ARPVP: Attack Resilient Position-Based VANET Protocol Using Ant Colony Optimization

The position-based routing of Vehicular Ad hoc Network (VANET) vulnerable to various security attacks because of dependency on computing, control, and communication technologies. The Internet of Things (IoT)-enabled VANET application leads to the challenges such as integrity, access control, availability, privacy protection, non-repudiation, and confidentiality. Several security solutions have been introduced for two decades in two categories as cryptography-based and trust-based. Due to the high computation complexity, cryptography-based solutions are outperformed by recent intelligent trust-based mechanisms. The trust-based techniques are lightweight and effective against the well-known security threats in VANET. The objective of this paper has to design a novel position-based routing in which the conduct of vehicles assessed to accomplish reliable VANET communications. Attack Resilient Position-based VANET Protocol (ARPVP) proposed to detect and prevent malicious vehicles in the network using the trust evaluation technique and artificial intelligence (AI). In the first phase of ARPVP, the periodic self-trust assessment algorithm has designed using various trust parameters to detect unreliable vehicles in the network. In the second phase of ARPVP, the position-based route formation algorithm has designed using the AI technique Ant Colony Optimization (ACO). ACO solves the problem of reliable route formation by neglecting the attacker's using a trust-based fitness function. The trust parameters of each vehicle as mobility, buffer occupancy, and link quality parameters had measured in both phases of ARPVP. Simulation outcomes of the proposed model outperformed state-of-art protocols in terms of average throughput, communication delay, overhead, and Packet Delivery Ratio (PDR).


Introduction
Recently automotive manufacturing has been proven as one of the fastest rising industries due to rapid advancement in technology, especially the wireless communications among the vehicles enabled by the Internet of Things (IoT). The IoT-enabled VANET is a collection of vehicles and infrastructure devices like Road Side Units (RSUs) connected through the Internet [1]. The IoT has more than the most recent pair of a long time, become an issue in the insightful world, and transversely over different ventures. While getting logically unavoidable, IoT supports a thorough depiction of the state of being and not all that terrible degree of relationship with the actual world [2]. Zones, for instance, collaborations, Intelligent Transportation Systems (ITS), business/measure the heads, and e-prosperity are several instances of conceivable application fields where this novel point of view will be significantly important. The certification of IoT tremendously depends on various rules, for instance, the routing's design, networks, and interchanges, data dealing with, and inescapable figuring developments that help profitable, strong, and physical and computerized interconnectivity. A significant fundamental driving force of IoT that empowers the interconnection of plans is arranging, and unequivocally, guiding the framework. It incorporates the creation of traffic courses and communicating the directed paths from the source to the last destination in a framework. With billions of things interconnected in the framework, an extreme test confirming that the framework from various sorts of perils and attacks. Customers will feel unreliable about their data if they are powerless against attacks from unapproved individuals or machines over the framework. Accordingly, security is by an expansive edge may be the best test in IoT networks [3]. A huge part of the security threats is performing at the routing layer, which suggests all through the data transmission method, in this way the keeping an eye on strong insurance from such perils will depend upon the security framework organized in routing helpfulness. Nonetheless, the cross-layer attacks additionally expanding these days to upset the correspondence networks to a huge broadening. Accordingly, making the strategy of secure interchanges in IoT is significantly also testing. This central prerequisite for confirming the routing method flanked by an assortment of IoT gadgets over various heterogeneous networks needs exploration responsibilities [4,5]. To reduce the troubles of secure routing in IoT engaged networks, different courses of action arranged at the routing layer from the latest decade. The security plans expected to recognize and ease the online protection risk, for instance, botnets, Denial-of-Service (DoS), malware, Distributed Denial-of-Service (DDoS), Man-In-Middle (MIM) attacks, sticking attacks, and so on. All such attacks are performed inside the network with the wrong intentions. To mitigate such attacks, not only strong measures are required but also minimum computation efforts for guaranteed network QoS. The cryptography-based methods introduced [6][7][8][9][10][11][12][13] for secure end-to-end data transmission, but these methods yet to address the challenges of key generation, encryption, and decryption processes with minimum computation overhead and higher security against the various threats. To overcome the problems related to cryptography-based techniques, recently the trust-based mechanisms gained significant attention.
Apart from the challenges of VANETs like highly dynamic network topology, load balancing, and fault tolerance, the nodes reliability analysis is also a vital factor. The unreliable vehicles need to be mitigated and select the reliable nodes for the transmission of data. In VANET, vehicles can communicate mistaken data both inadvertently, for example in case of a sensor disappointment, and purposefully when an attacker rolls out unapproved improvements to the product and equipment parts of the vehicle. These dangers are called figment attacks [14]. The counterfeit data broadcasting suggests that different participants change their course, mobility, and plan further activities based on the data received. It can be utilized by an attacker both to falsely diminish the gridlocks on his course and purposely produce an auto collision. Conventional routing strategies and data trustworthiness affirmation in remote networks are not compelling against such attacks just as having extreme calculation trouble [15]. As an approach to counter attacks focused on relevant data uprightness, the components based on notoriety and trust can be utilized when each network part has its degree of notoriety in the framework, and based on this level, the remainder of the participants conclude whether to trust the data got from this part. Nowadays the IoT applications [16][17][18][19] need reliability methods to protect from various security threats.
As the trust and reputation-based methods effective with minimum computation efforts, we proposed a novel secure position-based VANET routing protocol. The Attack Resilient Position-based VANET Protocol (ARPVP) is proposed to identify the attackers through weight-based trust calculations followed by AI-based stable and reliable route discovery. The novelty of the ARPVP protocol lies in the optimized position-based routing convention that evaluates vehicles to accomplish reliable VANET correspondences. The reliability achieved using the self-trust evaluation model. In the primary stage, the unreliable nodes recognition is detected, and in the subsequent stage, the reliable paths discovered using ACO for data transmission. Section 2 presents a survey of recent trust-based strategies for VANET. Section 3 presents the system of the proposed convention. Section 4 presents the reproduction results. Section 5 presents the end and future headings.

Related Works
Several methods proposed so far for reliability analysis and security in VANET communication using a trust-based approach. Some recent works are reviewed during this section. In a unique VANET environment, there is a lot of vulnerability in concluding who to trust. Existing trust models in VANETs incorporate substance-arranged trust models and data-situated trust models. Element-arranged trust models center around the modeling of the trustworthiness of companions. Data-arranged trust models put more accentuation on assessing the trustworthiness of data. A review of the recent state-of-the-art methods is presented in two parts such as trust-based security solutions and trust with Artificial Intelligence (AI)-based security solutions.

Trust-Based Security Solutions
The author analyzed two kinds of conventions for secure routing in [20] for VANETs such as geography-based and position-based routings. Geography-based methods are customary ones for Mobile Ad-hoc Networks (MANETs); they utilize a source for objective data that was put away in the routing table. The routing table in proactive conventions refreshes habitually on powerful geography, and they pick the most limited way calculation for routing. Receptive routing was utilized for large networks as they propose high portability and dynamic nature. At the point when the source speaks with the objective hub, the course was found continuously. Afterward, the sink node sends an acknowledgment message to the source. The half-breed method (Zone Routing) represents a mix of the initial two sorts and it is utilized by the network situation. The similitude-based trust and notoriety structure for VANETs was proposed in [21]. The message must be checked in the wake of getting it.
A similitude mining calculation was utilized to process the likeness between non-direct comparative data. Suggestions of operators and direct insight into the outcome were coordinated as notoriety assessments. When the message content was checked, the estimations of trust and notoriety are refreshed. Another calculation utilized for trust the executives proposed and named BARS (Blockchain-based Anonymous reputation System) in [22]. Vehicles utilize two blockchain instruments for verification based on evidence of essence and nonattendance. Public keys were utilized as pen names to secure vehicle protection. The communicated messages were recorded in one blockchain to assess the standing of vehicles. The outcomes have shown that BARS adequately improved the trustworthiness of communicated messages and ensures vehicle security effectively. The LSOT (Lightweight Self-Sorted out Trust) system was proposed in [23] as a calculation to secure VANETs interchanges. In the model, the nodes are self-composed; the total trust declaration-based and proposal-based execution. The ART (Attack-Resistant Trust) model proposed in [24] assesses the trustworthiness of both traffic data and vehicle nodes for VANETs as two separate measurements, specific data trust, and hub trust. Data trust was utilized to check data, yet the hub trust shows how trustworthy the nodes in VANETs are. To assess the proficiency of the proposed model, tests were conducted. The outcomes demonstrate that the Craftsmanship model adapts to pernicious attacks. The author utilized three markers for trust and proposed the trust assessment model based on boundaries of notoriety, experience, and information in [25]. Notoriety shows how well the trustee has traded data with the entirety of the substances up to this point. Experience demonstrates how well the trustor has achieved trading data with the trustee recently. Information renders perception on the trustee (the vehicle which is giving data) as immediate trust. The detailed proofhypothetical of the trust and notoriety model for VANETs with an augmentation of the normal derivation analytics (un)secured was introduced in [26,27]. Utilizing a calculation, they could qualify the activity passed as a protected message through quite a few vehicles by checking at every cooperation that consistency is saved. Hence, the standing model is based on an assessment of defined input messages, as far as the fleeting measure and positioning of the relevant help normal for each message. The modified AOMDV convention with an RLLMR (Reliable Low-Latency Multipath Routing) technique proposed in [28] based on multipath interface unwavering quality, equipped for deciding the reliable courses pre-emptively for VANET interchanges. The blockchain-based trust the board and data sharing answer for VANETs was designed in [29] called DrivMan. The utilization of DrivMan gives every vehicle a one-of-a-kind crypto-unique mark that had been utilized to build up data provenance. Declarations given by foundation units are abused to save the protection of the vehicles. DrivMan can be utilized as a compelling answer to forgive both data provenance and data uprightness to shrewd vehicles in VANETs for their safe and reliable activity.

Trust-with AI-Based Security Solutions
The novel trust-based multicast routing convention proposed was called MTAODV (Multicast Trust-based AODV) [30] to secure against numerous attacks and improve routing effectiveness. They processed direct trust utilizing the Bayesian hypothesis and roundabout trust utilizing assessment validity and action. The fuzzy rationale hypothesis was applied to fuzzify the immediate and roundabout trust esteems, and afterward, the absolute trust estimation of the hub is gotten by defuzzification. Another recent trust-based methodology for secure data forwarding in VANETs was proposed in [31] using AI and trust evaluation. They used the fuzzy logic technique and neural network for the trust evaluation algorithm. The authors of [32] suggested a privacy-preserving cloud-fog-based traceable road surface condition monitoring approach based on trust and AI technologies. In [33], the authors have proposed a reliable VANET communication mechanism using a novel Java macaque algorithm with a trust-based fitness function. In [34], another hybrid routing solution had proposed for secure VANET communications. The authors suggested an Enhanced Hybrid ACO Routing Protocol (EHACORP) in [35] to increase the efficiency of the routing process by utilizing the shortest path. The Raccoon Optimization Algorithm-based Accurate Positioning Scheme (ROA-APS) had presented in [36] for enhancing the accuracy of range measurement estimates to establish the actual position of nodes. Apart from this, we have reviewed the recently presented various security solutions [37][38][39][40][41][42][43] using trust-based and AI-based approaches for wireless communications.

Research Motivation and Contributions
The above studies show that reliable communications can be using trust and reputation-based techniques effectively. However, considering the challenges of positionbased routing related to the intersection management system, it becomes difficult to design trust-based threat detection in VANETs. The communication overhead incurred due to inefficient intersection selection mechanism along with unreliable forwarding mechanism in present solutions. It is mainly due to the high mobility of vehicles that leads to excessive data loss and increases communication costs. Therefore, the research gaps noticed in position-based and trust-based methods are summarized below: • Inaccurate selection of intersections leads the network performance degradations regardless of optimal path availability. • Inefficient intersection selection approach due to consideration of just immediate intersection while next intersection selection. • The communication overhead was incurred due to an inefficient intersection selection mechanism. • Unreliable forwarding mechanism due to the high mobility of vehicles and congestions caused by unreliable nodes leads to excessive data loss and increases communication cost.
Considering the above research gaps, we focused on intersection management free secure position-based routing in this paper in novel ARPVP protocol. The novelties of ARPVP protocol are summarized in below contributions: • Self-trust computation model to detect the unreliable vehicles in the network before link disNotized mechanism for RSU and intersection points. • Reliable and Stable position-based routing using the nature-inspired algorithm ACO regardless of intersection management and RSU devices monitoring. • Extensive performance evaluation of proposed protocol with state-of-art trustbased VANET protocols by considering the different network scenarios and traffic patterns.

Methodology of ARPVP
The proposed protocol ARPVP has based on a trust-based approach for continuous reliability monitoring and reliable data transmission. In both cases, a trust-score is computed for each vehicle to achieve security against the various threats in VANET. As discussed earlier, the ARPVP has designed in two phases such as (1) attack detection phase and (2) positionbased data transmission phase. This section presents the design of both phases. Table 1 presents the mathematical notations and their significance. The speed parameter has been used to discover the stable vehicles in the network for the data relaying process. As the vehicle with high mobility is considered an unreliable node, the current speed of the vehicle is an important trust parameter for attack detection. Furthermore, due to the malicious behavior of the attacker, the Medium Access Control (MAC) layer parameter called link quality is greatly affected and leads to a poor MAC layer trust score. This trust score helps to discover the cross-layer attackers easily, therefore the aforementioned trust parameter is computed for attack detection. The final parameter is related to network load or congestion on the vehicle. Attacker nodes are frequently broadcasting the false message that creates heavy bandwidth utilization around it and leads to packets drop. Thus estimate the current bandwidth level of the vehicle helps to discover the behavior. We used all three trust parameters to compute the trust score of each vehicle and then compared it with the predefined threshold value.

Trust-Based Attack Detection
If the trust score of vehicle N at the current time t is less than the predefined threshold value, then it is detected as an attacker node, and then necessary prevention measures are taken. Otherwise, node N is declared as legitimate and can be available for data forwarding processing. The entire process of attack detection mathematically presented in Algorithm 1. As noticed in Algorithm 1, every vehicle analyzed according to their trust value T computed using three self-trust parameters. These parameters computed as below: Mobility Trust: This trust parameter returns the current moving speed of the vehicle N at time t using below Eq. (1).
The max speed considered for each vehicle is 120 km/hr. This delivered the T1 t N that in range of 0 to 1. Higher the T1 t N value, better the chance to become legitimate node. Traffic Trust: This parameter estimates the trust score of N according to its bandwidth allocation to define its behaviour in network. The computation of this parameter is concerns the two important factors such as current bandwidth requirement and allocated bandwidth. We allocated the maximum and minimum trust values to define the current level of traffic around node N.
where BW A and BW D are current available and demand of bandwidth around vehicle N at current time t . The BW max represents the maximum bandwidth value allocated to each vehicle in-network is 2 Mbps.
Link Quality Trust: This parameter computes the MAC layer link quality trust of node N between time intervals. It is computed as: where N recv and N exp total number of packets received and expected number of packets during the time interval (t i−1 , t i ) at node N . Higher T3 t N , good chances to become legitimate node for vehicle N.
Common Trust Score: Once the three trust parameters computed for vehicle N at time t, then the common trust score estimated using the weights. This periodic trust value computed for each vehicle and compared with the predefined threshold value as shown in algorithm 1 and Fig. 1. It is computed as: where w 1 , w 2 , &w 3 are weight parameters applied to normalize the trust score in range 0 to 1. The selection of these values should satisfy the condition of w 1 + w 2 + w 3 = 1 . These weight values used as w 1 = 0.35, w 2 = 0.35, &w 3 = 0.3 . T t N value is compared with threshold which is set to 0.4 for this protocol. This value is set through the performance analysis of proposed protocol.

Position-Based Data Transmission
The process of attack detection is periodically updating the list of legitimate and attacker nodes in the network, which is helpful during the cycle of course disclosure and data transmission to forestall data misfortune. The course revelation fundamentally set off by any source hub S ∈ N that wants to send the data towards the objective hub D ∈ N . Disclosure of the steady and reliable course between the source and objective is a difficult examination issue for VANETs. The ACO used to find the ideal course by figuring the trust of each forwarder hub.
The problem of discovering the stable route is formulated by applying the ACO to search for the best relay node using trust analysis of each vehicle. The idea of the ACO remains the modeling of the ant behavior related to their ability to find the best available route from the anthill to the food source as demonstrated in Fig. 2(1) where the shortest route is preferred. The process of ACO is iterative in which initially the ant broadcasted the way with pheromone, and this information utilized by other ants to select the route ( Fig. 2(2)). The final route traveled by the ant visible when all the possible options are evaluated ( Fig. 2(3)).
The main steps of ACO to select the best route are: • Initialize the ants at the source from which the data transmission will perform. • Construction of the acceptable alternative forwarders.
• The rule determining the probability of the ant transition from the current node to the next node. • The rule of Pheromone update once after the selection of the best next-hop node in the network. • Pheromone evaporation rule.
Using ACO, S triggers the reactive route formation process by finding its neighbors and broadcasting set ants At to all its neighbors. The set of ants At generated based on the number of neighbors discovered at the current node, each ant works to find the best solution among all available neighbors based on their trust score value evaluation. Similar to attack detection, the trust score is registered to utilize three key boundaries of every vehicle, for example, speed, geographical distance from the current vehicle to the next forwarder node, and traffic estimation level of a vehicle. This process has presented in Algorithm 2. As showing in the algorithm, the next-hop node selected based on their current probability value computed using the parameters such as speed, geographical distance, and traffic conditions. The best part of this algorithm is that it checks periodically for the node's behavior against the attacker's list. The node is analyzed for the selection of forwarder only if it's not belonging to the periodically updated attackers list. The trust score T t Nb i for ith neighbouring node Nb i at time t is computed as: where T1 t Nb i and T2 t Nb i computes the speed and bandwidth level trust scores as discussed in above section. The weight parameters are similar as used in attack detection algorithm. Apart from this as this position-based routing we computed the geographical distance among the two vehicles as: The distance i based trust score for a node N i is computed as:

Fig. 2 Illustration of ACO working
where X and Y represents the height and width of the VANET network, the outcome value in T4 t Nb i in the range of 0 to 1. Higher the distance i value of node N i better the chance to become the next forwarder node. Figures 3,4 and 5 illustrate the secure and dynamic data forwarding mechanism of the proposed protocol. Figure 3 shows the VANET with source vehicle and destination vehicle or base station node in the network. All other vehicles are moving across the urban area. When the source node triggers the route discovery phase, the trust evaluation of neighboring nodes is performed according to Algorithm 2. As observed in Fig. 4, nodes 8 Node 5 is selected due to the higher trust value for vehicular data forwarding toward the intended recipient. Also, nodes 8 and 12 were part of the attacker list. Thus, the 5 becomes the next current node, and the same process is repeated to select the next forward node until the destination node has reached. Finally, Fig. 5 shows the final estimated path to transmit data from the source vehicle to the destination using the proposed approach. Figures 3, 4, and 5 demonstrated the behavior of the proposed mechanism effectively. In Fig. 3, the vehicles are deployed in the VANET scenario with highlighted source and destination vehicles in red color. Figure 4 demonstrates that the source vehicle started the discovery of the forwarding relays according to the proposed Algorithm 2. The integrated fitness value of each neighbouring node is computed which is 0.23, 0.33, and 0.41. As per the technique of selecting forwarding relay, the node with higher fitness value will be selected as the next forwarding relay. Therefore, vehicle number 5 is selected as forwarding relay. This process is repeated until the destination node reaches. Once the destination node is discovered by the proposed route discovery algorithm, the route is established between the source and destination vehicle as shown in Fig. 5.

Simulation Results
The ARPVP protocol is implemented and analyzed in the NS2 tool. The presentation of ARPVP protocol compared with recent trust-based routing protocols of VANET such as ART [24] and MTAODV [30] as the methodology of both protocols overlapping the concept of ARPVP and position-based routing. Also, the reason behind selecting these protocols is that they were evaluated regardless of VANET topology and RSU points. The performances of these protocols are compared using the parameters like average delay, average throughput, packet delivery ratio (PDR), and communication overhead (determines time & space complexity). The network scenarios used for the evaluations are varying numbers of vehicles (Table 2) and varying mobility speed (Table 3). For both scenarios, we used different mobility models namely random walk and Manhattan grid models with CBR and VBR traffic patterns. According to the simulation parameters section, A presents the results for density variations, and section B presents results for mobility variations. The simulation parameters show the presence of 10% malicious vehicles in the network.   Figures 6, 7, 8 and 9 demonstrates the results of throughput, PDR, delay, and communication overhead with varying numbers of vehicles using both CBR and VBR data traffics. From Figs. 6 and 7, we saw that as the network thickness builds, the exhibition of throughput and PDR diminishes because of a higher number of hops to send data from source to sink. The increased number of vehicles leads to more time to discover the forwarders and transmit data, and hence the throughput performance gets degraded. Another point that has been noticed here that the throughput ratio between CBR and VBR traffics. For VBR the traffic very much higher due to its high data rate capability of transmitting the multimedia traffic as compared to low data rate-based CBR traffic. Apart from this, the throughput performance in Fig. 6 shows a significant improvement in proposed protocol performance as compared to ART and MTAODV protocols under the attendance of 10% malicious nodes. It is since of the approach of discovering the reliable nodes and establishing the stable route for data transmission with periodic   Fig. 7 demonstrates the PDR performance that is overlapping the throughput performance. The existing protocols ART and MTAODV only focused on trust-based attack detection, however, it is possible that during the route discovery nodes may change their behaviors, and hence leads to unreliable communications in a network. In the proposed protocol, we applied the trust-based route establishment using ACO which resulted in the reduced number of packets dropped. Therefore, the proposed protocol achieved improved PDR and throughput using both CBR and VBR traffic patterns.  These results show that as the density an increasing, the delay and communication performance becomes worst. This is a possibility due to the increased number of re-transmissions caused by a higher number of vehicles in the network. The proposed protocol able to reduce the delay and communication overhead compared to both existing methods. The proposed protocol has taken a fewer number of routing packets to perform the communications because of minimum route formation operations. It leads to less communication overhead achieved compared to ART and MTAODV protocols.

Results of Mobility Variations
Mobility of vehicles is the vital problem of VANETs because it leads to unreliability in the network due to frequent operations of route discovery, route formation, and data transmission. The performance of VANET is significantly affected by the mobility of vehicles. Therefore, in this paper, we investigated the mobility variations and measure their effects on ART, MTAODV, and proposed protocols. Figures 10, 11, 12 and 13 demonstrate the outcomes for throughput, PDR, delay, and communication overhead respectively. From these results, the clear effect of higher mobility speed has been analyzed on all performance metrics. The increased mobility speed leads to decreased network throughput, PDR, and increased communication delay and overhead. It is mainly outstanding to the detail that severe mobility speed leads to unstable routes and hence the frequent route formation operations performed in the network.
The results of throughput (Fig. 10) and PDR (Fig. 11) demonstrate that the proposed protocol achieved improved tolerance against the increasing mobility speed as we consider mobility as a key trust parameter in both attack detection and route formation process. It leads to reduced route formation tasks and re-transmissions in the network. However, the existing protocols evaluate the nodes for attack detection only, and the route formation process is non-optimal.
The communication delay analysis by considering the mobility variations showing in Fig. 12 using both the CBR (A) and VBR (B) traffic patterns. The communication delay, as well as communication overhead (Fig. 13), is higher in the case of VBR traffic patterns as compared to CBR patterns. The proposed convention accomplished diminished correspondence delay when contrasted with existing conventions because of clear reasons uncovered in the above conversations. Along with trust-based attack detection, the reliability and robustness of the proposed protocol achieved using the optimal selection of  Furthermore, the communication overhead performance showing in Fig. 13 for both CBR (A) and VBR (B) traffic patterns disclosing that the proposed reliable and secure protocol able to keep the minimum computational overhead compared to existing protocols. As the consolidated approach used in the proposed protocol for attack detection and optimization-based method where the focus is on the discovery of more stable routes in the network which assures the guaranteed data delivery with the minimum overhead of paths formation compared to both ART and MTAODV protocols.

Time and Space Complexity
At last, the time and space complexity of all the protocols has analyzed in this section. The time complexity has analyzed using the communication delay parameter and space complexity has analyzed the average space required to perform the routing operations. The simulations were performed on Ubunto OS with RAM 4 GB and processor I3. The outcomes of time and space complexity are showing in Table 4. From these results, the proposed  protocol achieved a reduction in overall computation complexity compared to other protocols. It is mainly due to the reduction of operations related to route formation and frequent re-transmissions in the proposed protocol. The space requirements are computed as an average of each simulation scenario considering all TDMA time slots that show the number of routing packets generated. As the routing operations minimized in the proposed protocol, it leads to reduced space requirements as well. We calculated the time and space complexity of the ART, MTADOV, and ARPVP protocols based on the PDR and delay results. According to the PDR metrics, m-number of retransmissions are required utilising any protocol for the n number of nodes. Each protocol's space complexity (t1) is calculated as: where t1 proto represents the space complexity for each prootocol proto ∈ ART, MTAODV, andARPVP . As the PDR of the ARPVP is higher other protocols, it takes less space.
We assessed the average delay (ad) required to transfer data from source to destination using each protocol for time complexity. The temporal complexity (t2) for a network of n nodes with the ad is calculated as The delay performance mentioned Table 4 for the proposed protocol is lower, therefore, the time complexity performance of the proposed protocol mimimum compared to existing protocols. The limitation of this protocol is that it is not yet applied for the secure multimedia data transmission [46].

Conclusion and Future Work
This paper planned the novel position-based VANET routing protocol with objectives of reliability and robustness against the various security threats. As we know that VANET is vulnerable to several challenges that affect the reliability of the network, designing a reliable routing protocol to protect against malicious behaviors of vehicles is the main motive of this paper. The protocol was proposed with a two-step approach such as trust-based attack detection and trust-based route formation. In attack detection, three parameters have been used (speed, link quality, and congestion level) to discover the malicious behaviors of vehicles. In data transmission, the route discovery has performed by trust evaluation of vehicles using three parameters (speed, congestion level, and geographical distance). The consolidated approach leads to optimal solutions for reliable VANET communications. The simulation outcomes demonstrate the ARPVP had improved the performances regarding QoS, time, and space complexities. The throughput performance of ARPVP has increased by 13% and PDR by 12.5%. The communication delay and overhead have reduced by 11% and 14% respectively using ARPVP. For future work, first, we suggest researching various types of attacks, also we recommend applying the other artificial intelligence methods for reliable route formation and data transmissions. Data Availability Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Declarations
Conflict of Interest All authors declares that they has no conflict of interest.
Ethical approval This article does not contain any studies with human participants performed by any of the authors.