition-based POC-MAC protocol for multi-hop transmissions was developed in (Zhang, Hassanabadi, and Valaee 2014) for the cooperative POC-based forwarding (CPF) protocol for highway vehicular networks. To adhere to the POC-MAC, several collaborating relays create a virtual relay at each forwarding hop by scheduling their broadcasts to coincide with a single POC codeword. CPF takes use of geographical diversity while reducing the impact of concealed terminals. The CPF protocol lessens interference between multi-hop packets and the periodic broadcast of safety heartbeat packets by establishing separate POC-based schedules for each. In the aforementioned articles, the authors considered that the location of vehicles is almost static.
To improve the performance of the multi-hop cluster vehicular network, the best path from source vehicle to destination was obtained in (Alam, Adhicandra, and Jamalipour 2019) by using a close-optimal and optimal intermediate vehicle selection strategy. In this work authors considered instantaneous Signal to Noise Ratio (SNR) for the selection of optimal relay. The optimal path obtained with this approach may not be reliable since nodes in the vehicular network are not stationary.
Authors in (Al-Kharasani et al. 2020) have proposed a Cluster-based ADEPT Cooperative Algorithm (CACA) based on the quality of service. This method obtains the tradeoff between quality of service and mobility constraints by evaluating mobility factor and quality of link parameter, and it tries to improve routing scalability by choosing CHs and picking multipoint relays while keeping QoS requirements and mobility constraints in mind. The proposed technique determines the link quality for every pair of nodes based on values of signal strength and distance parameters. The relay vehicles are chosen based on the highest possible QoS value.

Dahmane and
Lorenz (2016) P-GPSR It considers link stability and node mobility for the relay selection mechanism. QoS was neglected.

Poursajadi and
Madani (2021) GORS and AORS In GORS, an optimal relay is selected based on privacy capacities. Further, they developed an incremental version of GORS called AORS with respect to the existence of the direct link and the availability of CSI. It was unable to swiftly calculate the next hops.

Alam, Adhicandra, and
Jamalipour (2019) Coop V2V It uses the instantaneous signal-to-noise ratio for optimal relay selection. The effect of mobility was neglected.
Numerous clustering algorithms were developed to address the OLSR protocol's scalability to reduce routing costs in a dynamic network. The goal of our proposal is to apply a clustering technique to choose the best multipoint relay in terms of link quality. Further, an optimization mechanism is incorporated in each hop to obtain the optimal number of cooperative vehicles.
The framework of the proposed approach is illustrated in Figure 1. The goal of our proposal is to minimize the cluster heads to further achieve minimum network overhead and maximum cluster heads to achieve the lowest network overhead and the maximum PDR possible. An optimization mechanism is implemented in each hop and an optimal number of cooperative vehicles.

System Model
The system model of the downlink cooperative vehicular network with NOMA is shown in Figure 2, where certain intermediary vehicular nodes are used in this context to enhance communications between source and destination vehicles by decoding the source messages and re-transmitting them to the destination vehicle. Every node, including the source ðv s Þ, K intermediate vehicles ðv 1 , v 2 , :::::::::v K Þ, and destination ðv d Þ, is equipped with a single antenna. In full-duplex mode, each node simultaneously transmits and receives data to prevent spectral efficiency loss. Further, we assumed that the channel is accurately known, as the estimation technique is out of the scope of this research. Furthermore, we assumed that the channel between each transmitter node p 2 v s , v k f g and receiver node q 2 v k , v d f g , for k ¼ 1, 2, ::::::::, K, i.e., h pq , is observed as the quasi-static flat fading with zero mean and variance r 2 pq : Figure 1. Solution framework.
As per the system model, NOMA is considered with a two-way DF relaying protocol. It is assumed that S, r, and D use similar transmit power ðPÞ: The signal received at Candidate Relay (CR) set nodes from the source and destination vehicle is expressed as: h v d , v k x 2 þ g v k k ¼ 1, 2, :::::::, K In the cooperative phase, the relay node broadcasts the signal x with transmission power P, the received signals at v d can be depicted as At the destination node, the Maximum Ratio Combining (MRC) strategy is used to combine the encoded information from various paths to obtain the information with minimal probability of error.

Reliable Cluster-Based Cooperative Routing Algorithm
In this section, Reliable Cluster based Cooperative Routing (RCCR) algorithm for the vehicular network is presented. This routing strategy improves the scalability of Optimized Link State Routing (OLSR) by considering capacity, link quality, distance, and mobility metrics.

Cluster Formation and Cluster Head Selection
The shortest path method of the OLSR routing protocols used in (Yamada et al. 2010;Li, Yu, and Deng 2008;Benabbou et al. 2015) are developed to decrease the number of multi-point relay along with the control messages size. These heuristics don't always deliver the best option since they ignore other routes with the same hop path length and reachability of the link. Those routes may be superior in terms of end-to-end latency, PDR, and network overhead in many circumstances. One of our goals is to prioritize the ideal path by selecting as many one-hop neighbors as possible.
To avoid repeated transmissions within the same zone, each source node emits a beacon signal and control messages regularly. A routing table is updated regularly to keep paths with a limited number of forwarding neighbor vehicles up to date. The quality of the path is analyzed when a vehicle receives a beacon message from its own one-hop neighbor's vehicles, by taking into account metrics like bandwidth, connection, speed, and distance. The bandwidth parameter is taken into consideration to provide dependability, the connection factor is considered to ensure a larger coverage region, and speed and distance are taken into account to ensure route stability. Let v s be a network source node and v k be a two-hop vehicle. The metric values are assigned to the link between ðv s ; v k Þ : dis v s , v k is the distance between v s and v k , and WF v s , v k is the cooperative weighting factor of both v s and v k : The capacity available between v s and v k is denoted by C v s , v k : The link quality for v s is LQðv s Þ, and the representation to the source vehicle neighbors is Nðv s Þ: WF v s , v k is proportional to the distance and inverse of mobility factor. The proportionality constant is the ratio between the CR of v k to the total CR. WF v s , v k can be given as shown in the equation below: The source node will compute WF v s , v k using periodic beacon signals and the distance between two vehicles as indicated by Equ. (5) provided by (Jarvis et al. 2018).
Where, k is the wavelength of the carrier. / is a complete phase obtained from signals which are communicated with fixed carrier frequency and B is an integer.
In the proposed work, low-speed moving vehicles are best suited as CR vehicles to rebroadcasting the information. Eq. (6) shows the mobility factor average value which depends on the speed of one's own vehicle ðvÞ: The computation of the following hop takes precedence in this equation.
where V r depicts the speed of the receiver vehicle. V min and V max are the minimum and maximum speed of the vehicle, respectively.
Algorithm Reliable Cluster-Based Cooperative Routing Input: A new flow request from source vehicle to destination Output: Multi hop Cooperative routing path from source vehicle to destination 1: While v d 6 2 V s ðvÞ do 2: Find V s ðsÞ 3: Source vehicle v s calculates the LQ v s of all the vehicles in V s ðsÞ 4: Forms the cluster based on LQ v s 5: selects the vehicle v k with a high LQ v s as CR 6: k th hop CR vehicle will act as the source vehicle for the ðk þ 1Þ th hop 7: end The product of capacity and weighting factor ðWF v s , v k Þ is used to determine the route quality. This is because, in the case of a high mobility factor, the MF v s , v k will be low and resulting in a smaller value of LQ v s , v k , as shown in Eq. (7). If the denominator value MF v s , v k is small, the WF v s , v k produced by Eq. (1) is large, resulting in a larger LQ v s , v k : In general, the new MPR selection algorithm prioritizes vehicle nodes ðv k Þ with a greater number of multi point relay linkages to become an multi point relay of v: As a result, LQ v s , v k selects the vehicle v k with the highest multi point relay linkages while keeping the number of multi point relay in v s low.
Our approach picks the source vehicle's CR set based on the LQ v, v k parameter; the algorithm selects the vehicles in v k with the greatest LQ v, v k without repetition. Other vehicles in the CR set help the source vehicle to forward the information toward MPR vehicle which are called as Candidate Relay vehicle. The flow diagram of the proposed algorithm is given in Figure 3.

Evaluation of the Optimal Number of Relays
In this section, we calculated the optimal number of cooperative nodes using cooperative multi input single output transmission model. After obtaining the route information between source and destination vehicles, in each hop data will be transmitted in two phases i.e., the broadcast phase and the cooperative phase.

Broadcast Phase:
In the first phase, data is disseminated to all n nodes in the cluster, where n can be obtained by Where PrðLQ v s Þ is the probability that the number of vehicles having Link quality LQ v s , v k greater than the threshold and A is the considered road area. For M-QAM modulation, the average energy use may be represented as (Devulapalli, Pokkunuri, and Maganti 2021): , b is the bit rate, BW is the Bandwidth, G Tx and G Rx are the transmitter and receiver gains respectively, M l is the link margin, carrier wavelength is denoted by k, N f Noise figure, l path loss exponent, P Tx , P Rx are the transmitter and receiver circuit power respectively, E avg, ph1 is the average received energy per bit during broadcast phase.
Cooperative Phase: In this phase, n nodes consist of n À 1 intermediate vehicles and one source vehicle re-transmit the data to the candidate relay. The average energy consumption in cooperative phase can be obtained by The upper bound of E avg, ph2 can be obtained by applying the Chernoff bound (11), expressed as: and E avg, ph1 can be obtained according to (11) by substituting n ¼ 1 The average energy consumption per bit for every hop ðE hop Þ can be obtained by summing the average energy consumption in two phases i.e., By approximating the bound (11) as equality, analytical expression for the average energy consumption per bit for a hop can be expressed as: From Eq. (8), maximum distance from source vehicle to the cluster vehicles is r 2 ¼ An pVPrðLQ vs Þ : The energy consumption per bit can be written as Therefore the analytical expression for energy consumption per bit for a hop is , Q e ¼ 4 bP e and Q p ¼ P Tx þP Rx bÁBW : According to the proposed algorithm, CH should be in the transmission coverage region of source vehicle. Hence the distance among the two CHs d max r, so we have average number of nodes n pd 2 max V A PrðLQ v s Þ: When d 2 max ! An pVPrðLQ vs Þ we can evaluate optimal n for the optimization problem given in Eq. (17), otherwise n ¼ 1: By performing the derivative for E hop with respect to n, E hop is a convex function with n when n is positive integer.
Since the parameters in above Eq. (18) are all positive, n should be minor than log ðQ e Þ: Let Q min ¼ minð log Q p , pd 2 max V A PrðLQ v s ÞÞ and n 0 be the real solution of (18). The approximate optimal number of intermediate vehicles can be obtained as

Simulation Results
This section compares and contrasts the experimental findings acquired with existing approaches to demonstrate the practicality of our proposed method. The parameters used in the simulation of the proposed algorithm are listed in Table 2. For the simulation, 60 vehicles are distributed randomly and moving at a steady speed of 15 m/s in various directions. The data packets of 512 bytes with a stable bit rate generated by a traffic generator are used to exchange information between vehicles.

Impact of Traffic Density
To investigate how traffic density impacts the network performance, we vary the number of vehicles to 100 in the network. The number of possible vehicles for cooperative relay vehicles grows as the number of nodes increases. As a result, the aggregate throughput increases for all routing systems, as illustrated in Figure 3. Among all the routing algorithms, the aggregate throughput increases with the number of vehicles under our proposed approach. With a welldesigned algorithm for path selection and relay selection, our approach can more effectively exploit the resources of cooperative vehicles to achieve high cooperative gain when the number of vehicles is large. At traffic density 100, compared with the CACA (Al-Kharasani et al. 2020), AORS (Poursajadi and Madani 2021), and Coop V2V (Alam, Adhicandra, and Jamalipour 2019) approaches, our approach improves the aggregate throughput of the network. The impact of node density on aggregate throughput i.e., Figure 4 is tabulated in Table 3. The performance estimation on the designed cooperative vehicular communication network in terms of outage probability is evaluated over various approaches as illustrated in Figure 5. The recommended optimal relay   selection RCCR algorithm shows the elevated performance by reaching the lower rate that demonstrates the superiority of the designed technique.

Impact of Communication Range
To analyze the influence of communication range on network performance, the aggregate throughput of the network is evaluated by varying the communication range from 100 m to 600 m by assuming other parameters as listed in Table 2. As shown in Figure 6, the aggregate network throughput decreases as the node's transmission range increases and remains the same until the transmission range reaches a large value. The reasons behind this can be depicted as; on the one hand, increasing the communication range and therefore the number of network links allows for additional options in terms of higher capacity routes with better cooperative vehicles and MPR (CH) vehicles. The boost in transmission range also aids network connection and the discovery of a better transmission channel. As a result, using too much transition power is counterproductive. Extending the  communication range for more relays comes at the expense of diminishing the interference range. When the transmission range is more than 200 m, all nodes are considered inside the transmission range and in the interference range of other nodes, since all 60 nodes are randomly positioned within the limited region of 1420 m Ã 1200m. As a result, when the transmission range is increased from 200 to 600, the performance of all routing systems remains the same. The impact of vehicle communication range on aggregate throughput i.e., Figure 6 is tabulated in Table 4.

Conclusion
An efficient Reliable Cluster based Cooperative Routing (RCCR) algorithm was presented in this paper. This algorithm has been introduced to resolve a tradeoff between mobility constraints and QoS requirements and to improve routing scalability based on the link quality metric which is used to select the cluster head and cooperative vehicles. For reliable communication, cluster head and cooperative vehicles are selected based on mobility factors and distance metrics. In each hop, optimization mechanisms are incorporated and obtained an optimal number of cooperative vehicles. The simulation results validated the efficiency of our suggested method, particularly in terms of aggregate throughput, and outage probability.

Author Contributions
The paper conceptualization, methodology, software, validation, formal analysis, investigation, resources, data curation, writing-original draft preparation, writing-review editing and visualization, have been done by Sravani Potula. The supervision and project administration have been done by Sreenivasa Rao Ijjada.