NAV2V: Navigation Assisted V2V Routing Protocol for Urban Areas

Vehicular AdHoc networks (VANETs) allow connectivity in dispersed vehicle environments and in other areas of diﬃcult access, where conventional network systems are not satisfactory. One of the most signiﬁcant technical challenges faced in implementing VANET is to design an eﬃcient routing protocol able to provide a reliable path between the source and destination of the information. This paper discusses a routing scheme that incorporates the control strategy for transmitting messages and Global Navigation Satellite System (GGNS) information to optimise network routing. This scheme uses geolocation information to select the best path to forward the messages. To simulate a communication in real-life scenario, we used the Simulation of Urban MObility (SUMO) and Network Simulator-version 3 (NS-3) platform to compare our proposed algorithm to the traditional routing protocols in scenarios where the number of nodes as well as the number of source-destination pairs vary. Our results show that the proposed NAV2V algorithm can decrease the packet loss rate, end-to-end delay, and enhance network eﬃciency.


Introduction
The fundamental requirement for the future of vehicular communication is that the network arrangement is to be done in an autonomous, intelligent and optimised way so that the cars forward the information among themselves without needing infrastructure. These communication methodologies combine to present a design of functions on the roads and improve the transport system [1]. The vehicles are mobile nodes, linked in a wireless set through AdHoc systems. Although these nodes are not displayed in a similar range, the information can pass through multiple intermediate nodes with the assist of routing protocols [2]. These moving cars are utilised as routers to provide a reliable mobile communication structure among the vehicles [3]. Interest in these networks is increasing, mainly in terms of research aimed at working with government entities, the automotive industry and transport sectors that ensure the security and performance of the transport system area [4].
The following open questions have inspired the research: • High Mobility: As the speed and route of vehicles within a given area are different, the problem of congestion arises, as well as topological changes. Data transmission on vehicle-to-vehicle (V2V) networks is difficult due to these irregular topological changes. The connection between nodes (vehicles) can take a few seconds, causing communication problems. Therefore, there is not enough time for the routing protocol to exchange information and build an efficient route table. • Connection instability: Starting and maintaining the connection is one of the necessary criteria for routing in modern VANET architecture. Because of the dynamic configuration, the connection lifetime is limited. An effective routing scheme would aim to provide the life of a connection for as long as possible. • Autonomy: Each node works autonomously, dealing with the information delivered across the network. As this type of network has no infrastructure, so it transmits information to all nodes, allowing it to withstand communication failures. It allows autonomous nodes to distribute information to their neighbours. Tracking vehicles' position can help to redirect data in an optimal way, which will reduce delay, network overhead, and use the limited time available to communicate with useful information. • Fault tolerance: As VANET is exposed to architectural disconnections, it is difficult to establish a stable link between nodes. An efficient routing algorithm should consider these factors and include a fault tolerance mechanism as part of the design. Therefore, an autonomous alternative routing protocol is required as a complement to the routing solution. • Scalability: The design of the vehicle network depends on the number of vehicles on the road. When the number of cars decreases, the need for assistance increases. However, as the number of vehicles decreases, the number of transmission demands is also expected to decrease. Vehicles require continuous data transfer, thus, in these situations the connection must always be preserved. The network must therefore be able to adaptively increase or decrease, based on requirements and other important factors. The networks are defined along the entire axis of the highway generating many nodes, therefore, the reception range needs to be expanded [4]. Mainly because we never know when it will be necessary to send information, especially if it is an emergency or a warning signal. • Low latency: As the transmission window is low and with ITS functions, the amount of requested transmission data can be very high, therefore, it requires a network with low latency. • Limited capacity: Bandwidth is known as the amount of data or information packets that are sent over an existing communication network. As Ad-Hoc VANET has no infrastructure, a weakening of the signal occurs due to electromagnetic interference in the radio signal and signal blocking, such as buildings and communication outside the line of sight [4,5,6]. The existence of dense buildings on the sides of the streets can degrade the performance of VANETs applications [2], resulting in a reduced capacity in relation to the guided networks. • QoS: The quality of service at VANET is extremely important due to the use of a wide variety of distributed mobile applications, including the delivery of traffic alerts, route planning and file sharing [7]. However, maintaining the priority option for delivering packages at an appropriate level represents a challenge, as it is difficult to guarantee the necessary resource reserve, due to the dynamic change of topology [3].
Despite all the problems presented, we hardly believe that there will be a single solution for routing protocol for vehicular networks. The development of efficient routing schemes is a challenging task and has a great impact on communication, leading researchers to experiment with different strategies to solve it.
In this article, we examine and analyse the performance of four routing algorithms for VANETs: AdHoc On-demand Distance Vector (AODV), Destination Sequenced Distance-Vector Routing (DSDV), the Optimized Link State Routing Protocol (OLSR) and Greedy Perimeter Stateless Routing (GPSR), as they are the most used and researched in AdHoc communication in vehicular networks today. Routing algorithms require optimisation for AdHoc vehicle networks as traffic congestion increases and a millisecond error in the traffic management network can lead to disastrous results [8]. Then, an optimised routing protocol scheme is proposed to reduce packet transfer times which can also be used for faster communication in a real environment, analysed in network simulator. This research sought to propose a solution that uses the position of vehicles at each moment of time as a parameter and dynamically updates the routing table to help each vehicle act as a network node and share information more quickly and efficiently. Therefore, a graph theory algorithm modified to help find the shortest path, based on the transmission time, is proposed to act as a catalyst when executed with existing algorithms.

Methods
Here, we present additional methods to improve the performance of our proposed algorithm, including: (1) Using information from the position of neighbouring nodes to make the selection of the best path and try to avoid nodes that do not connect to the final destination, to avoid loops of packages; (2) Selection of the best path, using a modified shorter distance algorithm, which takes time as the main weight and (3) We compare the performance of our proposed NAV2V with traditional protocols, analysing the performance metrics, such as packet loss rate, end-to-end delay and network performance.
Major contributions of this paper are summarised as follows: • We have introduced the most used routing protocols in vehicle AdHoc communications nowadays; • Then, we propose a new network routing scheme to improve data transmission using the information available on the vehicles' GNSS, which exploits the positions of neighbour nodes to forward the packets. We call this solution as NAV2V algorithm, which makes use of the concept presented in this work; • Subsequently, we present the comparative results between the current protocols and the new concept presented in this research using computer-based simulation software.
We organise the rest of this research paper in the following way: Section 2 reviews the related research work on network route protocol. Navigation Assisted V2V Routing Protocol for Urban Areas approach are described in Section 3. In Section 4, we explain the simulation environment. We explain the evaluation of the simulation results in section 5. Finally, Section 6 concludes the paper with discussions.

Related Work
VANETs are distributed and self-organised communication structures, configured through automobiles as nodes to transmit messages. These nodes are mobile, but have restricted ranges of mobility and communication, due to the direction of the roads, width of the lanes and limitations of pedestrian zones and buildings. The primary aim when researching VANETs is to create a fast and economical vehicle communication system to allow the distribution of data for the safety and comfort of passengers [9].
For the connection to be efficient, broadcast communication must be avoided, as it is seriously affected by propagation problems, such as signal loss, rapid fading due to multipath transmission and a hidden terminal problem [10]. To increase the efficiency of the V2V contact, a new smart contact relay scheme is expected as a research field. For the communication to take place, the vehicle communicates with another vehicle directly if there is a direct wireless connection available between them, forming a V2V communication. When there is no direct connection between the vehicles, it utilises an assigned routing protocol to route data from one vehicle to another until it reaches the destination point, forming the communication between the multi-hop vehicle [11]. To perform packet transfers, routing protocols use local information from their neighbourhoods on the network to decide which relay nodes will be used in routing the data. These protocols must be flexible to the peculiarities and resources of VANET, where vehicles can face many obstacles, such as crossings, buildings, traffic lights, trees, resulting in insufficient channel quality and connectivity. Therefore, it is necessary to use an efficient and reliable routing protocol to have an effective communication without losing data [12]. Various systems are involved through this modern notion of VANETs, for example, areas of traffic management, routing, handover, signal propagation, autonomous vehicles, etc. [13].
We can still apply the existing AdHoc routing protocols in other types of networks to VANETs. However, the result of the simulation indicated that they suffer from performance problems for rapid vehicle changes and reduced chances of transferring information [14].
Wireless VANET communication provides vehicle-to-everything (V2X) and the topology can vary from dense to very sparse [7]. With dense network typologies, vehicles on the transmission route make it possible to provide end-to-end multipoint connectivity between the origin and destination of the node, rather than transferring messages through base stations.
Communications on a sparse network can be achieved using the store-carry-andforward paradigm, inspired by Delay Tolerant Networks (DTN). It was originally designed for communication between spaces but is also recognised for use in scenarios where the telecommunications infrastructure is unstable or unavailable due to disconnected areas, natural disasters or emergency conditions. Based on DTN networks, Vehicle Delay Tolerance Networks (VDTN) were created. The framework consists of vehicle nodes and other nodes, providing a low-cost connectivity solution in challenging scenarios, where a telecommunications infrastructure is unreliable or unavailable [15,16].
This technology has a highly variable topology, with frequent partitions and, perhaps, low node density. Thus, delay-tolerant routing uses a message storage, loading and transferring paradigm, in which messages have a useful hop limit or time-to-live (TTL) and are stored until a satisfactory contact opportunity appears [17]. In this method, the node first stores the messages if there is no vehicle where it can be transported. Then, when other vehicles enter the transmitted radio line, the data will be relayed to the others [18]. VDTN routing algorithms make routing decisions by building and updating routing tables whenever mobility occurs [19].
Multiple replicas of messages can be generated to increase the likelihood of delivery at the cost of increasing network congestion. However, some nodes in a cooperative network may develop a selfish attitude to protect their resources, such as memory and energy, and not collaborate, that is, they do not forward messages. One solution to this dilemma is to adopt the Honesty Based Democratic Scheme (HBDS) introduced by [20] to find selfish nodes in a cooperation for relaying messages. They test performance metrics, such as package delivery probability, package delivery delay, overhead rate, and number of packages lost. The results indicate that the system can make nodes work together in a community to develop the performance of the network. The proposed scheme can be used in the future for cooperative distribution of information between nodes connected in the network to avoid problematic nodes [7].
Maintaining quality of service (QoS) on VANETs is crucial, while packet routing is a big challenge, especially when bandwidth is limited [21,3]. Many protocols aim to provide high-quality services (QoS) while saving more resources [22].
Another problem that exists in this type of approach is the hidden terminal: When there is no centralised communication coordination, the problem of the hidden terminal occurs in the VANETs as shown in Figure 1. This takes place when two nodes are not in the same communication range and transmit data for the same receiver [23,24]. This issue is a serious problem in the design of MAC protocols, because it deteriorates the performance of the MAC layer. Only a few researchers have discussed this problem for MAC protocols on VANETs. Therefore, this issue needs to be studied and integrated in the design of MAC protocols in VANET [25]. However, in our work, this problem is bypassed by using a synchronised solution, where now the hidden terminal is accessible via a bridge from the terminal that accesses both nodes.
Next, the main routing protocols available and currently studied on VANET networks, which are part of this study, will be described.

Routing Protocols
As routing is the central issue in networks where data transfer between network nodes is the fundamental requirement, this is an ongoing area of research and development. New routing protocol schemes continue to emerge as an improvement over established protocols [1].
Due to the high mobility of the nodes and the rapid changes in topology, designing an efficient routing protocol that can deliver a packet in a minimum period with few discarded packets is considered a critical challenge at VANET. In addition, many researchers have focused on designing a routing protocol suitable for dense environments that have a high density of vehicles with close distances between them. Designing an efficient routing protocol affects improving many factors; the first is to increase the reliability of the system, taking advantage of the percentage of package delivery, and the second, reducing the extent of interference caused by tall buildings in the city environment. The third factor is that taking scalability into account is essential to avoid conflict if a simultaneous unicast routing request operation has been initiated. Another issue is to deliver the package in the shortest time, especially in an emergency; this factor is considered being a very critical [11].
There are three different types of routing protocols for AdHoc data networks, proactive, reactive and hybrid protocols. The first is a proactive routing protocol that is based on the periodic transmission of the data network topology. Here, the protocol ensures that nodes always have an up-to-date knowledge of the paths to other nodes. The second is a reactive routing protocol that only looks for a route when one is needed. Finally, the hybrid routing protocol represents a mix between sensitive and practical protocols [13].
These protocols can also be classified into the following categories of approaches. The first category of routing protocols uses multiple on-demand paths to automatically switch broken paths to another routing protocol. The second category of routing protocols is position-based routing (or geographic routing). Most vehicles on the roads are equipped with GGNS and digital maps for navigation. These vehicles can therefore be easily aware of their geographic locations and then use this information to improve routing performance [26]. Now we will discuss the techniques and differences of the main routing protocols most used today.

AODV
The AdHoc Demand Distance Vector (AODV) protocol was developed in 1999 and, was based on Dynamic Source Routing (DSR) and DSDV protocols [4]. Currently, AODV is the most common routing protocol that provides route discovery and maintenance techniques, with reduced bandwidth and CPU usage being one of its most significant characteristics, as it sends packets only on demand [3]. It is, however, a protocol known as reactive and the most widely used of all currently active protocols [27]. When it needs to send data packets to a destination point and has no path to the destination node before sending the data packets, it initiates a route scan. This network requires three procedures: (1) the path discovery method, (2) the generation of path messages and (3) route management. The route is created only when necessary; in relation to constructive routing requirements, it needs less overhead. Consequently, the low overhead required for data packets is one of the main advantages of an AODV. In this protocol, the routing information, which is also efficient in terms of bandwidth, is not modified after a certain time. This strategy addresses the consequences of the initiated roads, along with the need for unused paths to preserve routes. In [28] the authors suggested that the use of path information for each node as a parameter to choose the next hop during a route exploration process, in order to increase the efficiency of an AODV on a VANET. Its analysis, however, does not provide end-to-end latency, transit rate and packet arrival rate and, it has not yet discussed the propagation effects to alleviate the flood problems [3].
AODV employs a route exploration method in the transmission mechanism. It also supports unicast, multicast and manages messages such as Route Request (RREQ), 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 Route Responses (RREP) and Route Errors (RERR) and uses the messages mentioned above to share information between the source and destination. "Hello" messages can be used to identify connections to and monitor neighbours where each active node periodically emits these messages to all neighbours. Since nodes periodically transmit these messages, a connection break would be detected if a node does not receive several "hello" messages from a neighbour. It transmits a Route Request (RREQ) to that destination when a source has data to be sent to an uncertain destination. A path to the origin is formed at each intermediate node when an RREQ is sent. If the transmitting node has not already received this RREQ, it is not the endpoint and has no existing path to the destination, the RREQ is retransmitted. If the receiving node is the destination or has a current path to the destination, a Path Response (RREP) is generated. The RREP unicast is hop-to-hop to the source. Every intermediate node establishes a path to the destination as the RREP propagates. It records the path to the destination when the source receives the RREP and can start transmitting the data. If the source receives several RREPs, the path that has the lowest hop count is chosen. Each node along the path updates the timers associated with routes to the source and destination, keeping the routes in the routing table. Data flows from source to destination. When a path has not been used for some time, a node cannot be sure whether the path is still valid; thus, the node excludes the route from its routing table. A path error (RERR) is sent to the data source in a hop-by-hop way if data is flowing and a link break is observed. Each intermediate node invalidates the routes to any inaccessible destination as the RERR propagates towards the origin. It invalidates the path and restarts the path discovery when the data source collects the RERR [29].
In AODV, networks do nothing except wait for links to be created, that is, nodes in the network that need to connect send a connection request. Nodes that are not part of a particular route must not hold precise route information. Consequently, these nodes are not capable of allowing the topology upgrade bundle to flow, or, in other words, they include only the path information that they are on [1].
Studies by [3,28]modified the AODV routing protocol adding details on the speed, orientation and location of the vehicle. For nearby cars, the routing tables now provide additional driving detail. To pick the next hop, they used the path parameter and then implemented the transport and routing system to bring the information to the vehicle nearest to the destination, as long as the source can send the parcels to other vehicles that are heading in the same direction, making the route more stable.
When there are more nodes, AODV is more efficient, as it is continuously finding new routes. Furthermore, AODV has less end-to-end delays as it has a higher number of collisions [4].

DSDV
Destination-Sequenced Distance Vector routing is a constructive routing protocol based on a table which deals with the route estimation Bellman-Ford algorithm. For the network nodes, DSDV establishes a routing table. The key contribution of the routing algorithm DSDV was to solve the routing loop problem. Since the routing table is managed by each node, if a router receives new information, the routing 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 table will need to be changed more regularly. With this information available in the routing table, through intermediate nodes [30], the source transmits the packet to the destination. The next hop to the collector and the number of hops away from the collector just need to be identified by each node. This information is contained in the [31] table.
If a path already exists prior to the arrival of traffic, it would be conveyed without delay. Otherwise, traffic packets must wait in a queue before the routing information referring to their destination is obtained by the node. The node automatically changes the sequence number and transmits the information to its neighbours if a path fails to the next node. It checks its routing table when a node receives routing information. If the routing table does not locate this entry, the routing table will be modified with the routing information found. If the node discovers that it already has an entry in its routing table, the sequence number of the obtained information is matched with the entry in the routing table and the information is modified. The information can be discarded with the lowest sequence number if it is smaller than the obtained information. The node would hold the knowledge that has the shortest path or the least number of hops to the destination if the two sequence numbers are the same. In DSDV, the routing table maintains a routing table for the destination node. The table will include all the other nodes that are known directly or by any neighbours known to you. The entry in the routing table for each node includes details about the IP address of the node's last known sequence number and the hop count to access that node. The table also holds a record of the next hop neighbour to enter the destination node along with this information, the timestamp of the last update obtained for that node [32].
In DSDV, the distribution pace of the package falls dramatically and one of the key reasons for this is the use of already closed routes in the event of broken connections. The existence of paralysed routes in DSDV does not actually mean that the path to the destination cannot be identified by [1]. The protocol provides a temporary connection to the desired destination through a neighbour that has a valid path [33]. However, as the number of vehicles is expanded relative to the other protocols, the outcome indicates a bad influence of the DSDV protocol. DSDV is also not a reasonable option if the number of vehicles rises considerably [31].

OLSR
The OLSR (Optimized Link State Routing Protocol) is a proactive routing protocol created for AdHoc mobile network types. To boost efficiency, OLSR uses a genetic algorithm, adjusting the parameters and variations in the experiments tested. This routing protocol depends on the use of efficient periodic flooding of control information, because it is a classic link state routing protocol and when using specific nodes that act as Multipoint Relays (MPRs).
Inside each node in the network topology, OLSR maintains a routing table to create a data transmission path. The basic principle used in this protocol is that messages are exchanged during the flooding process by the nodes chosen at the front. To discover its one-hop neighbours and even its two-hop neighbours through its replies, OLSR uses "hello" messages. This approach greatly decreases the overhead of the packet, contrasting a conventional flooding process in which all nodes retransmit each packet until the first copy of the message is received [9].  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 The features of this protocol are the delay times created by sending data packets are short and ideal, are well adaptable to topology changes, and are easily integrated with various types of systems.
This protocol has two key functions: to discover the neighbourhood for each node and to distribute the topology that transfers three different kinds of messages [13].
When using OLSR, in the event of faults and failures, there is a frequent exchange of messages about the topological details of the whole network. To minimise air traffic control, it uses the Multipoint Relay (MPR) technique. The OLSR operation regularly produces two messages on the network: • Hello message: These messages allow up to two hops for each node to know its neighbours. Each node makes use of this information to pick its multipoint relay nodes. • Topology control message: Each node transmits control messages known as network topology control messages to preserve the database necessary for routing packets. In order to build a set of selectors MPR [34], this message is transmitted periodically by separate nodes. As a consequence, for transmitting data packets, more bandwidth would be available. This protocol seems to be able to provide improvements in terms of packet transmission rate, latency, overhead routing and throughput [34], based on the simulation analysis.

GPSR
Owing to the prevalence and performance of satellite navigation systems, researchers have suggested various routing protocols tailored to vehicular networks in order to develop routing protocols in VANET. This has become a hot spot for testing, and the Greedy Perimeter Stateless Routing Protocol (GPSR), is the most promising routing protocol category using this method. It uses a traditional routing technique based on a location that uses information from nearby vehicles to determine which neighbouring node to receive the data would be chosen [12].
It uses greedy forwarding and forwarding by perimeter. GPSR usually forwards data packets using a greedy algorithm; where there is a local optimisation problem, it uses perimeter forwarding to send data packets, which ensures that the packet forwarding does not pick the next hop with a greedy algorithm. In comparison, GPSR belongs to the protocols of greed routing and does not include a table of maintenance routing. The protocol functions practically stateless and has routing multipaths capability [35].
Through the short-range position and location system, it is assumed that each node has its own position coordinate information available. Nodes regularly share this information through beacon messages with their neighbours in a single hop. Therefore, within the contact spectrum, each node has the position information of all its neighbours at any time, as well as the position of the destination by beacon messages and location services. Based on the response of the beacon messages, according to the greedy routing, the current node selects the best neighbour that is the nearest to the destination. However, after a timeout interval, if the current node does not obtain an answer from a neighbour, it finds the contact relation to be broken and deletes all entries from the neighbour table. Some conditions may   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 occur in which there is no better neighbour than the node itself, which is regarded as a maximum local state.
The GPSR can no longer keep the greedy forwarding technique in this situation, but instead transforms into a recovery mode to forward the packet to the next node. Both nodes obey the right-hand rule in the recovery mode technique to transmit the packet to the next node. Each node tests the packet header field after receiving the packets, either in Greedy mode or in Recovery mode.
In the perimeter routing mode of the GPSR algorithm, the right hand rule is able to ensure that the GPSR protocol can get out of the "hole", but makes the choice arbitrary to a certain extent, which often leads to finding greater leaps in the routing. An improved GPSR routing algorithm using the two-hand rule is proposed to solve the GPSR "hole" problem in wireless sensor networks. The double hand rule includes the left and right hand rule. The right hand rule will guarantee that the GPSR protocol can get out of the "hole" in the perimeter routing mode of the GPSR algorithm, but makes the decision random to some degree, always leading to larger routing leaps.
In [35], the authors proposed an update to the GPSR routing protocol, which preserved its functionality, such as low complexity and simple realisation. The results of the simulation in the NS-2 software show that the improved GPSR algorithm dramatically improved performance and demonstrated the efficiency of the proposed solution in the delay and packet arrival rates, compared to the original GPSR. A stable and improved variant of the Greedy Perimeter Stateless Routing (GPSR) protocol is introduced by the overlay. This protocol consists of two modules that incorporate an upgrade that minimises transmission delays and message control in the GPSR routing protocol [12].
Cross-layer data is often used as complement under practical limitations of the protocols in general. A practical implementation using this information was used in the work of [36] which identified barriers and other unforeseen causes of fading, such as interference, contributing to the phenomenon of radio irregularity, to update and improve the GPSR protocol. This update was called GPSR over Symmetrical Links (GPSR-SL). The experimental review shows the effectiveness of the GPSRSL compared to the original GPSR, where the results indicate that the proposed protocols allow a high packet delivery ratio without increasing energy consumption, maintaining a reasonable delay for the end-to-end application in relation to the original GPSR.
Next, we will cover other protocols, not so well known, but that make use of geolocation information and served as inspiring ideas to carry out this research work.

Vehicle Position and Route
There are many ways of relaying packets between various network nodes, as seen in previous protocols. In general, "hello" packages are often used between devices. Give these parcels to identifiable adjacent centres and become acquainted with their locations. The time to submit the course demand packages occurs as the initiating hub distinguishes its extensive hubs and their locations using GGNS.
Another technique that improves the rate of "hello" message generation in rapid motion was developed by [38]. They have an OLSR extension called Proactive-OLSR (P-OLSR) that uses the details available from GGNS to weight the parameter of the expected transmission count (ETX). The relative speed and direction of motion of the unmanned aerial vehicles (UAVs) [34] are also considered.
Solutions suggest the joint optimisation of essential device parameters that reside systematically in various layers of IOS to achieve the best detection capabilities. Nodes are responsible at the network level for gathering data from adjacent nodes to direct attitudes, while they are responsible for receiving warning notifications at the application layer level. Subsequently, with the aid of external knowledge found on other computers, such as the GGNS system, the local decision is taken.
Going on to [39], the vehicle cross-message authentication scheme in Wireless Access in Vehicular Environments (WAVE) of the IEEE is proposed to validate the protection application in the received message. The authentication process is the generation of a signature, the transmission of a periodic security message, and the authentication of the message obtained [40].
In order to create a path, [41] suggested an AODV routing protocol based on a global positioning system (GBAODV). To enhance routing performance, the influx of AODV routing packets using GGNS devices was restricted. It was found that their GBAODV technique decreases the load of the network more than the AODV technique. As a consequence, it decreases the number of broken connections and the amount of packet failure. Also, the average end-to-end delay is shorter than AODV when using GBAODV. However, the efficiency of the GBAODV system for packet loss was not adequate when considering various path scenarios. Also, this analysis was focused on only eight nodes.
In order to increase the efficiency of the routing protocols, [42] introduced vehicle movement details into the route discovery process based on an AODV for VANET applications. In the proposed protocol to achieve more reliable routing, they considered the total route weight (TWR) (based on location metrics) and the expected expiration period. The suggested protocol decreases the routing burden even more and guarantees more reliable links, they found. However, there is no substantial change in the percentage of packaging disposal in the current protocol [3].
This article proposes a complementary scheme for the routing protocol in enhanced V2V networks that incorporates time savings and collision prevention at relay stations. The simulated results assuming an intersection show that this significantly increases the average proportion of package delivery with the proposed scheme. To make this possible, we use the route information available in the geolocation equipment installed in vehicles today. Next, the proposed scheme and its implementation will be discussed.
We divide this section into three parts: Section 3.1 describes how we can use the position information available in the vehicles to improve the routing deliver packet scheme, while Section 3.2 introduces our proposed solution to build all possible route communication using graphs and 3.3 we formalise our proposed algorithm to solve the problem and the rest details.

Position information
Geographic and Location-Based Services (LBS) for ITS are services that exploit knowledge of where a user is located in order to deliver location-specific services. Therefore, LBS must identify its geographical location [43]. GNSS, short of Global Navigation Satellite System (GPS, Galileo, GLONASS) is the most accurate means of identifying location (where the sight of enough of their satellites is known). ITS applications are location-based facilities. Time-critical and many other safety facilities, including collision avoidance, all rely on the precise location of the vehicles involved. However, the role LBS plays in providing security services where knowledge of the exact location is a vital necessity.
With the city map and the aid of GNSS on board the vehicles, it is possible to determine if the vehicles are on the same street, to determine the crossing points, and if it is possible to have data communication between the hosts. This information is important to define the message routing hops, and calculate the distances between the vehicles, as shown in Figure 2. The communication area of a given vehicle is calculated using the area formula of an ellipse, according to equation 1. Where a is the Line-of-sight propagation (LOS) and b is the StreetLength/2 [44,30].
To determine whether two vehicles are within communication range, the following equation 2 was used. If the inequality results are less than or equal to 1, then the vehicle is within the communication range, and if the inequality is not satisfied, the car2 is outside the communication range [45,30].
As we can see in Figure 2 the vehicle in the crossroad can be a bridge to the other two cars communicating.
In this work we list some assumptions and illustrations: Instead of infrastructure mode, vehicle-to-infrastructure (V2I), the vehicle sends the packets only in AdHoc mode, it means vehicle-to-vehicle (V2V) communication.
Every vehicle knows its own coordinates, and a main controller of the perimeter knows all vehicle positions. Then, the controller calculates the best route using our proposed solution and distributed the route table to the hosts that are part of a certain communication perimeter, for example, the city centre.
With the information of the routes of the vehicles that are part or will enter the perimeter of the controller, it is possible to create a graph with all the possibilities of communication between all hosts and use the time that this communication can occur, as the graph's main attribute. Now we will show the formalisation of the route problem and detail the proposed algorithm to solve the issue.

Algorithm
Another important property is the DeliveryT ime of the packet, that is, when the package will be delivered to the destination. This value is calculated using equation 4 and is used in the evaluation chapter to compare the gain of our proposed solution with other existing protocols. The maximum value of t is when the package will be delivered to the destination P k (n D−1 , n D , T, W ).
It may be possible to tie the shortest time in the transmission window. In such cases, use the following rules to choose the best path: 1 Which path has the least hops; 2 Which is the path with the smallest Aggregate Path Loss as shown in equation 5 [46]. Path loss is an attenuation that decreases the power density of any electromagnetic wave as it transmits through the channel. There are distinct elements of path loss varying from natural spread of the radio wave, diffraction path loss grows because of interference, to saturation path loss that exists for existence of a signal that is not transparent to electromagnetic waves [47]. We use the sum of all path loss between the communication hops. As a variable in the equation, we have D as Distance between the transmitter and the receiver and the λ is the free space wavelength defined as the ratio of the speed of light in meters per second to the carrier frequency in Hz as shown in equation 6. The equations are based in [48] work.
For example: As shown in Figure 3 bold edges, Let P 1 be the set of available transmission path for a source node n 8 and destination node n 12 , while n 4 , n 0 , n 13 , n 12 is the forwarding nodes in the considered scenario, so P 1 = {n 8 , n 4 , n 0 , n 13 , n 12 }. Among all the available paths, the algorithm tends to choose the one with the shortest time t, that is, the end-to-end delay will be the shortest possible, taking into account the time t attribute, where the time t of the neighbour must be greater than the t of the current node. Because of this, it is not possible to use the path between n 4 and n 0 marked in dashed red line, because the attribute of the edge (n 4 , n 0 , 10, 1), T = 10 is less than the T = 12 of the previous edge (n 8 , n 4 , 12, 2). So, it is necessary to use the route available at time T = 18 in (n 8 , n 4 , 18, 1). Therefore, the bold path represents the best route selection to transmit the information.
As we can also see in the Figure 3, nodes 6 and 9 are disconnected from the rest of the nodes, this means that they can only communicate with each other.
The following Algorithm 1 represents the pseudo-code used to represent the V2V routing protocol assisted by controlled navigation for urban areas.
We divide the algorithm into three steps. First step identifies the vehicle positions at each time stamp, based on the input files of the city map and the gpsDat, which are the routes of the vehicles are in the same perimeter. Calculate the distance between vehicles and create a list with neighbours in different timestamps. Based on the distance between neighbours, it is possible to calculate the path loss, based on equations 5 and 6. In the second step, we create the graph with all neighbours in unique time periods. As the aim of this algorithm is to deliver the package as soon as possible, the smallest delivery time available between the vehicles is chosen and we discard the other information. In the third step, we choose the shortest path using the Short Path Temporal (SPT) algorithm, between the source and destination entry, based on delivery time information. If there are alternative paths, first, the 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 Algorithm 1 Steps of the NAV2V algorithm: Input: gpdDat, map.xml, source, destination Output: path route between source and destination Step 1 : Identify positions and crossing points ∀ time/position available: Calculate distance between cars in the same street if distance <= communication range then create a list of possible neighbours end if Calculate the time t neighbours will be available Calculate the P Loss neighbours Step 2 : Graph Generate all possible communications, with minimum time t Step 3 : Run SPT Algorithm Select the short path, based on time t if There is more than one option: then Choose the minimum Hop if There is more than one option: then Choose the minimum PLoss end if end if Result: Return the short Path END of algorithm algorithm chooses the one with the least number of hops and then the one with the lowest aggregated path loss, as explained before. Next, we will explain the Short Path Temporal algorithm.

Short Path Temporal algorithm -SPT
To solve our graph problem, we need to find the shortest path of the graph, considering the time as the most important attribute. Because of this, it is not possible to consider other algorithms for short path, as they try to minimise the weight, without taking into account the time dependence. In other words, the time constraint means that the next edge must have a time equal or greater than the current node time, as the function f e (t) in equation 3.

Simulated environment
Researchers typically use computer simulation programs to test and analyse the results of their research, as they are more flexible and less costly compared to the real environment.
We conducted a computer-based simulation to evaluate the performance of the proposed scheme. Network Simulator 3 (NS-3) [50] was used as the simulation tool. NS-3 is a discrete event network simulator, written in C ++ and Python and is a computationally efficient and practical computer language.
The proposed research work has used network simulator version 3.30 for the purpose of simulating the network and generating required data for analysis. Under this, the module for NetAnim 3.108 is included for the purpose of animating our network scenario and observing the packet flow in the network while our simulated vehicles communicate among each other. Using NS-3 the vehicle movements are simulated by uploading the map and vehicles setup first. The mobility-trace.cc file is utilised from the NS-3 directory to incorporate the Net-Anim simulator code to be run directly from the NS-3 simulator. Further the object.tcl file is created in the same directory, which would help us in creating our own defined scenarios for each of the protocols that are used from the same source file with just a simple creation of new object instance [1].
A generic grid city map was used as a scenario in our simulation, called Manhattan [51] grid. The time agreed by us in this simulation work was 50 seconds, and the Urban Mobility (SUMO) [52] simulation was introduced in order to obtain a realistic scenario.
The SUMO software is an open source software which provides an easy platform for simulating the land mode of transportation. It creates a simulation wherein the vehicles will be treated as nodes and with pedestrians in view, a network is created. It comes with an improvised tools setup which provides a very handy platform for creating various scenarios which a researcher could think for their experiment [1]. Figure 4 represents the urban scenario set in the 5 x 5 grid topology. In this model, movement pattern of mobile nodes in urban environments is simulated using a grid road topology. This mobility model contains horizontal and vertical road, where nodes are randomly placed on the map at the beginning of simulation and they are allowed to change their lanes. When a node reaches the intersection, it can continue by turning left, turning right, or going straightforward randomly [23].
The Diagram Process flow for experimental setup, is represented in Figure 5, and several steps are necessary to set up this model. First, we used the Netedit tool to create the map (Manhattan grid style) and the buildings identified as the map.xml and buildings.xml files. These files are also used in NAV2V and NS-3 respectively. OpenStreetMap and Polyconvert could be used for the same purpose, as used in other works [53,54,55,2]. After that, the randomTrips tool was used to generate the random routes of the vehicles and the number of vehicles in each simulation, identified as route.xml. To finish the first part of the simulation in the SUMO tool ,   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63 64 65 the configuration file sumo.cfg was used. As a simulation output in SUMO through fcd-output, the simtrace.xml file is generated. Through the traceExporter tool it is possible to create the gpsDat file that will be used in the python simulation of the proposed NAV2V algorithm and the mobility.tcl file that will be used in NS-3. Beside the buildings.xml file it is necessary to generate the obstacles and the mobility.tcl file, which is the vehicle's movement file, it is necessary to use the vanet.cc file. This file is a modification of the vanet-routing-compare.cc file available in version NS-3.30.1 and contains all the configuration for the simulation. Table 1 represents the parameters used in the simulation.
Because of the NS-3 simulation, we have the pcap files for each host (vehicle) in the simulation as output. Therefore, it is possible to filter, using the tcpdump tool, and the extracted information will be used for evaluation and comparison between the existing protocols and the proposed solution.
For the simulation done in Python, the map.xml and gpsdat files were used and processed in the proposed NAV2V algorithm, as explained in the previous chapter. The relevant result information is filtered and used for comparison with the NS-3 results.
In this work, we consider that BSM (Basic Safety Message) [56] information will be distributed to all vehicles involved in a given urban area. Our simulation uses the mobility trace files produced by SUMO simulations in NS-3 during which every vehicle emits a BSM 10 times per second. For example, if we have 20 vehicles traveling in a common area and we intend to send BSM information in that specific area, that means a distribution of 380 messages, that is, the P ermutation(20; 2), as shown in equation 7. Where a given car is the source of the BSM message and all other vehicles involved in the same coverage area, as the destination of the information. In the NS-3 simulation, the protocols used are AODV, DSDV, OLSR and GPSR to compare with our proposed NAV2V protocol. P (n,r) = n! (n − r)!
Next, we will describe the obstacle model used in this paper.

Obstacle Model
Obstacles, such as buildings and trees, interfere with radio wave signal propagation by contributing fading and shadowing effects. To produce results that accurately reflect real-world typologies, models must address the radio-interfering conditions that obstacles present. Failing to account for the effects of obstacles can therefore inaccurately overstate network performance. We implemented an obstacle shadowing model for the NS-3 network simulation tool set and tested using a script for wireless vehicular AdHoc network (VANET) scenarios and obstacle data from a map [57]. Figure 6 represents the obstacle model used in our simulation.

Results and discussion
To show the efficiency of the proposed algorithm, simulations were carried out with four existing protocols: AODV, DSDV, OLSR, GPSR and our proposed scheme NAV2V. Each protocol was simulated with different numbers of cars on the map; 5 to 30 cars being evenly distributed and using random travel to generate traffic, as explained in the topic of the simulation.
The following metrics were used to evaluate the performance of the proposed protocol: (1) Percentage proportion of packets delivered rate, that is, the number of packets that are received correctly at a destination in relation to the number of packets that were sent by the origin. (2) Delay time in seconds, the end-to-end delay between source and destination. In addition, we use the (3) average aggregate delay time, where we average the delivery time of messages with different vehicle densities, to compare each protocol in a generalised way. Figure 7 show us the total possibilities of successful communication between all vehicles, measured in percentage. As the result show, in our proposed scheme it is the same as the AODV protocol when the network is sparse (with only 5 vehicles on the map). However, as the number of nodes increases in the simulation, the performance of the proposed NAV2V scheme is noticeably better, reaching 30% gain.
The estimates point and confidence intervals of the aggregate number of possible communications of all simulation tests can be seen in Figure 8. According to the obtained results, NAV2V can offer until 25% higher communication possibility compared to AODV. Because, as the density of cars increases, so does the number of possible paths to transmit the data, and our algorithm is capable of making the optimised selection.
After estimating if communication was possible, according to the previous graphs, we measure the delivery time for these messages. The following figures represent the values in seconds in absolute time of the simulation. We assume in this work that all vehicles send the BSM message in a broadcast, since the zero time of the simulation and continue sending the message at a rate of 0.1 messages per second. We used this method only for the purpose of comparing the protocols. In the real environment, messages are triggered, according to the occurrence of security events.
We illustrate the average delivery time of the packages in Figure 9, separated by simulation. As many more packets are served, the average number of the delay can also increase, but even so the value is close to the other protocols.
Although the sum of the average package delivery time gain is small, as we can see in Figure 10, we must remember that this scheme reached up to 30% more packages successfully delivered.
Then, to view the delivery time of the first package between origin and destination, we generate a boxplot, illustrated in Figure 11. As the results show the differences in the time delivery packages between the routing protocols used in this work. And we can verify the time savings that the NAV2V scheme provided.
In order to evaluate the behaviour of NAV2V, we measured the time required to deliver the message from the source to the destination. The results of Figure 12 was generated to compare the behaviour of using useful bandwidth, that is, the Goodput in the simulations. As you can see, the NAV2V solution has a better use of the bandwidth when compared to the other protocols.
Packet reception rate is an important performance metric to compare the efficiency of different VANET routing protocols. As well as the useful bandwidth, we can see that the packet reception rate is also noticeably better than the proposed algorithm, highlighted in Figure 13 per simulation time and in Figure 14 as average rate in all simulation time.
A novel automobile antenna working in a broadband frequency is proposed. The antenna obtains the good omnidirectional radiation pattern and the impedance matching without any additional matching network, which are improved by loading the toothed capacitor and the impedance matching disk. The electric field distribution and transmission coefficient of the antenna installed on the vehicle roof are simulated, which shows high gain with low loss. The good omnidirectional radiation performance enables the antenna to effectively connect the mobile cellular networks and the IoV systems. Therefore, the network coverage performance is improved by the automobile antenna. Since the antenna without any additional matching network, the whole antenna structure is compact and easy to be installed on the top of the vehicle. Due to the high reliability and low cost, the proposed automobile antenna could be widely used for the V2X communication system.

Discussion
A new algorithm for AdHoc communication between vehicles has been proposed. The algorithm uses the GGNS information, already available in the vehicles' onboard equipment. Simulations were made using the latest applications available for this environment. The results and improvements in performance of the algorithm was proven in the simulations. This proposal can be widely used for the V2V communication system and later adapted for the V2X system as future research.

Conclusion
In VANET, the nodes travel quickly, and the topology also changes, so the routing architecture is a challenge. The next generation of vehicular network communication systems must work collaboratively to provide excellent performance. In this article, we have investigated in depth the routing protocols AODV, DSDV, OLSR and GPSR, which are the most used and adequate today in this type of system. A scheme to complement the routing, called NAV2V, has been proposed. This scheme uses the geolocation information available in a Global Navigation Satellite System (GGNS), which are devices embedded in automobiles nowadays, and which are an essential item for the evolution of communications between vehicles until reaching their evolution, which are full autonomous cars.
The simulated results show that the proposed scheme significantly improves the performance of V2V communication. The results of the simulations showed that there were problems with connectivity and selections for the next hop. NAV2V addressed connectivity and next hop selection problems and also used route mechanisms to forward packets to the next available node for forwarding to the destination. We compared the performance of NAV2V with the other routing protocols implemented in the NS-3 simulation tool. The terms of package delivery fees and end-to-end delays have been considered. The simulation results showed that the package delivery rate for NAV2V can reach up to 30% better in some cases.