TECHNOVATION A Hybrid Model for Performance Evaluation of Fixed VANETs using Novel 1C3N and Topology-Based Ad-Hoc Routing Protocols with Packet Loss Control Methods

: Vehicular ad-hoc Networks (VANETs) play a significant role in Intelligent Transportation Systems (ITS) Design. Intelligent Transportation Systems are the first mandatory requirements for any smart city. Researchers are vigorously working on ITSs for smart cities and so VANETs have received a lot of attention. In VANETs, Vehicle to Vehicle (V2V) and Vehicle to Infrastructure (V2I) message transmissions are carried out using wireless access technologies like IEEE 802.11p and IEEE 1609 WAVE family of standards. The crucial challenge in the implementation of VANETs involves the task of deciding the routing protocol because unlike MANETs, handover in VANETs is extremely high. In this paper, a novel routing technique, One Caption for 3 Nodes (1C3N) algorithm is proposed. This algorithm is implemented along with other topology-based existing routing protocols for the implementation of VANETs in the Coimbatore-Urban Area (Indian Smart City). The performance evaluation is carried out by comparing metrics like goodput, Overhead, Packet delivery ratio (PDR), Packet loss ratio (PLR) and end-to-end delay for four existing VANET routing protocols. The results show that a proper combination of packet loss model with routing protocol enhances the goodput and reduces the overhead for a fixed VANET. It is observed that the proposed 1C3N routing technique provides 60-65% better goodput than the other four algorithms.


Introduction
India is the first densely populated country in the world with a stretch of 1, 15,435 km roadways.The maintenance of roads is very difficult when the major population of the country uses roadways.Moreover, controlling the mushrooming vehicular population on roads in not an easy task.When the majority of population prefer roadways to reach their destination, it is found that burgeoning cities like Mumbai, Delhi, Chennai and Kolkata could not bungle in providing reliable transportation.This results in various difficulties like traffic jams, violation of traffic rules and some serious and fatal road accidents.It is officially reported that nearly 1.77 lakh deaths per year occur on Indian roads which are far higher than the deaths of Indians in wars.This can be rectified with the implementation of VANETs in all smart cities of India.
Vehicular ad-hoc network is currently the most vibrant research domain, due to developing technologies like driverless vehicles, lane detection, collision detection and various driver safety systems.This paper aims at suggesting the implementation of VANET in Coimbatore city, one of the smart cities in South India.VANETs support V2V and V2I communication with the help of 5.850-5.925GHz frequency band allocated by the federal communication commission for dedicated short range communication.It uses IEEE 809.11p and IEEE 1609 family of standards.
Wheeb et al [1] UAV adhoc networks performance were analyzed with difference protocols.VANET is enhanced with the help of 5G networks to increase the efficiency and stability Naeem et al [2].
The article is designed as follows: section II explains about related work; section III explains about simulation setup, concerned with the loss models, routing and interfacing protocols; section IV dwells on proposed 1C3N Algorithm.;section V focuses on the results analysis; section VI sums up the conclusion; section VII is acknowledgement and section VIII is references.

Related Work
There are various advanced routing protocols such as TLRC and GLSR-L posited by Xia et al [3].AQRV is specially designed for ant colony optimization in VANETs by Guangyu et al [4].There are other routing algorithms that are specially designed for VANETs such as short video distribution in simplex format and 5G assisted content distribution according to Leo et al [5].Intelligent OLSR and situation-aware routing models have been proposed by Toutouh et al and Hasham et al [6,7].But these routing models have not considered the importance of loss models that improvise the performance of VANETs.In this paper the proposed method is designed to use Friis and ITU.1411 loss model by which VANET performance can be observed and improved.GPSR protocol uses greedy forwarding techniques in numerous vehicular nodes.It is mostly restricted and designed to operate in the city environment.VADD, recommended by Zaho et al, is the Vehicle Assisted Data Delivery system that works by means of direction probe D-VADD and Location probe L-VADD [8].Yet the edge selection for transmitting a packet has been found to be difficult in this method.

Int
A-STAR algorithm scheduled by Raissi et al has been initially developed to work in wireless sensor networks [9].Based on its performance, it is currently positioned in VANETs for its ability to find the optimal path for the packet delivery.It uses packet forwarding technique and it is more applicable for city environment.Most packet forwarding techniques are suitable for city environments as their packet delivery ratio will be higher compared to the rural areas.CAR proffered by Neumuo et al., is a connectivity aware routing protocol which is designed for communication between vehicles in highway environment where the mobility is high [10].In this case, a technique 'Guards' is used to track the mobility and applied for real road maps in Switzerland.Jarupan et al put in place PROMPT which is a positionbased or geo-location-based routing convention that uses source routing mechanism [11].
It helps in finding the shortest paths by relying on independent vehicle positions.It is a cross-layer and delay aware protocol.This protocol performs better than GPSR, A-STAR and CAR, since the priority flag in the packet overtakes all these algorithms.The results of GPSR, A-STAR and CAR algorithm is also discussed with OLSR, DSDV, DSR and AODV algorithms [12][13][14][15][16][17].

Simulation set up
Simulation of Urban Mobility (SUMO) is used in our model to generate the realistic urban environment of Coimbatore city.It uses discrete time and continuous space traffic mobility modeling simulator which helps to stimulate different vehicles like cars, buses, trucks, motorcycles and even pedestrians.Traces and routes are carried out with the help of NS-3.29 simulator [18].SUMO helps with the direct import from Open Street Map (OSM).The authors of this paper have used OSM to generate the mobility created by SUMO for a fixed VANET with the help of wireless vehicles and 10 fixed RSU's which are equally distributed.Figure 1a shows mobility model created and Figure 1b shows the traffic congestion.

Friis Loss Model
This model works only in vacuum or free space.The model is defined by where , C=299792458 m/s and f=5.9 Ghz in our model.
L is system loss, for which in many cases it is very less and negligible.

ITU-RP.1411
An International Telecommunication Unit (ITU) has been designed for the protocol for 300 MHz to 100 GHz.This model works with upper and lower bound using the break point distance to decide the bounds.The breakpoint distance is given by (H is height and ƛ is wavelength)

AODV
Ad-hoc On-Demand Distance Vector (AODV) is a routing convention specially designed for MANETs and now it is currently used for VANETs.In fact, AODV supports all ad-hoc networks.For any routing protocol, route discovery and route maintenance are the two major responsibilities.In route discovery, source code does not take full responsibility for identifying the destination because each node has to maintain route cache which contains the information of previous and next hops.Here, the routes are created on-demand and a route discovery packet RREQ (Route Request) is sent from the sender that consists of SID, DID, Recent Sequence number, Broadcast ID, Hop count and Time to live (TTL) [19,20].TTL is also used to keep the control of traffic and security in the network.Once the route is discovered, RREP (Route Reply) is used to establish the path.Since AODV uses frequent update messages to maintain route cache, the memory used is very less and thus it can be implemented in Zigbee protocol and any other low rate protocols.AODV has various implementations such as AODV-UU, AODV-UIUC and Kernal-AODV.

OLSR
Optimized link state routing protocol is a convention that is along with optimization in number and size in the control packets.Less number of control packets results in less traffic and congestion.This is carried out by using multipoint relays (MPR's) which shows a huge difference between OLSR and LSR protocols.MPR is a node which is selected by one-hop neighbor to the source node mentioned in Figure 2. Usually flooding of routing information to the neighbours is used in LSR [21].Instead of this, OLSR uses flooding from one-hop neighbor which reduces the route search time by N x to N x-1 times.Routing information is maintained withperiodic updating of routing table at MPR. MPR will be selected only if the willingness of the node is 1.
Here for node 1, node 3 is one-hop neighbor and vice versa.Similarly for node 2, node 4 is one-hop neighbor and vice versa.

DSR
DSR is a routing convention that is reactive.The proactive routing protocol updates the routing table by flooding up routing information, whereas reactive routing protocol always attempts to identify a new path from the source.Since the routing information is not shared, there is no need for route maintenance as routes are discovered dynamically.A re-request packet sent from the destination node ensures the dedicated route to the source node so that the packet gets ensured delivery.Destination computes the shortest path upon receiving the packets from different routes based on broadcasting done in each node [22][23][24][25][26].Then, a re-request packet is sent through the shortest path to start packet transmission.Generally, the physical layer overhead of DSR will be higher when compared to any other protocols especially with less number of nodes.So, this protocol needs to be handled carefully.

DSDV
Destination Sequenced Distance Vector (DSDV) uses Bellman-Ford routing.For every 5 seconds, a routing table is broadcast from every node to every other node [27].It is usually called as periodic updates.Sequence numbers are used to avoid loopings in the network.The route used for broadcasting is decided based on the latest sequence number.When mobility of the route is high, route fluctuation increases which further increases the congestion and decreases the PDR in the network.

Proposed Model
In the proposed VANET model, the realistic Coimbatore city traffic scenario, involving traffic signals, road segments, highways and other building blocks, is taken into account.This model considers equally distributed vehicles and 10 RSU's in the network.MAC layer interfacing protocols IEEE 1609 WAVE STD and IEEE 802.11p standard are used in this work.As mentioned earlier, network layer protocols AODV, OLSR, DSDV and DSR along with 13CN are applied to the proposed fixed VANET model.Therefore the performance can be improved by means of identifying the best loss, fading and shadowing models for the channel in the implementation of routing protocols.The assumptions and simulation parameters used in the work are given in Table 1.

Why 1C3N upon other algorithms?
One Caption in Three Node (1C3N Algorithm) is designed in such a way that it provokes the routing design of all the four algorithms.Algorithms like AODV, OLSR, DSDV and DSR work efficiently under the foreign conditions, for example, Wi-Fi enabled road structures, well equipped road side units, driver sides, etc.But when implementing this algorithm for the Indian road systems, it gradually reduces the packet delivery rate at very high level (i.e. out of 100 packets, only 10-19 get delivered based on algorithms.It varies in this range and the remaining packets are lost).This heavy packet loss results in some difficulty during implementation, such that each router in the road side unit fails to deliver nearly 80% of its packets and also helps to reduce/avoid accidents by considering the driver's careless mistakes [28][29].It causes network congestion, overload in CRC check at the receiver side and a lot of security and implementation problems.

Design and Working of 1C3N
The 1C3N algorithm enables dynamic to send instant information, Auto-execute when the vehicle is switched on and connected vehicles provides an easy hop routing between the mobile nodes wishing to establish a VANET.1C3N allows every node to obtain routes as quickly as possible for the newest destinations.1C3N allows nodes to respond to links.It also allows the dynamic change in network topology to be updated in a timely manner.There is no separate node is needed to maintain the routes in the table for each and every cluster linked to the network for both active and passive communication.The operation of 1C3N is loop-free.This algorithm is designed in such a way that it avoids the Bellman-Ford, "counting to infinity" problem.An Easy convergence technique is used in this algorithm to adapt itself to the current topology changes (at times when a node moves in the network from one cluster to another).When any node breaks the link and moves out of the network, it notifies every other node in the network so that it can modify the routing table by deleting its entry.For the 2 nodes in the network, there is one captain and this captain node takes responsibility for routing all of its associated nodes (2 nodes).It takes all the computation process of its associated node that is required for routing the packet and reduces the complexity of the network by 2 n-3 /2 n .

Int. Res. J. Multidiscip. Technovation, 5(5) (2023) 20-29 | 24
The overall packet structure like packet design, type, size, routing table parameters, encoding and decoding methods and various other important designs of 1C3N are decided dynamically by the sumo tool in Figure 3.

Results and Analysis
The intention of most researchers focuses on the performance evaluation of the network in two categories: 1) How does the information propagate?2) How much time is taken to spread information to all vehicles in the network?It is also equally important to know about the performance of communication protocols used in the network.In this approach, observational areas are considered separately and implementing new model for improving propagation in VANET and New way of transportation systems were proposed to improve the health systems with the help of drones [30,31].This work evaluates the routing protocols combined with loss models.

Average Goodput (GP)
GP is total end user meaningful bits that are captured by the vehicles in the designed network at the given amount of time.It plays a pivotal role in deciding the network performance.

Average Phy / MAC layer Overhead (OH)
The saturation point of the network is decided by this value.A lower OH indicates that the network has a higher performance.It is given by Avg.Overhead (Bytes) = Total Overhead messages / Total transmitted packets.

Average Packet Delivery Ratio (PDR)
Ratio of total number of packets sent by the start node and the number of packets received by final node.
For a higher PDR, a higher network performance is required.

Average Packet Loss Ratio (PLR)
The percentage of packet loss is, ratio of difference between the total packets sent and received packets to total transmitted packets.
Sometimes, PLR is also called as collision ratio because collision of packets leads to packet loss.In other form, PLR is given by using PDR, PLR (in %) = 100 -PDR.

Average End to End Delay (EED)
The data packets that are successfully transmitted and received are counted and the time taken for every packet to pass through the destination vehicle is recorded.The average EED is given by Avg.EED (ms) = Average time taken of delivered packets / Total number of packets delivered.

Goodput Analysis
For the present scenario, it is found that when the number of vehicle increases, ITU model performs better than the Friis model.It is clearly seen in the graph in Figure 4 a, b, c, d, e, f.For 20-Vehicle input network, OLSR with Friis model performs better than any other combination of routing protocol with loss model used this study.Similarly, for 40-vehicle OLSR with ITU model and 60-vehicle, DSR with ITU model performs better.But deciding the algorithm and loss model for the VANET depends not only on goodput but also on various other parameters.1C3N performs much better than any other algorithms as shown in the graph in Figure 4.1 g and h.
For any loss model, DSR has shown considerable performance in terms of goodput for our VANET and New way of transportation were proposed during the COVID based diseases with safety travel using deep Learning algorithms [32].So, for a static VANET, DSR can be implemented without considering overhead.But for the random structured VANET,

Overhead Analysis
For our scenario, DSR working with source routing, which has the highest overhead compared to any other routing protocols in any type of loss model, increases the load to the physical layer.It is always important to compare the overhead and goodput to analyze the best suitable routes.So based on overhead, DSR can be totally omitted as observed from Table 2. OLSR performs better comparatively than the other protocols for all 20, 40 and 60 vehicle scenario.It is also found that 1C3N performs very well with ITU when the number of vehicles increases.

PDR, PLR and End to End Analysis
PDR and PLR are related to each other and here in terms of PDR, FIIS model gives good results for upto 40 vehicles and for more than 40 vehicles ITU will be implemented.It gives better results especially when OLSR is implemented with ITU as shown in Figure 5 a, b, c, d.For more vehicles, end to end delivery in Friis model is better when compared to ITU.ITU can be implemented for less than 40 vehicles by considering only the end to end delivery time in 1C3N.

Conclusion
The analysis discussed so far is based on the simulation results of the fixed VANET system in using different network sizes and different RPs.In this paper, the authors have presented the performance of the adhoc routing protocols of AODV, OLSR, DSDV, DSR and the proposed 1C3N.The model offers significant insights on how to enhance the VANET performance and how to implement it in intelligent transportation systems.As a result, OLSR with Friis loss model performs very well for the designed scenario of up to 40 vehicles and 1C3N with ITU outperforms with 60+ vehicles.The results corroborate and strengthen the importance of 1C3N and offer instructive guidelines to implement the modern traffic system, followed by safety applications and connection problems in VANETs.

Table 1 .
Assumptions and Simulation Parameters

Table 2 .
Comparative Analysis for Avg Goodput and Avg Overhead

Table 3 .
Analysis of PDR, PLR and E2E delay of traditional and proposed 1C3N