Hybrid routing protocol for VANETs: Delay and trust based approach

Abstract

VANET is mainly aimed at providing safety and security related information and traffic management. In future, VANET contributes to smart transportation system. Based on vehicle mobility, different routing protocols and traffic models were developed. In routing, trust between vehicles place an important role to forward safety related information. This paper aims at design of trust and delay based routing for hybrid communication in sparse VANET to avoid network attacks by malicious nodes. The proposed hybrid routing protocol works on the computation of trust in between vehicles and message reachable time (MRT). Route selection is done by considering the highest trust factor and minimum MRT. The performance effectiveness of the proposed scheme is evaluated by comparing with the Delay-aware and Backbone-based Geographic Routing for Urban VANETs (DBGR). The proposed scheme exhibits better performance in terms of packet delivery ratio, bandwidth utilization, end-to-end delay and control overheads.

Keywords

Vehicular Ad hoc Network (VANET)Vehicle-to-Vehicle Communication (V2V)Vehicle-to-Infrastructure Communication (V2I)

1. Introduction

A VANET is an emerging technology with vehicles equipped with a communication device called on-board unit (OBU), and a set of stationary units along the road, referred to as road side units (RSUs) which act as a portal to connect different communication networks. The OBU of each vehicle has wireless interface to connect with all vehicles and RSUs in its communication range. By employing Vehicle-to-Vehicle (V2V) and vehicle-to-RSU (V2R) communications, many applications are supported by VANET such as road safety, passenger infotainment, and vehicle traffic optimization for which VANETs have received significant support from industry, government,and academia over the globe. Due to VANET unique feature such as high speed, dynamic variations in the topology takes place which leads to many challenges such as data aggregation, clustering, data validation, data dissemination, routing, security, etc. [25,45].

Due to highly dynamic topology, on time data delivery and exchange of data packets is a challenging issue in VANET. Hence exchange of emergency alert messages are required to transmit against intentionally inserted attacks in the network. The different classes of attacks and their identification in VANET is presented in [46]. The main attacks affecting the VANET performance are Denial of Service (DOS), Distributed Denial of Service (DDOS), sybil and timing attack. The security related attacks are application attack, timing attack and monitoring attack. Node impersonation attack, blackhole attack, timing attack, social attack and monitoring attacks are mainly caused in V2V communication whereas DOS, DDOS, sybil and application attacks are affecting hybrid communication. The algorithm related to security mechanism in AODV to detect and prevent blackhole attack is proposed in [49].

Due to such attacks induced in VANET, security is an important factor to protect the driver and vehicle from flawful messages and also to get network services. One of the network attack related to safety applications is attack due to malicious nodes. Malicious nodes send false information to the vehicles. These malicious nodes prevent the genuine users to get the emergency services such as accident alerts, traffic jam notification etc. Thus in our proposed work, secured routing with minimum delay for hybrid communication is addressed. The proposed scheme prevents the network from malicious node attack (also known as denial of service attack).

To be specific, our contributions are as follows. (1) Identification of routes based on trust of neighboring nodes, (2) Selection of route with minimum delay and maximum trust, (3) Enhancing the basic routing performance metrics such as Packet Delivery Ratio (PDR), end to end delay, control overhead and throughput.

The remaining part of the paper is summarized as follows: Details about the literature survey is given in Section 2. Section 3 proposes our hybrid routing protocol for VANETs. Simulation inputs and simulation procedure are presented in Section 4. Simulation results are discussed in Section 5.

2. Related works

Many routing mechanisms for VANET have been proposed. Proactive and reactive routing algorithms for VANETs with different mobility models and network scenario are presented in [6,18,23,32,50]. Position and topology based routing protocols are explained in [9,29,38]. In [12,33,41,51], routing protocols using some of the intelligent techniques such as fuzzy logic, software agents, bio inspired, game theory, etc., are proposed to improve delay, packet delivery ratio, security, and throughput for high dense VANET environment.

User mobility and node density (ESRA-MD) are used to get the efficient and stable route for urban VANET environments [8]. A centralized routing scheme by predicting the mobility of vehicle is proposed in [10,47] to minimize the delay for VANET using an artificial intelligence powered Software-Defined Network (SDN) controller. Based on the information of network collected from RSUs and base station (BS), switches computes optimal routing paths to minimize the overall vehicular service delay. An Ant colony optimization (ACO) is used to compute optimal route in VANET [30] using local road segment delay and global delay from current intersection to the terminal intersection of the destination. In [14], an Artificial Spider-web based Geographic Routing protocol (ASGR) is addressed by using the different parameters of the paths like connectivity, number of hops and delay.

Applications and security approaches for intelligent transportation system of VANET such as authenticity, integrity, availability, confidentiality, anonymity and non-repudiation to provide the secure communication between Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) are presented in [2,36]. The routing protocol based on trustworthiness of the path and the number of hops in order to find the optimal route for transmitting reliable information is proposed in [26,48]. Due to selfishness, some vehicles exhibit various misbehavior such as packet dropping leads to degrade the efficiency of network. Hence it is necessary to stop such vehicles to participate in network communication by proposing an effective revocation scheme for disconnected delay tolerant VANETs [5].

To prevent the attacks in VANET from malicious nodes a trust management is approached in [3]. In [19] secured routing protocol with minimum delay is presented to prevent the VANET from Denial of Service attack. The swarm intelligence based secure routing in VANET is proposed in [28] which uses Intelligent Water Drops (IWD) method. This method protects the VANET from blackhole and wormhole attacks.

To decrease the hop count and delay, and to increase the network throughput and the packet delivery ratio, a routing protocol based on cost metric is proposed in [39]. In some situations where the target vehicle is moved away from the expected range and in sparse traffic conditions connectivity is the main issue for routing in VANETs. To overcome such situations Connectivity aware Minimum-delay Geographic Routing (CMGR) [42] and Connectivity Aware Intersection-based Routing (CAIR) [13] protocols are proposed to select an optimal path with high probability of connectivity and low delay. Light weight Intersection-based Traffic Aware Routing (LITAR) protocol [15] for V2V communication in urban scenario uses two algorithms where route decision is made based on density of the network, network connectivity and distance from destination to minimize overhead of the network due to traffic condition measurements.

Many time critical applications competing for wireless access leads to congestion and “hotspot” problem in urban areas with high traffic scenario. The algorithm proposed in [20] choose the optimal route with the knowledge of traffic conditions by calculating the load status in order to avoid congestion. In Real Time Intersection based Segment Aware Routing (RTISAR) [7], the traffic segment status based on connectivity, density, load segment, and cumulative distance toward the destination is used to choose the next intersection on road to forward the packets. To support Internet access in urban environment using cellular networks, a novel infrastructure-based dynamic and Connectivity Aware Routing protocol (iCARII) is proposed in [4]. Considering the real time traffic conditions for route selection precess in case of link connection/disconnection and historical traffic condition is explained in [22].

To exchange the information over distances more than communication range of the vehicle, multi hop communication is used. To ensure stable multi hop communication with low delay for high mobile vehicles a routing protocol is presented in [35], based on medium access control (MAC) protocol. Time Division Multiple Access (TDMA) scheduling algorithm is used to select the intermediate vehicles to transmit/receive packets for longer distances. Secure routing is an important challenge in VANET as some malicious nodes try to disrupt the route discovery process in routing. In [16], survey of various techniques to improve routing protocols for secure routing process by enhancing the trust among different nodes in VANETs. In VANET all vehicles must work with full cooperation to generate and broadcast some important messages for safety purpose. Hence trusted and secured routing protocol named as Trusted Vehicular Ad-hoc On-demand Distance Vector (TVAODV) is explained in [37]. A secure group authentication technique for VANET is proposed in [34]. It operates in three phases: (1) a group is formed which consists of several vehicles within the communication range using beacon message, (2) the message is checked for reliability, and (3) the co-operative message authentication is performed.

To select the next-hop in vehicle communication the “bridging approach” for message forwarding is used in [21] i.e., vehicles on the east select from west and vice versa to improve the Quality of Service (QoS) parameters like packet delivery ratio and throughput. In city scenario with buses and cars as nodes, a hybrid protocol is proposed in [44]. It uses the greedy forwarding approach and the store-carry-and-forward approach to improve the performance parameters.

Blockchain is the recent technology used now a days to build trust and track asset in network. There are two types of blockchain: Private and Public. The routing using both private and public blockchain called as regional blockchain is used to prevent a minimum 51% attack in VANET is proposed in [43]. In [24], MAC layer attacks in VANET are reduced by using block chain method to increase the security and privacy to the users. Fog computing is a method where computing resources are placed between data source and data center. It uses edge devices for computation. The current research issues and future opportunities of fog computing in VANET are presented in [11] along with its characteristics and services. Streaming of accident and disaster videos to the nearby vehicles for V2V communication using fog computation is presented in [27]. It uses speed, position, and video recording angle parameters of each vehicle to minimize unnecessary exchange of video data. Fog computing along with intelligent optimization algorithms is described in [40] to secure the VANET. Back propagation with feed forward neural network is used to differentiate the authenticate vehicles and attacked vehicles.

Machine learning is an application of Artificial Intelligence (AI) provides the ability to automatically learn and improve from experience without being explicitly programmed. Essential data for routing in VANET using Deep Reinforcement Learning (DRL) method is described in [53]. In this work, the vehicle movement is predicted by DRL and transmits the packets with proper route. In [52], the improvement of Greedy Perimeter Stateless Routing (GPSR) using Support Vector Machine (SVM) method is used named as Greedy Machine Learning Routing (GMLR) algorithm. In [1], a Q-learning based routing in VANET to improve the speed of routing process is presented.

Comparison of existing routing protocols is presented in Table 1. Limitations of these routing protocols are listed as follows: requirement of trust between vehicles, needs better connection between vehicles and Internet through RSU when trusted nodes are not their in the neighbors, takes more delay to send the packets, minimum flexibility for rapid change in important information parameters, and more loss of data packets due to rapid change in topology and link disconnection.

Table 1
Summary of routing protocols in VANETs

Sl. No. Routing Scheme VANET Scenario Mechanism Trusted Vehicles PDR End to End Delay Control Overhead Throughput

1. DBGR [31] V2V Based on present and previous traffic No More Less Less –

2. LPRV [10] V2V By predicting future localization of nodes near destination No More More – –

3. IDRA [30] V2V Based on Ant Colony Optimization (ACO) No Less Less Less –

4. ASGR [14] V2V Based on artificial spider web No More Less – –

5. CSRP-TDA [48] V2V Based on trust and channel state Yes – Less Less –

6. LITAR [15] V2V Based on road density, network connectivity and distance towards destination No More – Less –

7. RTISAR [7] V2V Intersection based No More – Less –

8. TVAODV [16] V2V AODV with trust between vehicles Yes More Less Less More

9. SGA [37] V2V Group authentication for group of vehicles Yes More – Less More

10. RBVNPA [44] V2I Regional block chain Yes – Less Less –

11. CPRV [11] V2V Fog computing Yes More Less – More

12. SIVNFC [27] V2V Feed forward back propagation algorithm Yes – Less Less More

13. GMLR [53] V2I Support vector machine learning No More Less – –

Sl. No.	Routing Scheme	VANET Scenario	Mechanism	Trusted Vehicles	PDR	End to End Delay	Control Overhead	Throughput
1.	DBGR [31]	V2V	Based on present and previous traffic	No	More	Less	Less	–
2.	LPRV [10]	V2V	By predicting future localization of nodes near destination	No	More	More	–	–
3.	IDRA [30]	V2V	Based on Ant Colony Optimization (ACO)	No	Less	Less	Less	–
4.	ASGR [14]	V2V	Based on artificial spider web	No	More	Less	–	–
5.	CSRP-TDA [48]	V2V	Based on trust and channel state	Yes	–	Less	Less	–
6.	LITAR [15]	V2V	Based on road density, network connectivity and distance towards destination	No	More	–	Less	–
7.	RTISAR [7]	V2V	Intersection based	No	More	–	Less	–
8.	TVAODV [16]	V2V	AODV with trust between vehicles	Yes	More	Less	Less	More
9.	SGA [37]	V2V	Group authentication for group of vehicles	Yes	More	–	Less	More
10.	RBVNPA [44]	V2I	Regional block chain	Yes	–	Less	Less	–
11.	CPRV [11]	V2V	Fog computing	Yes	More	Less	–	More
12.	SIVNFC [27]	V2V	Feed forward back propagation algorithm	Yes	–	Less	Less	More
13.	GMLR [53]	V2I	Support vector machine learning	No	More	Less	–	–

3. Proposed work

This section explains the proposed trust and delay based routing for Vehicle to Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) i.e., hybrid communication under the VANET constraints such as rapid change in topology, high dense network, link failures etc. The proposed hybrid routing protocol operates in the following sequence: (1) based on the prior knowledge about the neighbors, determine trust between the source vehicle and neighboring vehicles, (2) identify trusted neighboring vehicles (i.e., vehicles having trust factor more than the threshold assigned), (3) the process repeats until the destination vehicle is reached, (4) compute the message reachable time (MRT) from source vehicle to the destination vehicle for identified routes of involved trusted neighbors, (5) select the route among the available routes based on highest trust factor and minimum MRT.

3.1. Network environment

The vehicular network shown in Fig. 1 has highway scenario with dual lane consisting of vehicles and RSU for communication. Few assumptions are made within network for communication purpose. Every vehicle is installed with GPS, digital map, sensors, network devices, which gives the information about road segment and computing devices. Wi-Fi is used as technology of communication between vehicles which have an ability of charging their batteries constantly. It is assumed that all vehicles are traveling in same direction.

Fig. 1.

Network environment.

3.2. Preliminaries

Trust originator node: It is the node that calculates the trust of other nodes to send the packet.

Neighbor nodes: The nodes coming in the communication range of originator node.

Trusty node: The node whose trust is above the assigned threshold value is called as trusty node.

Advisor node: The node which gives the trust information about the neighbor nodes based on the history in case of all neighbor nodes fail to meet the assigned threshold value.

Message Reachable Time (MRT): The total time taken by nodes to reach the message from source node to the destination node.

3.3. Delay and trust based routing scheme

The proposed hybrid routing protocol protects the network from denial of service attacks where malicious node activities are minimized. The different parameters used to identify the malicious nodes in network are received signal strength, less hop count and more sequence number, link life time, trust interval and number of transactions [17]. The transactions may be positive or negative. Positive transactions are sending the information regarding current position, road condition and traffic condition to the nearby vehicles. In our research work, we have considered the malicious node as node which has less number of transactions and trusted node as a node with more number of transactions.

The proposed routing protocol operates in the following sequence: (1) based on the prior knowledge about the neighbors, determine trust between the source vehicle and neighboring vehicles, (2) identify trusted neighboring vehicles (i.e., vehicles having trust factor more than the threshold assigned), (3) the process repeats until the destination vehicle is reached, (4) compute the message reachable time (MRT) from source vehicle to the destination vehicle for identified routes of involved trusted neighbors, (5) select the route among the available routes based on highest trust factor and minimum MRT.

Trust computation: Two types of trust are calculated, direct trust and advised trust. Direct trust is computed based on the vehicle individual trust which are above the assigned threshold value. Advised trust is computed depending on the feedback taken from the advisor node in the network. Advisor node gives the trust information about the neighbor nodes based on the history if all neighbor nodes fail to meet the assigned threshold value.

Direct trust computation: As shown in Fig. 2, the originator node (source vehicle S) fetches the trust value of its neighbors from RSU 1. Direct trust value of the neighbor vehicle (whose value is not fetched from the RSU 1) is computed using the previous knowledge that the trust originator node has upon the neighbor trusty node. The originator node has more number direct knowledge about its neighbors which may be positive knowledge or negative knowledge based on importance level of every message. For example in VANET, accidental alert message has more importance level than searching shopping mall message. If the neighbor trusty node confirms the alert message then the originator node considers this knowledge as positive for computing direct trust. Each knowledges are having different importance level. In the proposed work, direct trust computation ${TC}_{D}$ is computed using equation (1), if the value of total number of knowledge (TK) is greater than or equal to 1 else the value of ${TC}_{D}$ is equal to zero. $\begin{matrix} (1) & {TC}_{D} = round (\frac{\sum_{j = 1}^{n} (W_{j}) ({PK}_{j})}{TK}) \end{matrix}$

Where j represents knowledge, n represents number of neighbor nodes in range of originator node, $TK$ is total number of knowledge that the originator node may have on the trusty neighbor nodes, $W_{j}$ is the weight of knowledge that replicate the importance level of message. ${PK}_{j}$ represents number of positive knowledges. The value of $PK$ is 1 if the knowledge j is positive. Depending on weight $W_{j}$ the trust value is increased or decreased. The value of weight is predefined which lies between 0 and 1. Higher the weight value with more positive knowledges, trust is more. The function round is used to round the decimal value to its closest tenth. It is considered to get the values in between 0 and 1 as given in Table 2.

Fig. 2.

Trust computation.

Table 2

Degree of trust

Trust value	Semantics
0	New or unknown
0.1–0.2	Very low
0.3–0.4	Low
0.5–0.6	Medium
0.7–0.8	High
0.9–1	Very high

Advised trust calculation: It is used where the originator node has more neighbors. Whenever the trust originator node doesn’t have enough prior direct knowledge about trusty node in its neighbors, the originator node may send queries to advisor nodes to get advise about trusty node. Here it is expected that the advisor nodes are as of now having the trust value ${TV}_{j}$ about trusty node based on its own assessment. The originator nodes may inquiry few nodes in the network for exact assessment. The advised trust ${TC}_{A}$ is computed from equation (2). $\begin{matrix} (2) & {TC}_{A} = \frac{1}{n} \sum_{j = 1}^{n} ({TC}_{D j}) ({TV}_{j}) \end{matrix}$

Where, ${TC}_{D j}$ is the direct trust value of trust originator node on advisor node.

Mixed trust computation: When the originator node is having both trust values (direct and advised) then mixed trust $T_{mixed}$ (equation (3)) is used to assess the trust value. $\begin{matrix} (3) & T_{mixed} = {TC}_{D} + (1 - {TC}_{D}) ({TC}_{A}) \end{matrix}$

( $0 ⩽ {TC}_{D} ⩽ 1$ , $0 ⩽ {TC}_{A} ⩽ 1$ )

If ${TC}_{D} = 1$ , the originator vehicle is having adequate value to fulfill the trust necessities. Hence, the advised trust value is not required. If ${TC}_{D} = 0$ , the originator vehicle will totally relies upon the trust value given by the advised vehicle. If the direct trust value is more, the advised trust value is less and vice versa.

Computation of message reachable time (MRT): The Message Reachable Time (MRT) [17] between two nodes is the time taken by nodes to send the packet which is calculated as ratio of distance between two nodes and speed difference between those nodes as given in equation (4) and (5). $\begin{array}{l} (4) & S_{d} = V_{d} - V_{s} \\ (5) & MRT = \frac{D}{S_{d}} \end{array}$

Where $V_{d}$ and $V_{s}$ are velocity of destination vehicle and source vehicle respectively and D is distance between two vehicles given by Euclidean distance formula (equation (6)). $\begin{matrix} (6) & D = \sqrt{{(x_{d} - x_{s})}^{2} + {(y_{d} - y_{s})}^{2}} \end{matrix}$

Where $x_{d}$ , $y_{d}$ are co-ordinate of destination node and $x_{s}$ , $y_{s}$ are co-ordinate of the source node.

Once the MRT between two nodes is calculated, the total path delay is computed using equation (7) which is the time taken for the message to reach its destination from the source vehicle. It is calculated for all routes. $\begin{matrix} (7) & Pathdelay = \sum MMRT \end{matrix}$

Then the Minimum Message Reachable Time (MMRT) is computed using equation (8), which is the lesser delay amongst all total path delays calculated for trusted routes between source and destination. $\begin{matrix} (8) & MMRT = min (MRT) \end{matrix}$

Route selection: In our proposed routing algorithm, trust is computed locally by each vehicle. The trust data table must be kept update by every vehicle and the table consists of data about trusty vehicle’s ID number, direct, advised and mixed trust values. The trust data might be invigorated intermittently. The data will be lapsed if it isn’t revised within the time frame t as accepted by the system and the relating trust value becomes zero. Finally the trust decision TD is evaluated based on equation (9). $\begin{matrix} (9) & TD = T_{mixed} - {TR}_{t h} \end{matrix}$

Where, $TR$ th is the threshold value of trust risk for ongoing tasks. It is nothing but the maximum risk that can be considered for ongoing task. Hence the threshold value is defined for every task. To give an example the accident alert message should have high trustworthy value which starts from 0.7 and searching for parking slot should have low trust value which starts from 0.2 as mentioned in Table 2. If $TD ⩾ 0$ , the calculated value of trust fulfills the criteria of trust requirement. Otherwise it does not satisfy the criteria of trust requirement for an ongoing task. In that case, originator node forward the message to its corresponding RSU which was always trustworthy.

Table 3

Route decision

Trust value	Path delay	Route decision
Maximum	Minimum	Select
Maximum	Maximum	Select
Minimum	Maximum	Don’t select
Minimum	Minimum	Don’t select

Route decision: Once all the trust paths between source and destination are calculated the MRT for each path is calculated using equation (5). Then the path with maximum trust and minimum MRT is selected as final path to forward the message using Table 3. The trust value greater than the set trust threshold (0.6) is taken as maximum trust and lesser is taken as minimum trust value. The maximum and minimum path delay depends on the calculated value in seconds. Thus any safety message is delivered with minimum delay and more trust by avoiding the malicious node for routing. Sometimes even though the MRT is more the route with more trust is selected. If no trustworthy vehicles in the path then packet is forwarded through respective RSU.

3.4. Algorithm

See Algorithm 1.

3.5. Example scenario

The Fig. 3 shows the example scenario of trust and delay based hybrid routing protocol for VANET. A highway scenario with six vehicles is considered, where V1 is the source and V4 is the destination vehicle. The source vehicle V1 sends RREQ packet to RSU1 to get the information of neighbor vehicles (V2, V3 and V6). The information is trust value, number of positive transactions and negative transactions of vehicles V1, V2, V3 and V6. The RSU1 sends RREP packet consisting of these information values. Based on the values, the source vehicle V1 selects neighbor vehicle having high trust and more number of positive transactions. The trust value and delay between two vehicles is calculated using equation (3) and (5) respectively. The initial trust and delay values are tabulated in Table 4. Now there are two paths available from V1 to V4. The total path delay computed for first path i.e., V1-V2-V5-V4 is 0.47 seconds and for second path, V1-V3-V5-V4 is 0.59 seconds. MMRT is the path with minimum delay i.e., V1-V3-V5-V4. If V2 and V3 has low trust value, then the source vehicle transmits the packet to RSU1. In this way the proposed routing algorithm selects the best route based on minimum delay and maximum trust.

Algorithm 1

Proposed routing algorithm

Fig. 3.

Example scenario.

Table 4

Initial trust and delay values

Vehicle ID	Trust	Delay (seconds)
V1	0.61	0.11
V2	0.7	0.2
V3	0.64	0.32
V4	0.68	0.14
V5	0.71	0.16
V6	0.42	0.23

4. Simulation

The simulation of the proposed model is done using “C $+ +$ ” programming language on Intel Core i5 machine. The simulation model, simulation inputs, simulation procedure, performance parameters and result analysis are discussed in this section.

Table 5
Simulation inputs

Sl. No. Input parameters Specification

1 Length of road 5000 mts

2 Breadth of road 100 mts

3 n 20

4 Scenario Highway

5 R 80–100 mts

6 V 80–120 kmph

7 S 2 mts

8 ${TR}_{t h}$ 0.6

9 Data packet size 2 Kb

Sl. No.	Input parameters	Specification
1	Length of road	5000 mts
2	Breadth of road	100 mts
3	n	20
4	Scenario	Highway
5	R	80–100 mts
6	V	80–120 kmph
7	S	2 mts
8	${TR}_{t h}$	0.6
9	Data packet size	2 Kb

4.1. Simulation model

For simulation, n number of nodes are considered, whose position and ID’s are assigned randomly. We have considered highway network with dual lane and all vehicles running at a speed of V kmph with safe distance of S meters in the same direction throughout the simulation. The communication range of vehicle is R meters. It is assumed that every vehicle is installed with GPS and communication device to get the position, and its traveling speed. All vehicles are distributed uniformly with different speed and are assumed to be connected each other.

4.2. Simulation inputs

The simulation inputs taken for the proposed scheme are given in Table 5.

4.3. Simulation procedure

The process of simulation for the projected hybrid routing protocol is as follows.

Give the input as total number of vehicles.

Enter the source and destination vehicles.

Search neighbor nodes.

Calculate trust and MRT for all neighbors.

Select path with maximum trust and minimum delay.

4.4. Performance parameters

The performance metrics analyzed through simulation in our proposed routing algorithm are as follows.

Packet Delivery Ratio (PDR): The ratio of total number of packets delivered to destination to the total number of packets sent by the source. It is evaluated in %.

End to end delay: It gives the total time taken for a packet to reach from source to destination. It is measured in milliseconds.

Control overhead: It is the overhead of control packets in routing. It is measured in %.

Available paths: These are total number of trusted routes in between source and destination.

Bandwidth utilization: It is defined as the amount of data that can be transmitted in the fixed amount of time and it is expressed in bits per second (bps).

Throughput: It is defined as number of bits received by destination vehicle per unit time. It is measured in kilobits per second (kbps).

5. Result analysis

The above performance metrics defined are simulated and analyzed to check the performance efficiency of the projected routing scheme and compared with a delay-aware and backbone-based geographic routing for urban VANETs routing protocol (referred as DBGR in graphs) [31] for supporting vehicle-to-vehicle and vehicle-to-infrastructure communication in VANETs. In DBGR, the urban scenario is considered where the path with minimum delay is selected. In our research work, the trust between vehicles is computed to find the path to the destination. Trust plays an important role which secure the network from malicious nodes attack. Along with trust, the delay is also minimized in our proposed protocol.

Figure 4 shows increase in end to end delay with increase in number of vehicles by varying the speed of vehicles. It is due to more number of intermediate nodes between source and destination. The delay decreases with increase in speed of vehicle because less time is required to cross the range of the RSU. Further observation shows that the end to end delay is more in DBGR compared to our proposed algorithm.

Fig. 4.

End to End Delay Vs. Number of Vehicles.

From Fig. 5, as the number of vehicles increase within the range, the intermediate hops are more. Hence while transferring packets to destination, packet loss is more which intern decreases PDR. It also shows that as the speed increases rapid change in network takes place and while making hand-off from one RSU to another RSU packet drop increases. Thus PDR reduces with more vehicle velocity. In our proposed work PDR is better as compared to DBGR.

Fig. 5.

PDR Vs. Number of Vehicles.

Variation of control overhead by increasing number of vehicles is shown in Fig. 6 keeping the speed of vehicles constant. It is observed that control overhead increases with more vehicles because the exchange of control packets such as RREQ, RREP, and ACK are more for more intermediate vehicles. But it is less in our proposed work compared to DBGR. This is because of the involvement of RSU in identifying the trusted neighbors.

Fig. 6.

Control Overhead Vs. Number of Vehicles.

Number of available paths depending on trust and delay are calculated from source node to destination node. There are more paths as the number of vehicles increases in the network due to less failure of path finding process (as shown in Fig. 7). The paths in DBGR considers only delay. But our proposed algorithm can send packets with more trusted route in minimized end-to-end delay. The number of paths are decreased by increasing speed of vehicles, as there are chances of frequent network disconnections for high speed vehicles.

Fig. 7.

Available Paths Vs. Number of Vehicles.

Fig. 8.

Bandwidth Utilization Vs. Number of Vehicles.

Fig. 9.

Throughput Vs. Number of Vehicles.

Bandwidth utilized by the channel with increase in number of vehicles is given in Fig. 8. As the number of vehicles increase, intermediate nodes increase hence bandwidth utilization will be more. The bandwidth utilized in our proposed work is less as compared to DBGR. As the number of vehicles increase throughput decreases due to more number of intermediate nodes as given in Fig. 9. Throughput is better in proposed routing algorithm as compared to DBGR.

6. Conclusion

Vehicular Ad hoc Networks (VANETs) provide safety and infotainment services for traffic management. Because of unique characteristics of VANETs, routing with trust between vehicles is a challenging task. As the safety information is bounded with time, routing with trust between vehicles plays an important role in VANET communication. In this paper, we have developed trust and delay based routing for Vehicle to Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) i.e., hybrid communication by considering rapid change in topology, high dense network, link failures etc. The proposed routing scheme is simulated and compared with a delay-aware and backbone-based geographic routing for urban VANETs (DBGR), which shows better performance in terms of packet delivery ratio, bandwidth utilization, end-to-end delay, and control overheads. The QoS parameters for routing and the other parameters to identify malicious node such as received signal strength of node, the link life time of node can be considered as future extension of the work.

References

Z.A.

Abdulkader,

Abdullah,

M.T.

Abdullah and

Z.A.

Zukarnain, Malicious node identification routing and protection mechanism for vehicular ad-hoc network against various attacks, International Journal of Networking and Virtual Organizations 19(2) (2018), 153–175. doi:10.1504/IJNVO.2018.095419.

Agarwal, Technical review on different applications, challenges and security in VANET, Journal of Multimedia Technology & Recent Advancements 4 (2017), 21–30.

Ahmed,

Ur Rehman,

Ishtiaq,

Khan,

Ali and

Begum, VANSec: Attack – resistant VANET security algorithm in terms of trust computation error and normalized routing overhead, Hindawi Journal of Sensors 2018 (2018), 1–17.

Alsharif and

X.(S.)

Shen, ICARII: Intersection-based connectivity aware routing in vehicular networks, in: IEEE ICC 2014 – Mobile and Wireless Networking Symposium, 2014, pp. 2731–2735.

C.P.

Anyigor Ogah,

Cruickshank,

P.M.

Asuquo,

Lei and

Sun, Delay tolerant revocation scheme for delay tolerant VANETs (DTRvS), Springer Journal of Digital Communication towards a Smart and Secure Future Internet, Communications in Computer and Information Science 766 (2017), 143–164.

M.C.

Aswathy and

Tripti, An enhancement to AODV protocol for efficient routing in VANET – a cluster based approach, in: Proc. Springer Conference on Advances in Computer Science, Eng. & Appl., AISC, Vol. 167, Verlag, Berlin Heidelberg, New York, 2012, pp. 1027–1032.

Y.R.

Bahar Al-Mayouf,

N.F.

Abdullah,

Adil Mahdi,

Khan,

Ismail,

Guizani and

Hassan Ahmed, Real-time intersection-based segment aware routing algorithm for urban vehicular networks, IEEE Transactions on Intelligent Transportation Systems 19(7) (2018), 2125–2141. doi:10.1109/TITS.2018.2823312.

Y.R.

Bahar Al-Mayouf,

Ismail,

N.F.

Abdullah,

Wahid Abdul Wahab,

Adil Mahdi,

Khan and

K.-K.R.

Choo, Efficient and stable routing algorithm based on user mobility and node density in urban vehicular network, Journal of Public Library of Science (PLOS one) 11 (2016), 1–24.

Bala and

Rama Krishna, Performance analysis of topology based routing in a VANET, in: Proc. IEEE 3rd International Conference on Advances in Computing, Communications and Informatics (ICACCI), Delhi, India, 2014, pp. 2180–2184.

10.

L.N.

Balico,

H.A.B.F.

Oliveira,

R.S.

Barreto,

A.A.F.

Loureiro and

R.W.

Pazzi, A prediction-based routing algorithm for Vehicular Ad Hoc Networks, in: IEEE Symposium on Computers and Communication (ISCC), 2015.

11.

Bezerra,

Melo,

Douglas,

Santos,

Rosario and

Cerqueira, A collaborative routing protocol for video streaming with fog computing in vehicular ad hoc networks, in: Proc. International Journal of Distributed Sensor Networks, Vol. 15, 2019, pp. 1–19.

12.

Bitam,

Mellouk and

Zeadally, Bio-inspired routing algorithms survey for Vehicular Ad-hoc Networks, IEEE Communications Surveys and Tutorials 17(2) (2014), 843–867. doi:10.1109/COMST.2014.2371828.

13.

Chen,

Jin,

Pei and

Zhang, A connectivity aware intersection-based routing in VANETs, Springer EURASIP Journal on Wireless Communications and Networking 2014 (2014), 42. doi:10.1186/1687-1499-2014-42.

14.

Chen,

Liu,

Zhang and

Wang, A bio inspired geographic routing in VANETs, in: Proc. IEEE International Conference on Intelligent Transportation Engineering (ICITE), 2016, pp. 20–22.

15.

Darwish and

Abu Bakar, Lightweight intersection-based traffic aware routing in Urban vehicular networks, Elsevier Journal on Computer Communications 87 (2016), 60–75. doi:10.1016/j.comcom.2016.04.008.

16.

Debasish and

Das, Trusted and secured routing protocol for Vehicular Ad-Hoc Networks, Indian Journal of Science and Technology 10(17) (2017), 1–12.

17.

G.S.

Devi and

M.G.

Alex, Fast Emergency Message Dissemination Routing Protocol in VANET (FEMDRP), Journal of Network Communications and Emerging Technologies (JNCET) 7(2) (2017).

18.

Ding,

Chen,

Wang and

Yu, An improved AODV routing protocol for VANETs, in: Proc. IEEE International Conference on Wireless Communications and Signal Processing (WCSP), Nanjing, China, 2011, pp. 1–5.

19.

V.V.

Duduku,

Chekima and

J.A.

Dargham, Detection and mitigation of DoS attacks in VANET using secured minimum delay routing protocol, in: Proc.of the 8th International Conference on Soft Computing and Pattern Recognition (SoCPaR 2016) Springer, 2017, pp. 472–479.

20.

Feng,

Fan,

Xiao and

Wang, Load aware routing with delay threshold for Vehicular Ad Hoc Networks, in: Proc. IEEE International Conference on Parallel and Distributed Systems, 2010.

21.

Ghafoor,

Koo and

Gohar, Neighboring and Connectivity Aware Routing in VANETs, Hindawi The Scientific World Journal (2014), 1–10.

22.

Hadded,

Muhlethaler,

Laouiti and

L.A.

Saidane, TDMA-aware routing protocol for multi-hop communications in Vehicular Ad Hoc Networks, in: Proc. IEEE Conference on Wireless COmmunications and Networking (WCNC), 2017.

23.

Islam Chowdhury,

W.-I.

Lee,

Y.-S.

Choi,

G.-Y.

Kee and

J.-Y.

Pyun, Performance evaluation of reactive routing protocols in VANET, in: Proc. 17th IEEE Pacific Conference on Communications (APCC), Kota Kinabalu, Sabah, Malaysia, 2011, pp. 559–564.

24.

Kai,

Cong and

Tao, Fog computing for Vehicular Ad-hoc Networks: Paradigms, scenarios, and issues, Proc. The Journal of China Universities of Posts and Telecommunications, Elsevier 23(2) (2016), 56–65. doi:10.1016/S1005-8885(16)60021-3.

25.

M.S.

Kakkasageri,

S.S.

Manvi and

Pitt, Cognitive agent based critical information gathering and dissemination in vehicular ad hoc networks, Journal of Wireless Personal Communications, Springer 69(4) (2013), 1107–1129. doi:10.1007/s11277-012-0623-5.

26.

Kchaou,

Abassi and

Guemara, A new trust based routing protocol for VANETs, in: Proc. COMSNET, 2018, pp. 1–6.

27.

Kofi Erskine and

K.M.

Elleithy, Secure intelligent vehicular network using fog computing, MDPI Journal of Electronics 8(455) (2019), 1–25.

28.

Krundyshev,

Kalinin and

Zegzhda, Artificial swarm algorithm for VANET protection against routing attacks, in: Proc. IEEE Conference on Industrial Cyber-Physical Systems (ICPS), St. Petersburg, 2018, pp. 795–800.

29.

Lee and

Lee, A study on the location-based VANET routing scheme in urban environments, in: Prof. 3rd IEEE International Conference on Network Infrastructure and Digital Content (IC-NIDC), Beijing, China, 2012, pp. 209–211.

30.

Li and

Boukhatem, An intersection-based delay sensitive routing for VANETs using ACO algorithm, in: Proc. IEEE 23rd International Conference on Computer Communication and Networks (ICCCN), 2014, pp. 4–7.

31.

Liu,

Chen,

Ren,

Qiu and

Yang, A delay-aware and backbone-based geographic routing for urban VANET, in: Proc. IEEE International Conference on Communications (ICC 2018), 2018.

32.

Maowad and

Shaaban, Efficient routing protocol for Vehicular Ad Hoc Networks, in: Proc. 9th IEEE International Conference on Networking, Sensing and Control (ICNSC-12), Beijing, China, 2012, pp. 209–215.

33.

Mohsin Mehdi,

Raza and

Hussain, A game theory based trust model for Vehicular Ad hoc Networks (VANETs), Journal of Computer Networks, Elsevier Publications 121 (2017), 152–172. doi:10.1016/j.comnet.2017.04.024.

34.

Mostafaa,

A.M.

Vegnib,

Bandaranayakea and

D.P.

Agrawala, QoS-aware node selection algorithm for routing protocols in VANETs, in: Proc. Elsevier International Conference on Selected Topics in Mobile and Wireless Networking (MoWNet-2014), 2014, pp. 1–8.

35.

N.J.

Patel and

R.H.

Jhaveri, Trust based approaches for secure routing in VANET: A survey, in: Proc. Elsevier International Conference on Advanced Computing Technologies and Applications (ICACTA-2015), 2015, pp. 592–601.

36.

Ponikvar and

H.J.

Hof, Overview on security approaches in intelligent transportation systems: Searching for hybrid trust establishment solutions for VANETs, in: Proc. Secureware: The 9th International Conference on Emerging Security Information, Systems and Technologies, 2015.

37.

Punitha and

J.M.L.

Manickam, Secure group of authentication technique for VANET, Asian Journal of Information Technology 15(11) (2016), 1637–1644.

38.

Rajesh,

K.R.

Virnal and

G.S.

Raj, An efficient reactive location based Ad Hoc routing protocol for VANETs, in: Proc. 5th IEEE International Conference on Advanced Computing (ICoAC), Chennai, India, 2013, pp. 468–471.

39.

Sadatpour,

Zargari and

Ghanbari, A new cost function for improving anypath routing performance of VANETs in highways, Springer journal on Wireless Networks 25(4) (2019), 1657–1667. doi:10.1007/s11276-017-1620-0.

40.

Saravanan and

Ganeshkumar, Routing using reinforcement learning in vehicular ad hoc networks, Journal of Computational Intelligence, Wiley 36(2) (2020), 682–697. doi:10.1111/coin.12261.

41.

M.J.

Sataraddi and

M.S.

Kakkasageri, BDI agent based dynamic routing scheme for vehicle-to-vehicle communication in VANETs, in: Proc. IEEE International Conference on Smart Technology for Smart Nation (SmartTechCon2017), REVA University, Bangalore, India, 2017, pp. 665–670.

42.

Shafiee and

V.C.M.

Leung, Connectivity-aware minimum-delay geographic routing with vehicle tracking in VANETs, Elsevier Journal Ad Hoc Networks 9(2) (2011), 131–141. doi:10.1016/j.adhoc.2010.06.003.

43.

Shahid Khan,

Balan,

Javed,

Tarmizi and

Abdullah, Secure trust-based blockchain architecture to prevent attacks in VANET, MDPI Journal of Sensors 19(4954) (2019), 1–27.

44.

Shrestha and

S.Y.

Nam, Regional blockchain for vehicular networks to prevent 51% attacks, IEEE Access Journal 7 (2019), 95033–95045. doi:10.1109/ACCESS.2019.2928753.

45.

Sukumaran,

Ramachandran and

Rani, Intersection based traffic aware routing in VANET, in: Proc. IEEE International Conference on Advances in Computing, Communications and Informatics (ICACCI), Mysore, India, 2013, pp. 1414–1417.

46.

I.A.

Sumra,

Ahmad,

Hasbullah,

J-l.

bin and

Manan, Classes of attacks in VANET, in: 2011 Saudi International Electronics, Communications and Photonics Conference (SIECPC), IEEE, Riyadh, 2011, pp. 1–5.

47.

Tang,

Cheng,

Wu,

Wang,

Dai and

Shen, Delay-minimization routing for heterogeneous VANETs with machine learning based mobility prediction, IEEE Transactions on Vehicular Technology 68(4) (2019), 3967–3979. doi:10.1109/TVT.2019.2899627.

48.

Tolba, Trust-based distributed authentication method for collision attack avoidance in VANETs, IEEE Access 1(1) (2018), 99.

49.

Tyag and

Demblai, Performance analysis and implementation of proposed mechanism for detection and prevention of security attacks in routing protocols of vehicular ad-hoc network (VANET), Egyptian Informatics Journal 18 (2017), 133–139. doi:10.1016/j.eij.2016.11.003.

50.

Vidhale and

S.S.

Dorle, Performance analysis of routing protocols in realistic environment for Vehicular Ad Hoc Networks, in: Proc. 21st IEEE International Conference on Systems Engineering (ICSEng), Las Vegas, USA, 2011, pp. 267–272.

51.

Wu,

Ohzahata and

Kato, Flexible, portable, and practicable solution for routing in VANETs: A fuzzy constraint Q-learning approach, IEEE Transactions On Vehicular Technology 62(9) (2013), 4251–4263. doi:10.1109/TVT.2013.2273945.

52.

Wu,

Fang and

Li, Reinforcement learning based mobility adaptive routing for Vehicular Ad-Hoc Networks, Journal of Wireless Personnel Communication, Springer 101 (2018), 2143–2171. doi:10.1007/s11277-018-5809-z.

53.

Zhao,

Li,

Meng,

Gong and

Tang, A SVM based routing scheme in VANETs, in: Proc. IEEE International Symposium on Communications and Information Technologies (ISCIT), Qingdao, 2016, pp. 380–383.

Hybrid routing protocol for VANETs: Delay and trust based approach

Abstract

Keywords

1. Introduction

2. Related works

3.1. Network environment

3.3. Delay and trust based routing scheme

3.5. Example scenario

Table 5 Simulation inputs Sl. No. Input parameters Specification 1 Length of road 5000 mts 2 Breadth of road 100 mts 3 n 20 4 Scenario Highway 5 R 80–100 mts 6 V 80–120 kmph 7 S 2 mts 8 TR t h 0.6 9 Data packet size 2 Kb

4.2. Simulation inputs

4.3. Simulation procedure

4.4. Performance parameters

5. Result analysis

References

Table 5
Simulation inputs

Sl. No. Input parameters Specification

1 Length of road 5000 mts

2 Breadth of road 100 mts

3 n 20

4 Scenario Highway

5 R 80–100 mts

6 V 80–120 kmph

7 S 2 mts

8 ${TR}_{t h}$ 0.6

9 Data packet size 2 Kb