AI-Enabled Ant-Routing Protocol to Secure Communication in Flying Networks

,


Introduction
e most important requirement for any AI-based network should be security. However, there are two types of security threats that are widely used: passive and active. Passive attacks are sometimes used by intruders to get around heavily secured data. Whenever active attacks are utilized, data can be easily accessed and changed. To encrypt plain text, the AES and DES methods use secret keys. Heuristic function is reinforced to use repeatedly arti cial ants for new solutions. Binary ant colony optimization has been used to secure data from security attacks [1]. Hash functions, asymmetric keys, and symmetric keys are three types of cryptography techniques that encode and modify the security of information. Particle swarm optimization (PSO) based key generation algorithm is designed to make use of character code table to secure data, while ant colony optimization (ACO) algorithm produces unique keys and strategies to ensure pro ciency in communication [2]. e topological scenario of wireless sensor network nodes is static. To counter security attacks, routing protocols is the optimal solution which can easily overcome on data transmission vulnerabilities. Intruder makes use of security attacks which include blackhole, grey-hole, and false routing updates; also, hello data packet attacks are utilized to disturb the transmission channels. e mentioned countermeasure can be easily tackled by using appropriate routing protocols. Multihop communication within network bio-inspired technique called ant colony optimization is used to work on path selection and adaptive security nature. In addition, graph theory and arti cial neural networks will be quite helpful in routing techniques for securing communication standards. In addition, routing techniques can be used in smart grids [3,4].
In this research study, routing performed well in comparison with i-ACO and LEACH in terms of end-to-end delay, overhead, and data forwarding [5]. Existing routing protocols have several security aws. is research focuses on finding vulnerabilities in distributed hash table routing approaches. Security risks from the attacker's perspective include destination LID tempering, neighborhood attack, node joining, and authentication. is study has used digital signatures in order to swap logical identifiers and establish a trust between nodes [6]. e network life-time, on the other hand, is essential in any system. is research discusses a unique approach that focuses on energy efficiency, clustering, and fog nodes. e goal of this study is to develop a procedure based on the distance and residual energy associated with the network's cluster head [7]. e notion of AI-assisted routing in flying networks is visualized in Figure 1. Only encryption techniques can ensure the security of communication between aerial vehicles. Wireless communication channels connect unmanned aerial vehicles (UAV) to the base station. e information is stored using AI-based intelligent cloud computing. Using modern routing algorithms, secure communication in any network is achievable. e implementation of any algorithm is the most difficult task due to the dynamic movement of aerial vehicles. However, Table 1 illustrates the security techniques used in the area of aerial ad hoc networks.
AI-based flying network is designed to overcome on the issues related with real world applications of aerial vehicles. e technique of evolutionary computing is introduced in this research study which works on the basic principle of heuristic analysis. As drones need lightweight algorithms to secure communication medium in Internet of flying vehicles, a hybrid cryptographic security approach with high level of computing to ensure encryption using AES-256, ECC, and SHA256 is utilized for authentication [13]. For scaling up the network, convergence is the main issue which affect the energy level of every node. Enhanced pheromone method is considered to increase the life-line and path length [14]. e main contribution of this paper is as follows: (i) AI-based heuristic routing protocol is designed for Internet of flying vehicles (ii) Column mobility pattern is used to check the performance of secure routing schemes in the flying networks (iii) Network throughput, bandwidth utilization, packet drop rate, and quality of experience is used in this research study (iv) Aerial ad hoc networks are utilized in the field of information security and artificial intelligence (v) IoT based flying vehicles are used as a real-life application e organization of the rest of the paper is as follows: Section 2 discusses the literature review, followed by explanations of the proposed scheme in Section 3. en, Section 4 presents the simulation environment. Lastly, Sections 5 and 6 presents results and concludes the paper and suggests future work.

Literature Review
Secure communication scheme in aerial vehicles have to maintain and manage specific area to employ hierarchical identity-based broadcast encryption. Packet transmission using the HIBBE technique is broadcasted as preassigned key to encrypt data. Mutual communication of UAVs uses the signcryption method to verify and authorize every flying node. is technique is utilized for master drones to give resistance towards denial-of-service attacks [15]. On the military side, the employment of aerial vehicles is very popular.
Between the UAVs and the base station, a secure protocol is implemented. BAN-logic, on the other hand, tested that the suggested protocol is valid. e protocol outperforms and responds to denial-of-service attacks, man-in-themiddle attacks, and UAV-to-UAV security [16]. Unmanned aerial systems are either having connection with the ground station and satellite or store the information on memory chips. Key negotiation approach in cryptography is used to develop security protocol which encrypt the data stored in the system. e proposed implementation is installed on the FPGA boards which is utilized in a prototype [17]. e main problem with flying vehicles is network lifetime. To address this, a novel routing protocol is introduced in a dynamic pattern that can save one-third of the energy. Khan et al. worked on the ant colony optimization and incorporated energy stabilizing parameter to conserve energy. is is the first ever routing protocol in FANETs for energy efficiency which is named E-AntHocNet [18]. Due to mobile movement of aerial vehicles, wireless communication technology is used as a backbone. For indoor and openair communication in flying vehicles, path loss model and machine learning classifier called decision tree is used to improve signal power. 3D centroid algorithm is used for node localization to find the actual location and also calculate the estimated location [19].
Routing protocols can be employed in flying networks to reduce end-to-end delays by employing a random way point mobility model that pauses until a path is randomly selected. AnHocNet is employed with other routing techniques which shows better results in comparison with other routing schemes [10]. However, ACO-based routing protocol can be employed in many real-time applications which include searching, monitoring, and rescue operations using aerial vehicles [20]. FANET network is infrastructure-less and nodes move in three dimensions. Software defines networks integration with FANET makes STFANET [21]. UAV-network can also be used in smart grid [22].
For efficient communication, routing protocol plays an important role in health care and sports using Internet of drones [23]. erefore, maintaining the security level of IoTnetworks intrusion detection system filters queue length packets to minimize false alarm and missed detection probabilities [24]. Figure 2 illustrates how data encryption is used in cloud computing to protect information transmitted by UAVs. e diagram also depicts the mechanics of attacks such as denial of service. However, Table 2 describes the detailed survey of security-based routing protocols.
ACO algorithm is a multipath method that can be quite useful in aerial networks for secure routing. A multiobjective function can be utilized to save residual energy while assuring low-cost routing. Ant-based heuristic computation relies on the pheromone update process to improve and solve security attacks regarding networks [34].

Proposed Scheme
AntHocNet is a routing technique for securing communication between nodes in aerial networks.
is method is based on ant colony optimization, which is a hybrid methodology that combines reactive and proactive techniques. e three processes that make up the main working strategy are as follows: (1) Evaporation (2) Concentration (3) Reinforce learning e basic working of AntHocNet is explained in Figure 3, where the process of evaporation and concentration directly connects with pheromone update. If the concentration of pheromone will be high, then solution will be optimal. e term concentration is related with the evaporation of pheromone. e proposed technique easily overcomes coverage, packet drop rate, throughput, and bandwidth utilization. Also, due to quality of service, this routing protocol easily improves the overall life-time of flying network. However, reinforcement learning is used to train the aerial vehicles from environment which gives best results. is novel routing easily improves aerial vehicles performance in many applications which include forest monitoring, surveillance of border areas, filming, and criminal catching. Due to the dynamic pattern of flying vehicles, deployment of every new technique is a very tough task to perform. AntHocNet is compared with other advance techniques like ZRP, M-DART, DSR, and DSDV; in addition, AOMDV is utilized.
Column mobility pattern is used as the mobility model in aerial ad hoc networks for scanning and searching the network environment. Mobile node has a proper reference point which easily moves UAVs in the forward direction. One-time pad (OTP) encryption technique is technique enables a secure link between the drone and the ground control station (GCS), with improved accuracy and execution time in comparison with other systems He et al. [9] Encryption bridging technique between IBE and ABE A new encrypted data switching protocol was introduced that minimizes client computation and communication costs, especially when the client's data are encrypted and outsourced to a faraway cloud Deng et al. [10] An identity-based encryption transformation (IBET) model Using identity-based encryption (IBE) and identity-based broadcast encryption (IBBE) for high-efficiency flexible sharing of encrypted data in the public cloud Puthal et al. [11] Selective encryption (SEEN) method Secure data streams were developed by encrypting data using a symmetric key block cipher with multiple shared keys to protect data confidentiality and integrity from malicious attackers Islam and Shin [12] BUS: a block chain-enabled data acquisition scheme

Simulation Environment
e simulation environment consists of thirty aerial vehicles and one base station which is distributed in ad hoc manner. Network simulator-2 is used to perform the evaluation of routing protocols. Topological scenario contains three-dimensional structure in flying vehicles which include x, y, and z. However, 1000 m is used for each dimension; also, the experimentation is performed in 180 seconds. Figure 4 elaborates the topological structure of aerial vehicles connected with the base station. Deployment of routing techniques in network simulator 2 is a very tough task. Also, the physical structure of the UAV-network is properly presented in Figure 4.

Results and Discussion
AntHocNet is a heuristic computing method that is inspired by ant behavior. When it concerns to evaluating routing algorithms, ant-based routing exhibits the most diversity in terms of throughput. Figure 5 shows how routing techniques monitor network performance to see how many data packets are received per unit time.   Figure 6 illustrates bandwidth utilization to sense the transmission capacity of the flying network. However, AntHocNet shows optimal results in terms of network/ bandwidth in comparison with dynamic source routing and other routing schemes.
Packet drop rate is visualized in Figure 7, where bioinspired scheme called AntHocNet shows less packet drop analysis on 40 to 140 seconds and 160 to 180 seconds. In comparison with other techniques, packet drop is very much reduced.

Conclusion
AI-enabled flying networks are a newly emerged area that provides a better understanding of artificial intelligence and UAVs. ere exist many applications such as rescue operations, disaster, searching, and monitoring. However, routing protocols are quite helpful, especially in optimal and secure communication. In this paper, the ant colony optimization-based routing technique is introduced which is used to secure communication standards within the network. Drone networks are regularly changing physical structure which is quite vulnerable. erefore, a comparative analysis of routing protocols is properly discussed where AntHocNet performed well in terms of packet drop rate, throughput, and bandwidth utilization. Around 93% of packet drop analysis is optimized due to AntHocNet. Also, bandwidth utilization is improved by 90% in comparison with other routing techniques. Moreover, throughput is used to check the number of received data packets with respect to time. erefore, the throughput level is enhanced by more than 95% in comparison with traditional routing techniques. e topology of flying vehicles is dynamic. However, to address this problem, security is a key consideration. Ant behavior routing is based on heuristic computing to secure communication between flying nodes. us, this research proposed a novel routing protocol with a focus on ant behavior routing, which assists in end-to-end security. e proposed algorithm is compared with other legacy protocols which include DSR, AOMDV, M-DART, ZRP, and DSDV. e simulation results illustrated that AntHocNet has shown better performance in accordance with throughput, network utilization, and packet drop analysis. e experience is evaluatedby combining the parameters quality to finalize the aerial network life-line. Column mobility model is deployed in the aerial networks.
In the future, particle swarm-based routing should be introduced to secure communication among aerial vehicles. Also, hybrid mobility models need to be designed for FANETs. In addition, artificial intelligence, machine learning, and computational intelligence will be the optimal solution to improve communication standards in FANETs.

Data Availability
is research study was performed on network simulator-2.
Disclosure is paper has been removed from SSRN.

Conflicts of Interest
e authors declare that they have no conflicts of interest. Applied Computational Intelligence and Soft Computing 7