Bio-Inspired AODV-ACO Routing Protocol for Optimal Routing in VANETs Networks

In this section, a comprehensive review of the literature on ad hoc networks using the AODV protocol with the ACO optimization method is presented, along with the pros and cons.

A new route selection mechanism that combines the AODV reactive routing protocol and the ACO algorithm to optimize QoS in mobile ad-hoc networks (MANETs)was proposed in the study [14]. Inspired by ant colonies, it selects the optimal data delivery route based on the path's pheromone value, calculated using factors like end-to-end reliability, congestion, hop count, and nodes' residual energy. This mechanism aims to identify the most efficient and reliable routes. However, residual energy is an outdated metric in VANETs, where vehicles typically have continuous power supply. Moreover, their protocol does not consider essential metrics such as RSS or end-to-end latency, making it poorly suited to the highly dynamic topology and rapid mobility of VANET environments.

ACO based on a Modified AODV (MAODV) protocol for secure data transmission in MANETs have been studied by Anantapur et al. [15]. MAODV utilizes ACO and considers four fitness functions: residual energy, distance, trust and node degree to detect optimal paths while mitigating blackhole attacks. While the integration of trust metrics enhances security, the approach primarily targets security at the expense of QoS. Additionally, the use of trust indicators and node degree significantly increases computational complexity, which is not ideal for embedded VANET systems requiring lightweight protocols.

Sindhwani et al. [16] proposed a novel approach that integrates the AODV routing protocol with the ACO technique for efficient data transmission in MANETs. In this research, a multicast-based path establishment strategy is introduced from source to destination, aiming to reduce congestion and optimize bandwidth utilization to improve the performance. However, this multicast-based strategy can lead to excessive routing overhead in high-density VANET environments, and the criteria used (e.g., hop count) remain conventional without considering radio link quality or physical-layer dynamics.

A new hybrid approach that modifies the AODV routing protocol by combining clustering K-Means methods with ACO to discover optimal routes in (MANETs) was presented in study [17]. The integration of these two techniques aims to leverage their respective strengths, enabling efficient route discovery while mitigating common MANET challenges like network dynamics, resource constraints, and security threats. Their approach's incorporation of K-Means clustering may increase computational complexity and processing time, which may be impractical for real-time embedded VANET units like OBUs.

Nemade and Pujeri [18] proposed a novel Emergency Data Transmission using ACO (EDTA) protocol. EDTA enables ambulances to utilize any available route, prioritizing minimal communication overhead, reduced delays, and maximized throughput for prompter patient care. It introduces an ACO fitness function focusing on current vehicle speed and data/network congestion levels to optimize emergency data transmission in vehicular networks. This approach is limited to emergency situations, however, and does not apply to other VANET services such as safety alerts or infotainment. In addition, important parameters such as signal strength and bandwidth are not taken into account in the decision-making process, which can limit the stability and efficiency of selected routes.

Simaremare et al. [19] enhanced the performance of their suggested secure protocol using an ant algorithm. Ant agents deposit positive pheromone on trusted nodes, enabling path selection based on pheromone values. In their approach they evaluate and compare the protocol's performance, with and without the ant algorithm, under DoS/DDoS attacks. The integration of the ant algorithm aims to enhance security and reliability by prioritizing trusted routes. It focuses exclusively on trust without considering QoS improvements.

Lee et al. [20] proposed an intelligent routing protocol based on the ACO algorithm, titled Multiple-Junction-Based Traffic-Aware Routing (MJTAR), specifically designed for urban vehicular networks. Their approach dynamically selects road junctions based on real-time traffic conditions, aiming to reduce congestion and improve routing efficiency. While MJTAR offers localized traffic-aware optimization, it presents notable limitations. The protocol does not rely on AODV and does not incorporate critical link-quality metrics such as RSS, available bandwidth, or end-to-end latency, which are essential in highly dynamic VANET environments. Moreover, by focusing exclusively on junction-level decision-making, the approach overlooks full end-to-end route optimization. In contrast, our AODV-ACO approach combines the reactive nature of AODV with the adaptability of ACO, integrating physical-layer metrics directly into the pheromone computation. This enables a more robust and QoS-oriented routing solution, suitable for a wide range of VANET applications. The summary of related work is shown in Table 1.

6.1 Simulation environment

In our approach, we simulated urban VANET scenarios from the downtown of Malaga, Spain [21, 22]. The NS-2 network simulator (version 2.35) was used for the various simulations, where the IEEE 802.11ext protocol was used for the physical layer [23]. The wireless channel was modeled with a Nakagami radio propagation model. At the network layer, we used the AODV routing protocol. Finally, TCP was used for the transport layer. The simulated geographic area covered an area of 2 km × 1 km, with a number of vehicles varying from 10 to 45. Table 2 summarizes simulation parameters.

Table 2. Parameters used in simulation

6.2.1 Packet delivery ratio

Packet delivery ratio measures the number of data packets that reach the destination compared to the number of packets originally sent by the source. More specifically, it's the ratio of packets received by the destination node to those emitted by the sending node station [24, 25]. This PDR is calculated by Eq. (5).

$\mathrm{PDR}=\frac{\sum_{\mathrm{i}=0}^{\mathrm{n}} \text { Packet received }}{\sum_{\mathrm{i}=0}^{\mathrm{n}} \text { Packet sent }} \times 100 \%$            (5)

6.2.2 Delay(D)

This measurement represents the average time taken for a data packet to travel from the transmitting vehicle to the receiving vehicle [25, 26]. This average is expressed in Eq. (6).

$\begin{array}{r}\text { Delay }=\frac{1}{n}\left(\sum_{i=0}^n \text { Packets received time }(\mathrm{n})\right. - \text { Packets sent time }(\mathrm{n}))\end{array}$        (6)

6.2.3 NRL

The NRL index is obtained by dividing the total number of control packets sent by each node by the number of data packets received at the destinations [27]. The NRL is given by Eq. (7).

$\begin{aligned} & N R L=\frac{\sum_{i=0}^n Routing \, control \, packets\, transmitted}{\sum_{i=0}^n Data \, packets \, received}\end{aligned}$          (7)

6.3 Results and discussion

This section compares the performance of the new proposed AODV-ACO protocol with that of the existing AODV protocol. Three criteria are evaluated: packet delivery rate, transmission delay and routing overhead. Results are shown as the mean values over 5 simulation runs with different random seeds.

Table 3. Performance analysis of PDR

Figure 7. Packet delivery ratio

Figure 7 and Table 3 show the evolution of PDR for the two protocols, AODV-ACO and original AODV, depending on number of vehicles, the results show a better PDR of our AODV-ACO approach compared to the original protocol.

To achieve a high packet delivery ratio, packet losses during transmission must be minimized, mainly caused by route breakages. In vehicular ad-hoc networks (VANETs), route breakages during transmission pose a significant challenge to maintaining efficient communication. When a route fails, packets in transit are temporarily stored in the queues of intermediate nodes while the network attempts to establish an alternative path. However, since node queues have limited capacity, if a new path is not established quickly, queues fill up, and packets are dropped. This leads to additional losses and degrades the routing protocol's performance.

The proposed routing approach improves VANET network performance by integrating signal quality and bandwidth into its fitness function. This strategy prioritizes paths with robust links and sufficient capacity, effectively excluding connections between vehicles with low signal strength from the routing process. By favoring high-quality links and taking bandwidth capacities into account, the algorithm significantly reduces the probability of packet loss due to link instability or insufficient capacity. This refined selection method results in more reliable data transmission paths, minimizing route breaks and reducing the need for frequent route re-establishment. As a result, fewer packets are dropped due to queue overflows or transmission failures, leading to a significant improvement in packet delivery rates.

Figure 8 and Table 4 illustrate the impact of increasing the number of vehicles on Delay for proposed protocol AODV-ACO and original AODV; we can see that the end-to-end delay of AODV-ACO is better than AODV original.

In AODV-ACO, the selection of routing paths is not based only on the number of hops. It takes into account various other critical factors such as the available bandwidth, the RSS, and the total latency from the source to the destination. This multi-objective approach allows AODV-ACO to steer clear of links with weak signal strengths, which are susceptible to instability and congestion. By avoiding such low-quality links, the protocol minimizes the necessity for packet retransmissions and reduces waiting times caused by queuing delays. As a result, AODV-ACO can deliver improved overall network performance and achieve lower end-to-end delays.

Table 4. Performance analysis of delay

Figure 8. Delay (ms)

The cumulative effect of these improvements is better overall network performance. Specifically, the AODV-ACO protocol reduces end-to-end delays, as packets take more reliable and efficient paths through the network. This reduction in latency, combined with the protocol's ability to maintain stable routes, contributes to the improved QoS and robustness of the VANET communication system.

Figure 9 and Table 5 represent the NRL of AODV-ACO and AODV original. The graphical representation clearly demonstrates the reduced routing load associated with the AODV-ACO method when compared to the original AODV protocol. The latter suffers from elevated routing load levels due to the transmission of a significant volume of control packets during the process of routing path establishment. In contrast, the AODV-ACO method does not contribute to additional control packet transmissions, as the pheromone level is incorporated into the fitness function calculation. This pheromone-based approach not only mitigates the initial routing load but also facilitates continuous adaptation to the frequent topology changes in the network through periodic pheromone evaporation. This evaporation mechanism prevents excessive routing load accumulation while dynamically maintaining optimal paths. As a result, AODV-ACO efficiently minimizes the control routing load while seamlessly adapting to the dynamic characteristics of the urban environment.