Secure bio-inspired optimization with intrusion aware on-demand routing in MANETs

Abstract

Mobile Ad-Hoc Networks (MANETs) are specialized networks that operate in a decentralized manner. As it has no centralized control over communication, the overall network is generally vulnerable to various security threats such as Blackhole, Wormhole, Flooding and Unauthorized access. Although several studies have analysed these issues, ensuring security in MANETs remains a significant challenge. Previous research has achieved security at the cost of significant energy and resource consumption. To resolve this issue, we proposed a novel Secure Bio-Inspired Optimization (Sec-BIO) methodology for routing over MANET. Only when the source node expresses a wish to do so, it builds route. Ad-hoc On-Demand Distance Vector Routing (AODV) is one of the applications that fall under the category of on-demand routing protocols. The proposed methodology is built upon the conventional AODV protocol which uses bio-inspired Elephant Herding Optimization algorithm. Here, Trust Value (TV) which is the important security measure is considered as the fundamental parameter of route selection. In addition, energy efficacy also plays vital role while selecting optimum path. Before that, the nodes which involve in the network are validated with the Digital Signature Algorithm based authentication procedure (DSA-Auth). It assures that no other unauthorized nodes are allowed in the communication. Next, the data is sent over the trusted secure route selected by proposed approach. To ensure high level security, we intend to use Elliptic Curve Cryptography (ECC) which encrypts the data sent over the network. Here, the normal and intrusion packets are classified at receiver end by means of K-means Clustering approach. In general, the proposed Sec-BIO methodology demonstrates superior performance, achieving a Packet Delivery Ratio (PDR) of 90% at 100 nodes, reducing end-to-end delay to 90ms, and lowering energy consumption to 1390 J compared to existing methods.

Introduction

A mobile Ad-hoc Network, also known as a MANET, is made up of a collection of mobile hosts that are capable of performing core networking functions like data communication without support of any central access^1,2. As a result of the limited reach of each device that is advertised to the host, the nodes that make up an ad-hoc network are dependent on a number of intermediary nodes in order to convey a data packet to its intended destination. When it comes to core network features like routing packets and forwarding, security in MANET is an essential component. During the early phases of the planning process for the network functions, unless the messages are intrinsic to the network’s capabilities, then these network functions may be put in a convenient position of risk. In a routing network, dedicated nodes are used to offer assistance for critical services such as network management, cluster organization and wireless communication. However, in ad-hoc networks, all of the available nodes will be responsible for performing these duties. Regarding the fundamental foundation of the safety vulnerabilities that are specific to Manets, this is an incredibly difficult challenge. It is no longer possible to rely on the nodules of an ad hoc network to ensure that critical networking characteristics are carried out in an appropriate manner. This is in contrast to the dedicated nodules that are present in a conventional network^3,4,5 Wireless networks have a significant need for security measures to be implemented. The proposed wireless routing protocol network’s adaptable performance was focused on eliminating the threat posed by the cancerous agent.

Regrettably, none of the ad hoc routing protocols that are frequently used contain any security concerns, and they rely on all of the participants to appropriately relay routing and data traffic. In the case of an ad hoc network that is dependent on intermediary nodes for the purpose of packet forwarding, this assumption may out to be very detrimental. In earlier surveys and review publications that presented comparisons of ad hoc routing protocols, security issues were largely disregarded^6,7. The topic of secure and resilient routing in mobile ad hoc networks is discussed in this article, which provides a summary of the methods that have been proposed to handle the issue. In a security system, vulnerability is a flaw that may occur. There is a possibility that a certain system might be susceptible to unauthorized data modification due to the fact that the system does not authenticate the identity of a user before granting access to data. Compared to a wired network, a MANET is easier to be compromised by the intruders. These vulnerabilities have the potential to constitute difficulties and problems for the security of MANETs.

Putting a halt to the transmission of data packets is the most common kind of assault that occurs in MANET^8,9. In the event that we take into consideration a malicious node that persistently waits for its neighbour node to begin an RREQ packet. On the moment when a node gets the RREQ packet, it will immediately transmit a fake RREP packet that has a high sequence number that has been updated. This is done in order to give the source node the impression that there is a new path that may be used to reach the destination. The RREP packet that is sent from the other nodes, including the right nodes, is ignored by the source node. Instead, it immediately rejects the other nodes and begins transmitting packets in the direction of the malicious nodes. After then, the malicious node will seize all of the routes that go to itself, and it will not let the packets to be forwarded to any other destination^{10,11,12,13,14}. In order to solve these assaults, we use detection methods. This kind of attack will occur often, which makes it difficult to discover. This kind of assault is referred to as a black hole attack since it consumes all of the data. With the assistance of network simulator (NS-2), the Blackhole attack, which is one of the potential assaults, was simulated in the study. This attack is one of the conceivable attacks on the protocol at hand. The simulation results demonstrate the packet loss, throughput, and end-to-end latency on AODV in MANET in both presence and absence of Blackhole. The analysis revealed that the network with a Blackhole node has more amount of packet loss.

There is a plethora of routing protocols available currently; unfortunately, many of these protocols lack security and are therefore vulnerable to attacks^15,16. Because of this, the network becomes vulnerable, which can drastically reduce the system’s performance. In mobile ad hoc networks, any one of the many possible types of attacks might potentially affect any given node. MANETs face several security risks such as blackhole, wormhole, Sybil, flooding and grayhole attacks which can damage network routing processes and data transmission pathways. The decentralized and dynamic nature of MANET causes severe impact on network performance as these attacks makes intrusion-aware and trust-based secure routing crucial for research^17,18,19,20.

MANET security faces essential challenges with intrusion-aware routing acting as one of the main issues. Modern routing protocols known as Ad-Hoc On-Demand Distance Vector (AODV) function by using a dynamic route-building system based on network usage requirements. All nodes receive automatic trust from these protocols because of which network disturbances and false routing data occurs, causing denial-of-service attacks. Security mechanisms in MANETs typically generate large overhead on network resources and energy consumption that creates conflicts between security levels and network performance. Security methods based on cryptography create additional processing requirements at the same time, protection against attacks through continuous monitoring consumes more battery power. Therefore, a challenge exists to achieve secure and energy-efficient communication systems for MANETs.

Research gaps and motivation

The security of MANETs has received research attention due to the implementation of encryption methods availability of distributed anomaly detection systems and trust-based routing protocols. The present security methods face several drawbacks which include:

• High energy overhead: The energy requirements of security mechanisms which depend on sophisticated cryptographic procedures or continuous network monitoring cause rapid reduction in network operational time.

• Vulnerability to routing attacks: Traditional MANET routing protocols show poor trust-based decision capabilities because of which they remain susceptible to blackhole, wormhole and Sybil attacks.

• Lack of intrusion detection at the receiver end: The receiver-side implementation of intrusion detection systems uses inadequate methodology to differentiate between regular and harmful network packets.

• Limited optimization in secure routing: The current security-enhanced versions of AODV need better optimization of secure routing because they lack trust-based route selection along with energy-efficient performance under congestion conditions.

Main contributions of the work

The paper presents Sec-BIO as a security framework which integrates biological optimization approaches with cryptographic authentication and AI-based packet identification to protect MANETs against routing attacks. The main contributions of this study include:

• Trust-based secure routing using elephant herding optimization (EHO): Sec-BIO selects routes through Trust-Based Secure Routing using Elephant Herding Optimization (EHO) by analysing trust values along with residual energy and congestion information to let only dependable nodes to handle the routing operations.

• Lightweight authentication via elliptic curve cryptography (ECC): ECC-based digital signatures found in Sec-BIO offer strong protection along with minimal performance impact because they avoid the heavy computational requirements of traditional encryption approaches.

• Receiver-end intrusion detection using K-means clustering: The proposed model implements K-Means Clustering analysis at the receiver-end for intrusion detection which enables real-time packets classification between normal and malicious traffic based on transmission characteristics.

• Adaptive security response mechanism: The system employs adaptive security mechanisms through Sec-BIO which implements real-time adjustments of trust scores and routes while separating malicious nodes by automatic reconfiguration in order to sustain secure communication.

• Extensive performance evaluation: Testing through extensive performance evaluation shows that the model performs better than current security mechanisms like AODV, E-AODV, SVM-based IDS in terms of security accuracy and efficiency along with attack resistance during network simulations.

As the secure routing protocol requires authentication, it is necessary to first establish a suitable authentication measure by including digital signatures from all the hubs. In addition to this, the data included inside the control packets must be modified. It is also typically accompanied by the use of hash algorithms that only work in one direction. In order to prevent this kind of attack, routing is one of the apt solutions. In addition to this, crypto systems are also playing a pivotal role in making the network more secure^21,22. In particular, encryption algorithms and hashing functions are widely used to secure the network transmissions. Although the working of both functions differs, both rely on agreed or shared public/private keys which can be generated by trusted authority. Many algorithms are adapted in network security and showed better security levels in this regard.

The rest of the paper is organized as follows: “Related works” presents a literature review of existing security approaches in MANETs. “Methodology” details the proposed Sec-BIO methodology, including node authentication, intrusion-aware routing, and encryption mechanisms. “Results and discussions” discusses experimental results comparing Sec-BIO with conventional methods. Finally, “Conclusion and future work” concludes the paper and outlines future directions for research.

Related works

Security challenges in MANETs

Mobile Ad-Hoc Networks (MANETs) perform well in disaster recovery, military operations and emergency response as well as in vehicular networks as they operate without infrastructure and automatically adapt their configuration upon requirements. Internet connectivity through Wi-Fi transmitter and receiver capabilities is standard in every mobile node operating in a MANET²³. Steps taken to safeguard open wireless networks against blackhole, wormhole, Sybil and grayhole attacks become necessary due to MANETs’ dynamic topology, decentralized network control, and exposed wireless connection point. Researchers have developed four strategies to fight MANET security risks including cryptographic encryption, trust-based routing and clustering techniques and also AI-driven intrusion detection. Current security approaches in MANET create high CPU demands and elevated energy usage and detection errors, so developers need to build an advanced yet simpler security solution.

Cryptographic security approaches

Data confidentiality and authentication are efficiently managed in MANETs through cryptographic methods. EAACK implements RSA encryption with P2P (Peer-to-Peer) acknowledgment protocols for enhanced security functions to authenticate nodes and stop attacks that causes packet loss²⁴. EAACK implements RSA-based encryption to boost security levels as it optimizes EIGRP hybrid protocol routing methods. EAACK establishes superior packet delivery performance accompanied by lower routing costs and speedier network transmission than TWOACK and Watchdog traditional acknowledgment procedures. The implementation of RSA encryption produces considerable amounts of computation that reduces power efficiency for limited-resource MANET nodes. The Adaptive Coronavirus Herd Immunity Optimizer (ACHIO) serves as a network security solution for MANET-IoT by implementing chaotic map encryption and decryption²⁵. The cryptographic method delivers 86.02% average accuracy but tackles larger text inputs at increased computational time that extends processing delays. The benefits from cryptographic methods for data integrity and confidentiality fade since they neither support insider threat detection nor it suits battery-powered MANET devices because of its increased computational requirements and energy costs.

Trust-based secure routing

The trust-based routing system safeguards the data transmission by allowing only trusted reliable nodes to take part thus protecting from attacks like blackhole and Sybil attacks. GWOA stands as the proposed algorithm for trust-based secure routing within MANETs. The routing system uses this approach to achieve the best possible results through a combination of node-to-node distance and trust value calculation. Secure routing is achieved through a k-disjoint path selection approach together with trust-based metric evaluation for evaluation and selection of the best routing paths. The effectiveness of GWOA in reducing costs and increasing delivery speeds of packets is diminished by its absence of encryption systems that expose network traffic to data capture attacks. The Distributed Clustering Algorithm Dependent Intrusion Detection System (DCAIDS) has been developed as a security-enhancing tool for MANET through its implementation of clustering procedures to discover malicious nodes²⁶. The model operates through clustering nodes into groups so that Cluster Heads (CHs) direct routing decisions and provide security management. The addition of an acknowledgment system in DCAIDS strengthens system delay tolerance while stopping unauthorized users from entering the network. The ongoing node behavior observation under this model results in increased energy consumption.

Military and tactical MANET applications utilize the Optimized Link State Routing (OLSR) protocol because of its successful handling of dynamic topologies¹. Proactive routing updates of OLSR ensure rapid telecommunications within mobile scenarios. Security together with QoS along with OSPF connectivity face substantial challenges in this system. Research using an 18-node testbed demonstrated that radio networks together with dynamic routing updates enhanced network responsiveness yet they also created routing performance problems and exposed systems to falsified attacks.

AI and optimization-based security solutions

The security applications of dynamic network conditions have been widely adopted in MANET through machine learning in conjunction with bio-inspired optimization techniques. RAHHO stands for Refined Adaptive Harris Hawks Optimization algorithm that adjusts security parameters by monitoring current network status regularly²⁷. RAHHO uses Harris Hawks Optimization (HHO) to enhance attack identification together with decreased network exposure. The simulated approach leads to security upgrades alongside remarkable enhancements in network operational metrics through decreased cost function overhead. The security method processes information at slow speeds which makes implementing it in real-time MANETs as challenging. AANNS (Authentication-based Associate Neighbor Node Selection) functions to detect and remove malicious nodes through its signal strength-based trust metric system²⁸. Node reliability in the proposed model gets better because it selects nodes based on their strong signal integrity to deliver secure communication. Signal-based trust assessment provides ineffective scalability when networks experience both large size and frequent topology changes. The Active Routing Authentication System (AAS) stands as a proposed resilient security solution for MANETs because it enhances defense against route spoofing and selective forwarding attacks²⁹. The Active Routing Authentication System (AAS) expanded data delivery rates by 33.9% and kept network connectivity 1.6× better than Cap-OLSR protocol operations. The implementation of authentication creates delays which need improvement for delivering real-time applications. Table 1 shows the comparison of existing security approaches in MANETs.

This research work is based on development of a multi-algorithm system which amalgamates nature-based optimization methods with bio-algorithms to enhance scheduling efficiency and for lower energy expenditures making it suitable for Underwater Acoustic Sensor Networks (UASNs). The research intended five hybridized algorithms with scheduling algorithms that delivers minimal energy costs while handling unscheduled nodes in communication networks. The optimization process incorporates hybridization techniques which includes Elephant Herding Optimization (EHO) with Genetic Algorithm (GA) together with Firefly Algorithm (FA) and Levy Firefly Algorithm (LFA) and Bacterial Foraging Algorithm (BFA) and Binary Particle Swarm Optimization (BPSO)³⁴. The implementation of optimization techniques required the Scheduled Routing Algorithm for Localization (SRAL) which focused on scheduling nodes together with localization optimization. The framework acted as the main driver to optimize Route REQuest (RREQ) and Routing Overhead (RO) while simultaneously reducing Average End-to-End delays, localization errors and improving data delivery. Node localization problems in combination with beacon-level RREQ reconstruction, higher Routing Overhead as well as End-to-End delays and unreliable data forwarding degrades the overall underwater network performance. A new proposed framework along with combined metaheuristic algorithms proved successful at enhancing both node localization techniques and scheduling optimization, energy savings and data delivery reliability in Internet of Underwater Things (IoUT)-based systems.

Table 1 Comparison of existing security approaches in MANETs.

Ad hoc networks required recovery solutions for self-healing processes in their sensor nodes. Such networks demanded new efficient routing protocols as they needed to fulfill their requirements. The mobile ad hoc network currently employs two routing protocols known as AODV and DSR. The investigation of routing issues under different parameters for Packet Delivery Ratio in real-time environments proved to be a challenging task for researchers. The success rate of routing protocols directly influences their overall efficiency based on the number of operational packets being delivered to their destinations. The presented study examined PDR functionality between the key protocols TCP and UDP through result-driven analysis. The team utilized Network Simulator (NS-2.35) as the evaluation tool³⁵. With TCP operation AODV achieved lower PDR effectiveness than DSR. DSR demonstrated an exceptional performance level in PDR measurement when the network environments used TCP or UDP protocols. Data transmission reliability operated at superior levels according to the analysis when DSR was used instead of AODV. This research offered significant knowledge to understand the selection process of suitable routing protocols when considering different network conditions and traffic types. The evaluation results of ad hoc network routing performance better showcased why researchers must choose enhanced protocols for wireless communication networks that experience dynamic changes.

The high energy consumption rates among sink neighboring nodes made the hotspot an essential point of concern for IoT-Enabled Wireless Sensor Networks (I-WSN). The address of preceding nodes got added successively to data packet headers when transmission moved from one intermediary node to other. Each increment of hop count expanded the data size which depleted node energy during transmission specifically within IoT device networks. The process of excessive data transmission resulted in sinking the neighboring nodes by using up their energy at a fast pace which caused premature failures. Node premature death created major network performance issues that led to unnecessary delays along with broken connections and poor packet delivery ratio³⁶. The whole communication network became isolated because of destination node failures. Researchers evaluated the hotspot phenomenon through performance tests performed before and after its manifestation. I-WSN experienced higher hotspot severity under the TCP protocol than UDP conferring to the simulation results. Energy depletion experiences of sink neighboring nodes drives the requirement for better energy management methods which enhance stable networks while reducing delay times and improving data transmission reliability across IoT-enabled wireless sensor networks.

The deployment speed of mobile ad hoc networks (MANETs) have increased because nodes used wireless technology to operate as both routers and researching entities along communication paths. Current security solutions for MANETs were insufficient to protect network systems which rendered their deployment in real environments impossible. The proposed solution addresses these difficulties through the development of Blockchain-Based Trusted Distributed Routing Scheme for MANET using Latent Encoder Coupled Generative Adversarial Network Optimized with Binary Emperor Penguin Optimizer (LEGAN-BEPO-BCMANET). The scheme established trusted decentralized routing through authenticated blockchain token transactions by deploying blockchain technology for proof-of-reputation system implementation. The LEGAN-optimized through BEPO, enhanced both security measures and routing efficiency³⁷. The key aspects include defending against double-spending while ensuring transaction uniqueness and route information reliability as well as protecting the system from self-alterations. Certain metrics were logged during the software implementation of NS3 since this method adopts those metrics. The implemented LEGAN-BEPO-BCMANET framework outperformed BATMAN-MANET and other MANET systems by delivering enhanced throughput at 31.02% while maintaining constant security standards. This makes the solution a secure choice for MANET network implementation.

Self-forming wireless networks that use multi-hops for communication define Mobile Ad hoc Networking (MANET). The specific characteristics of MANET produced an extremely exposed setting to different security threats and attacks. The cuckoo search algorithm helped to identify optimal routing hops to establish secure and efficient navigation through MANETs by using confidence-protected and energy-efficient methods. Trust-based protocols enabled secure data transmission through an energy-efficient and effective mechanism. The introduced method produced substantial improvements in routing security performance³⁸. The approach demonstrated minimal energy drain of 0.10 mJ along with 0.0035 ms delay, 0.70 bps traffic and 83% detection accuracy. Research findings proved that the combination of trust-based navigation with the cuckoo search algorithm effectively strengthens both MANET security and operational efficiency. The proposed solution made routing decisions more effective for improved wireless network resource management and communication reliability. The application of this security method provided strong protection against security breaches in addition to maintaining quick response times and accurate threat detection for secure MANET operations. The investigation provided substantial progress toward resolving mobile ad hoc networks security weaknesses and energy usage issues.

Research gaps and motivation for Sec-BIO

The MANET security field faces continuing challenges even after development of cryptographic, trust-based AI security systems.

High computational overhead: The encryption methods involving RSA and blockchain security introduce substantial computational complexity that reduces overall network performance efficiency.

Lack of receiver-end intrusion detection: Trust-based routing protocols fail to detect malicious activity at receiver nodes since end security check at receiver remain unattended.

Suboptimal routing decisions: The current routing systems use fixed routing parameters which do not dynamically adjust path selection criteria in accordance with energy performance, network load levels together with trustworthiness considerations.

To overcome these limitations, this study proposes Secure Bio-Inspired Optimization (Sec-BIO), a hybrid security model integrating trust-based routing, lightweight encryption, and real-time intrusion detection. Sec-BIO employs Elephant Herding Optimization (EHO) for optimized route selection based on trust, energy efficiency, and congestion levels, addressing routing inefficiencies in OLSR and GWOA. It utilizes Elliptic Curve Cryptography (ECC) for lightweight encryption, reducing computational overhead compared to RSA-based EAACK. Unlike AANNS, Sec-BIO enhances intrusion detection with K-means clustering at the receiver-end, ensuring real-time attack classification. Additionally, it features adaptive security mechanisms, unlike DCAIDS, dynamically adjusting security strategies to changing network conditions. Lastly, Digital Signature Algorithm (DSA) is integrated for faster authentication, minimizing delays seen in AAS. By combining these techniques, Sec-BIO ensures efficient, secure, and scalable MANET communication while reducing energy consumption and latency.

Methodology

This section broadly describes the proposed Sec-BIO routing optimization for the MANETs. The considered network is constructed with $\:n$ number of mobile nodes. The network structured as follows,

• Mobile nodes $\:N=[{N}_{1},{N}_{2},{N}_{3},\dots\:,{N}_{n}]$

• Certificate authority $\:CA$ (It issues certificate to all nodes in the network and it is assumed to be trusted)

• Malicious nodes $\:M=[{M}_{1},{M}_{2},\dots\:.,{M}_{m}]$

• Suspicious nodes $\:S=\left[{S}_{1},{S}_{2},\dots\:,{S}_{s}\right]$ (These nodes behave as malicious but they can be either malicious or normal node that can must be detected)

• The network area is defined as $\:a\times\:b$m

• All nodes are allowed to move around the network.

• The nodes move with varying speed $\:s$ with the maximum speed $\:{S}_{max}$

The overall network setup is illustrated in Fig. 1.

Fig. 1

Proposed Secure MANET.

Algorithm 1

Secure bio-inspired optimization (Sec-BIO)

Node authentication

Authentication is the very first step to validate the trust level and behavior of the nodes that participate in the network. This process eliminates mainly unauthorized nodes in the network. A Digital Signature Algorithm (DSA) for authentication (DSA-Auth) is proposed in this research work. In general, DSA is used to identify the authenticity as well as integrity of the data sent by the sender or source node. At first, all nodes are registered with the CA which is a trusted entity and get the certificates. The data and secret key (in the certificate) are used to create a cryptographic value that can be determined only by the source node whose signature matches with the digital signature. The block diagram of proposed DSA is given in Fig. 2.

Fig. 2

Block diagram of DSA-Auth methodology.

First and foremost, in the realm of cryptography, any individual who adopts this system has a public-private key pair. For each and every signature, the key pairs that are used for encryption or decryption, as well as signing or confirming, are distinct from one another. Within the context of this technique, the private key that is used for signing is referred to as the signature key, while the public key is referred to as the verification key. After that, the node that signed the message sends the data to the hash function, which then creates a hash of the data associated with the message. The hash value and signature key are then input into the signature algorithm, which generates the digital signature based on the hash of the message that has been provided. The data is then added with this signature, and both of them are then submitted to the verifier in order to ensure the safety of that communication. After that, the verifier will input the digital signature as well as the verification key into the verification algorithm that is included inside this DSA. The verifier which is responsible for validation has the procedure for signature verification. By this procedure the hash value is generated. When the data packet is received, then the destination node generates hash value for that certificate too. If both hash values are same, then the verification procedure is successful and the node is normal. Otherwise the node is considered to be malicious. Hence, the digital signature is generated using the signer’s ‘private’ key, which ensures that the data remains secure and cannot be accessed by anyone else. Additionally, in order to ensure the utmost security of the data via cryptography, the signer cannot later deny signing it.

A three-step procedure is included in the suggested authentication mechanism, which is organized as follows:

Step 1: Node Authentication using Digital Signature Algorithm (DSA).

All nodes must undergo DSA-based authentication before starting data transmission in order to prevent unauthorized routing participation.

• 1. Key Generation: Each node $\:n$ generates a public-private key pair $\:\left(P{K}_{n},S{K}_{n}\right)$.

• 2. Message Signing: The source node $\:S$ digitally signs the packet $\:P$ using its private key $\:S{K}_{s}$:

$$\:\sigma\:=Sign\left(S{K}_{S},P\right)$$

(1)

• 3. Signature Attachment: The signed packet $\:\left(P,\sigma\:\right)$ is then broadcast to neighbouring nodes for further routing.

Step 2: Signature Verification at Intermediate Nodes.

All intermediate nodes verify signatures prior to packet forwarding for authenticating participating nodes.

• 1. Upon receiving the packet $\:\left(P,\sigma\:\right)$, each node extracts $\:\sigma\:.$

• 2. Verification using Public Key: Each node verifies $\:\sigma\:$ using the source node’s public key $\:P{K}_{S}$:

$$\:Verify\left(P{K}_{S},\sigma\:,P\right)$$

(2)

• 3. Decision making

  • If $\:\sigma\:$ is valid, the node forwards $\:P$ to the next hop.

  • If $\:\sigma\:$ is invalid, the node discards $\:P$ and flags the sender as untrusted.

The process serves to allow only valid nodes to take part in routing operations.

Step 3: Intrusion-Aware On-Demand Route Selection using Elephant Herding Optimization (EHO).

The selection of an optimal secure path takes place following authentication through a process which combines trust, residual energy and congestion factors using Elephant Herding Optimization (EHO).

• 1. Initialization: Define the set of available paths $\:{R}_{i}$ from source $\:S$ to destination $\:D$.

• 2. Fitness Calculation for Each Route $\:{\varvec{R}}_{\varvec{i}}$:

$$\:F\left({R}_{i}\right)=\alpha\:{T}_{avg}+\beta\:{E}_{avg}-\gamma\:{C}_{avg}$$

(3)

where $\:{T}_{avg}$ is the average trust value, $\:{E}_{avg}$ is the average residual energy, $\:{C}_{avg}$ is the congestion level of nodes in $\:{R}_{i}$ and $\:\alpha\:,\:\beta\:,\gamma\:$ are adjustable weight parameters for optimization.

• 3. Apply EHO optimization

  • Matriarch Update: Nodes with high $\:{T}_{n}$ and $\:{E}_{n}$ are prioritized for routing.

  • Clan Separation: Nodes with low $\:{T}_{n}$ or $\:{E}_{n}$ are removed from candidate paths.

  • Calf Movement: New potential paths are evaluated dynamically.

• 4. Selection of Optimal Route $\:{\varvec{R}}_{\varvec{o}\varvec{p}\varvec{t}}$: The route with the highest $\:F\left({R}_{i}\right)$ is chosen for packet transmission.

Step 4: Data Encryption using Elliptic Curve Cryptography (ECC).

Elliptic Curve Cryptography (ECC) applies encryption methods prior to data transmission to provide secure communication with low performance impact.

• 1. Key exchange between source and destination

  • The source $\:S$ and destination $\:D$ generate public-private key pairs $\:\left(P{K}_{S},S{K}_{S}\right)$ and $\:\left(P{K}_{D},\:S{K}_{D}\right)$.

  • They establish a shared session key $\:K$ using ECC key exchange:

$$\:K=ECC\_KeyExchange\left(P{K}_{S},P{K}_{D}\right)$$

(4)

• 2. Encryption at source node

  • The source node encrypts packet $\:P$ using $\:K$:

    $$\:{E}_{data}=Encrypt\left(K,P\right)$$

    (5)

  • The encrypted packet $\:\left({E}_{data},{R}_{opt}\right)$ is transmitted over the selected secure route.

• 3. Decryption at destination node

  • Upon receiving $\:{E}_{data},\:$the destination node decrypts it using the shared key $\:K$:

    $$\:{P}^{{\prime\:}}=Decrypt\left(K,{E}_{data}\right)$$

    (6)

  • The integrity of the data is verified before processing.

Step 5: Adaptive security response mechanism.

• 1. Intrusion detection at destination:

  • The receiver node uses K-means clustering to perform normal or malicious packet classification when receiving data.

  • The system detects malicious packets which leads to marking the sender as an intruder.

• 2. Security alerts and trust updates

  • A security alert is broadcast to neighboring nodes.

  • The trust value $\:{T}_{n}$ of the compromised node is reduced, preventing it from future routing decisions.

• 3. Route re-selection in case of attacks:

  • A detected attack triggers the network to perform another EHO run for finding an alternative secure routing.

  • The system provides an uninterrupted method to maintain secure communication.

Consequently, a very important and useful method for ensuring information security in cryptography and crypto-analysis is the cryptographic analysis of digital signatures that use public-key cryptography. Therefore, in addition to the capability of providing non-repudiation of the message, it does not only offer authentication of the message but also ensures the integrity of the data in cryptography. The could be accomplished by the use of Digital Signature. For message authentication, the sender must possess the correct secret private key in order to make the verifier to ascertain that the digital signature was created by that sender. The secret private key will remain secure as this key is kept undisclosed to third parties.

Table 2 Notations and their descriptions.

Data Integrity: There are certain cases where a malicious node having access to the data can intentionally modify it. Consequently, this node would be unable to verify digital signatures at the receiving end. Due to this its signature won’t match the hash of the revised data or the verification algorithm’s output. The destination is now in a position to securely reject the message on the assumption that the algorithm’s data integrity would have been compromised.

Non-repudiation: When it comes to cryptography, only the signer can create a unique signature on any given message’s contents. Consequently, the signature key is usually only a secret number known only to the signer. So, if a dispute arises down the road, the receiver may provide the data and digital signature to an impartial third party as evidence that the data is secure.

In this manner, the proposed methodology validates the node’s authenticity before all authorized nodes enter into the network. The nodes which are not authorized are not allowed into the network and are sent to the CA. Table 2 shows the notations used in proposed technique with their descriptions.

Intrusion-aware on demand route selection

When all nodes are validated, the next stage is to opt for the optimum path for data transmission. In this regard, AODV is considered for route selection. For the purpose of security, the conventional AODV is integrated with Elephant Herding Optimization (EHO) algorithm.

As sociable animals, elephants live in social structures consisting of females and their young, known as calves. There is a matriarch which serves as the leader of an elephant clan made up of a number of elephants. The female members prefer to reside with other members of their family, whilst the male members often reside in other locations. Over time, they will progressively develop their independence from their families, and eventually, they will entirely leave their relatives behind. In the EHO, the following presumptions are taken into consideration.

• (1) The total population of elephants is based on the numbers held by a small number of tribes.

• (2) A certain number of male elephants will disperse from their family group and live in solitude, very far from the rest of the elephant population, at the start of every generation.

• (3) In every tribe, there is a matriarch who rules over the elephants.

The operator of clan updates

To a large extent, the matriarch of an elephant herd is the one who guides and leads the herd. For the purpose of this research study, the IEHO algorithm is modified to find the most efficient path. Elephants are used to represent the possible pathways in the beginning. That is, a route that has been upgraded is chosen to be the matriarch. The location of the $\:{j}^{th}$ elephant has been modified in line with the matriarch at the following time:

$$\:{x}_{new,j,m}={c}_{m,j+a\times\:({x}_{best,m}-{x}_{m,j})\times\:r}$$

(7)

In this case, the new location is denoted by $\:{x}_{new}$, whereas the previous position is denoted by $\:{x}_{m,j}$. The position of the matriarch in the clan is denoted by $\:{x}_{best,m}$, and the scale factor is denoted by $\:a\in\:\left[\text{0,1}\right]$. The best elephant is determined using the following formula:

$$\:{x}_{new,m,j}=\delta\:\times\:{x}_{center,m}$$

(8)

To determine the $\:{x}_{center}$ in the search space with dimensions $\:{d}^{th}$, the calculation is as follows:

$$\:{x}_{center,m,d}=\frac{1}{{N}_{m}}\times\:\sum\:_{j=1}^{{N}_{m}}{x}_{m,j,d}$$

(9)

In order to choose which matriarch to choose, the fitness function $\:\left(F\left(fn\right)\right)$ of the route is determined. The primary goal of the IEHO is to find the most optimum route by taking into account a number of different goals. In order to ensure energy efficacy without compromise in security, we considered the following routing metrics: total number of nodes that are provided along the route $\:\left(t\right)$, the trust value $\:\left(TV\right)$, the residual energy level $\:\left(RE\right)$, and the congestion level (CL). The formulation of the combined $\:F\left(fn\right)$ is as follows:

$$\:F\left(fn\left({R}_{i}\right)\right)=\sum\:R{E}_{{R}_{i}},T{V}_{{R}_{i}}-\sum\:C{L}_{{R}_{i}},t\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(10\right)$$

(10)

An evaluation of the $\:{i}^{th}$ route $\:\left({R}_{i}\right)$ is performed using the equation above that is on the basis of “residual energy, trust value, and congestion level”. The calculation for each metric is performed as follows:

Direct trust value

The trust model includes an evaluation of direct trust value. Numerous factors are included in this assessment, such as energy trust, data trust, and communication trust. The Communication Trust metric is used to evaluate the ability of sensor nodes to effectively communicate with each other and execute the approved protocol. The concept of data trust pertains to the evaluation of the reliability of data generated and updated by sensor nodes. This examination aims to assess data consistency as well as fault tolerance. The energy trust measure is used to assess a node’s residual energy reserves in order to determine whether they are enough for carrying out further communication and data processing tasks. The communication patterns and data consistency to create a reliable environment is examined that is impervious to errors and attacks. This ensures the dependability of the network’s operations and security. The environment is strengthened by this approach. In order to prevent the network from including nodes with low levels of competition, the energy component is considered that also protects network integrity. The whole methodology used in the direct trust computation is explained in next sections.

The assumption is that a node is capable of putting in place a watchdog mechanism that allows it to keep track on the activities of other nodes that are next to it and are within its communication range. In the node, it is anticipated that this capacity will be present. An example of this would be a node that has the ability to monitor the transmissions of other nodes in its immediate vicinity. This enables the node to assess whether or not it is actively sending the packets or just rejecting them. It is common practice in the overwhelming majority of contemporary trust models to define trust as the chance that a particular node would exhibit the behaviour that is anticipated of it and fulfil its responsibilities toward another similar node. This particular definition is used in a wide variety of settings. For the purpose of replicating the interactions that take place between sensor nodes, it is presumed that the Beta distribution will be used as the distribution that came before it. For the purpose of tracking and storing information on the exchanges that take place, a monitoring system is used nowadays. Through the utilization of this method, the amount of well-behaved or harmful behaviours that are shown between nodes can be determined, which is then used in the calculation of the trustworthiness value. On the basis of this value, a determination is made on whether or not a node may be trusted. To ascertain the likelihood that the nodes in the surrounding region will act in a positive way, the credibility model that is used at each node makes use of the Beta probability density function. This is done in order to assess the likelihood that the nodes will behave in a positive manner. Due to the fact that trust modelling comes with its own set of inherent uncertainties, this method has been selected in order to handle such concerns. There is a mathematical equation that represents the beta probability density function, and it may be used to explain the function.

$$\:f(x|\gamma \:,\beta \:) = \frac{1}{{B\left( {\gamma \:,\beta \:} \right)}}x^{{\gamma \: - 1}} \left( {1 - x} \right)^{{\beta \: - 1}}$$

(11)

The parameters that are indexed and the range of values that are permitted for $\:x$ is from 0 to 1. If the number of outcomes that are successful (well behavior) is represented by $\:\left(w\right)$, and the number of outcomes that are unsuccessful (malicious conduct) between nodes is represented by $\:\left(m\right)$, then the probability of outcomes may be found by setting the values for and in the following manner. Under the assumption that $\:\left(w\right)$ reflects the number of successful results (well behaviour), and $\:\left(m\right)$ represents the number of unsuccessful outcomes (malicious behavior), a conclusion may be drawn. Based on the premise that the variable $\:\left(w\right)$ represents the number of favorable outcomes and the variable $\:\left(m\right)$ represents the number of negative outcomes, this is the conclusion that can be drawn.

$$\:\gamma\:=w+1$$

(12)

$$\:\beta\:=b+1$$

(13)

Using the established model, it is possible to gain the communication trust, which is denoted by the notation $\:{T}_{C}\left({t}_{i,j}\right)$. This trust is derived from the direct observations that node $\:i$ has made on node $\:j$ at time $\:t$, which is defined as,

$$\:{T}_{C}\left({t}_{i,j}\right)=\frac{ST+1}{ST+UT+2}\left(1-\frac{UT}{ET}\right)\left(1-\frac{1}{ST+\tau\:}\right)$$

(14)

where the values $\:ST$ and $\:UT$ respectively provide a representation of the total number of successful and unsuccessful interactions that occurred between nodes $\:i$ and $\:j$ during the course of the time period t. On the other hand, these conversations are not the same as the one that was just finished taking place a few moments earlier. In that particular conversation, it was decided whether or not the node is forwarding or rejecting the packets. In this context, the communication pattern and the consistency of the data are considered in order to ascertain whether or not the interaction between the nodes was successful. This formula is referred to as the punishment factor, and it is written as $\:\left(1-\frac{UT}{ET}\right)$, where $\:ET$ is the total number of effect interaction data. $\:T$. It is the punishment factor that is responsible for penalizing the trust value. This component is determined by the variable quantity of harmful actions that are shown between separate nodes. In the context of mathematical expressions, the phrase “regulating function” refers to a function that demonstrates a rapid convergence towards 1 as the number of successful interactions rises. A positive constant is represented by the variable “α” in this case. This constant may be changed to influence the pace at which the expression approaches the value of 1. An extended period of time would be required for the procedure by which one node would increase its trust value for another node.

The implementation of the energy trust is being done primarily for the purpose of carrying out an exhaustive assessment of the performance of a node of the network. It is possible that the introduction of the energy component might serve as a suitable approach to offset the problem of diminished competitiveness among nodes that are involved in network activities. When a node’s energy consumption falls below a certain energy threshold, the node is restricted to executing only the most basic of activities. This is done in order to lengthen the overall lifetime of the network and to ensure that there is a balanced distribution of energy consumption across all of the nodes. These are the characteristics that define the energy trust concept:

$$\:E_{T} \left( {t_{j} } \right) = \left\{ {\begin{array}{*{20}l} {0,\quad {\text{if}}\:E_{{res}} < E_{{th}} } \\ {1,\quad {\text{else}}} \\ \end{array} } \right.$$

(15)

$\:{E}_{res}$ is the symbol used to represent the residual energy of the node, while $\:{E}_{th}$ is the symbol used to represent the threshold value. The following equation is used to determine the total level of trust,

$$\:O{T}_{ij}\left(t\right)=\frac{{T}_{C}\left({t}_{i,j}\right)+{E}_{T}\left({t}_{j}\right)}{2}$$

(16)

Separating operator

Within the EHO, the separating operator is enhanced by the incorporation of the genetic operator, also known as crossover, as the separating operator. Separation is then carried out in the following manner:

$$\:{x}_{worst,m}={x}_{min}+\left({x}_{max}-{x}_{min}\right)\times\:rand$$

(17)

Both the minimum and maximum numbers are used to determine the upper and lower bounds, respectively. The suggested EHO algorithm chooses the optimal route by considering all of the metrics mentioned above during the course of iteration. The flow of EHO is given in Fig. 3.

Fig. 3

Flow of EHO-AODV protocol.

When using AODV, the network is fully quiet as long as it is not necessary to establish a connection in order to transmit a data packet. In the event that it is necessary to look for routes on the network, AODV will implement the following packets, which are described by its protocol:

• 1. Route request (RREQ).

• 2. Route reply (RREP).

• 3. Route error (RERR).

Due to the fact that these communications may be implemented as straightforward UDP packets, the Internet Protocol (IP) continues to serve as the foundation for routing.

Data encryption by ECC algorithm

Commonly abbreviated as ECC, Elliptic Curve Cryptography is a key-based encryption technology. Using public and private key pairs to encrypt and decode internet communication is the main focus of ECC. Let us consider that ECC public and private key pair is available. Encrypting and decrypting data has to be done with these keys. Asymmetric encryption works as follows: after encrypting data using a public key, it is possible to decrypt it using the private key that matches the public key. There is an elliptic curve for cryptography over a finite field and that curve has a generator point G. By combining the two functions shown below, we may generate an encryption and decryption shared secret key. These functions are obtained from the ECDH scheme.

Steps involved in ECC based encryption and decryption

The destination node $\:D$ receives encrypted and routed packets which need security threat analysis. Received packets undergo K-Means clustering analysis to get divided into two distinct categories by the algorithm.

• Cluster $\:{C}_{1}$ (Normal Packets): Data packets adhere to standard protocols when they follow predefined network protocols.

• Cluster $\:{C}_{2}$ (Malicious Packets): Unusual packets detected because of abnormal time delays or irregular sender operations.

The classification is based on network features extracted from each received packet, including packet delay $\:\left({d}_{p}\right)$, source reputation $\:\left({R}_{s}\right)$, and anomaly score $\:\left({A}_{p}\right)$. These features are collectively represented as a feature vector:

The classification system analyzes three network features extracted from received packets which include packet delay $\:\left({d}_{p}\right)$, source reputation $\:\left({R}_{s}\right)$, and anomaly score $\:\left({A}_{p}\right)$. A feature vector includes these features all together:

$$\:{X}_{i}=\left({d}_{p},{R}_{s},{A}_{p}\right)$$

(18)

where $\:{X}_{i}$ is the feature set of packet $\:i$. The clustering system groups similar network traffic patterns together to identify security threats through the evaluation of behavioral data attributes. The K-Means algorithm starts by setting random starting points as $\:{\mu\:}_{1}$ and $\:{\mu\:}_{2}$ that serve as initial cluster centers for normal and malicious traffic types. Packet assignment occurs by finding its nearest cluster through Euclidean Distance that calculates packet-cluster similarity levels. The Euclidean distance computation between packet $\:{X}_{i}$ measures its distance to $\:{\mu\:}_{k}$ cluster centroid through the following formula:

$$\:d\left({X}_{i},{\mu\:}_{k}\right)=\sqrt{{\sum\:}_{j=1}^{n}{\left({X}_{ij}-{\mu\:}_{kj}\right)}^{2}}$$

(19)

where $\:d\left({X}_{i},{\mu\:}_{k}\right)$ is the distance between packet $\:{X}_{i}$ and cluster centroid $\:{\mu\:}_{k}$, $\:j$ represents the feature dimensions (i.e., $\:{d}_{p},{R}_{s},{A}_{p}$), and the packet is assigned to the cluster with the minimum distance. Once all packets receive cluster assignment the centroids get computed by finding the mean position of points inside each cluster. The new centroid $\:{\mu\:}_{k}$ of a cluster $\:{C}_{k}$ is updated as follows:

$$\:{\mu\:}_{k}=\frac{1}{\left|{C}_{k}\right|}{\sum\:}_{{X}_{i}\in\:{C}_{k}}{X}_{i}$$

(20)

where $\:\left|{C}_{k}\right|$ represents the number of packets assigned to cluster $\:{C}_{k}$. The clustering process runs sequentially until centroids stop changing substantially between rounds of execution.

The system performs categorization of packets through cluster allocation at the conclusion of clustering. The system accepts and forwards to processing any packet classified under Cluster $\:{C}_{1}$ (Normal Packets). The security response adapts based on packet allocation to Cluster $\:{C}_{2}$ (Malicious Packets). When a malicious packet reaches the network, the system identifies the sender node as an untrustworthy intruder which results in lowering its trust value $\:T$ to prevent further routing participation. The procedure for trust updates includes the following definition:

$$\:{T}_{n}^{new}={T}_{n}^{old}-\delta$$

(21)

where $\:\delta\:$ is a penalty factor that reduces the trust value upon malicious activity detection. Security alerts travel between neighboring nodes while the system operates route selection again to prevent compromised nodes from future transmission routes.

Advantages of integrating DSA, ECC, and K-means clustering

Security in Mobile Ad-Hoc Networks (MANETs) faces persistent challenges because of their lack of central control and dynamic conditions together with their susceptibility to multiple routing-based attacks. Existing traditional security solutions implement either encryption based on cryptography or trust routing combined with anomaly detection, yet they cannot provide complete protection against sophisticated attacks when used individually. An attack from within the network remains possible when employing encryption methods alone and unauthenticated anomaly detection systems cannot recognize genuine network members from fake ones. Sec-BIO establishes its uniqueness through the combination of Digital Signature Algorithm (DSA) along with Elliptic Curve Cryptography (ECC) along with K-Means clustering as security components. The multi-layered approach of this framework provides comprehensive authentication with lightweight encryption and adaptive intrusion detection which enables secure efficient MANET communication operations. Sec-BIO implements a complementary security network that makes security solutions stronger through multiple integrated security processes to deliver a complete solution.

Sec-BIO embraces three security aspects using DSA along with ECC protocol and K-Means clustering to deliver three-dimensional defense mechanisms for security, confidentiality safeguards and rapid intrusion detection. The techniques in combination stops Blackhole, Greyhole, Wormhole and Sybil attacks as these exploit security weaknesses in traditional MANET security models. The entry of malicious nodes into the network depends heavily on proper authentication procedures. DSA lets Sec-BIO establish trusted node verification which guarantees legitimate nodes access to routing activities. DSA digital signature creation keeps Sybil and Blackhole attacks at bay by stopping both fake route replies and node impersonation attempts. Digital signature validation becomes exceptionally challenging for attackers when they cannot access private key information needed for creation. Therefore, the risk levels for unauthorized access and routing manipulation decrease substantially. This authentication method stands crucial in MANETs because these networks include constantly changing trust relations along with regular node network entries and exits.

The security measure of authentication stops unauthorized access to networks but fails to encrypt or protect the integrity of data during transmission. Sec-BIO employs ECC as its encryption standard which provides robust communication security at a low computational cost. The use of ECC produces superior efficiency when compared to RSA encryption measures through its capability to provide equivalent security with compact encryption keys. The system provides optimal performance benefits during operations of power-limited MANET nodes which depend on efficient resource utilization. The encryption of packets with ECC at Sec-BIO creates an impervious environment for data through which an attacker must withstand because they cannot decode or alter captured information thus protecting against Wormhole and packet injection attacks. Despite successful authentication measures with encryption techniques the MANET network remains vulnerable to routing-based attacks which include Greyhole and Wormhole attacks through which nodes can control routing paths. The receiver end of Sec-BIO implements K-Means clustering as part of its intrusion detection system to detect these dynamic attacks through anomaly-based operations. The K-Means clustering system operates differently from static intrusion detection by implementing a dynamic classification process that groups packets into normal or malicious categories through analysis of packet delay together with source reputation and anomaly score. This machine learning model automatically detects evolving attacks whereas it swiftly isolates compromised nodes from routing protocols. The continuous updating of cluster centroids through Sec-BIO enables detection of selective packet drops that occur over time thus making it highly successful against Greyhole attacks in which attackers attempt sporadic packet forwarding to avoid detection.

Novelty of the work

The proposed Sec-BIO methodology introduces a novel approach for enhancing the security and performance of Mobile Ad-Hoc Networks (MANETs). By integrating the Ad-Hoc On-Demand Distance Vector (AODV) protocol with the bio-inspired Elephant Herding Optimization (EHO) algorithm, the model ensures optimal routing while addressing core security challenges in MANETs. Unlike conventional methods, Sec-BIO combines trust-based route selection with energy efficiency, ensuring secure and reliable data transmission. Additionally, the incorporation of the Digital Signature Algorithm (DSA) for node authentication and Elliptic Curve Cryptography (ECC) for data encryption adds robust layers of protection against attacks such as blackholes, greyholes, and wormholes. The innovative use of K-means clustering at the receiver end for intrusion detection further strengthens the security framework, ensuring that malicious nodes are effectively identified and isolated. This integrated design not only improves routing efficiency but also ensures high data delivery rates with reduced latency and energy consumption, making it a significant advancement in the domain of secure MANETs.

Results and discussions

Network Simulator version 2.34 served as the platform for conducting tests that simulated Mobile Ad-Hoc Networks (MANETs). The researchers use Network Simulator (NS2/NS3) as their evaluation platform for Sec-BIO performance testing because this platform is established within the MANET research field. Repeatable simulation tests occurred as all network characteristics such as node density along with mobility patterns, traffic models and attack scenarios with security mechanisms are defined before running multiple trials. The intrusion detection model using K-Means clustering applied real-world attack pattern datasets for training in order to ensure genuine security threats are reflected in its results rather than artificial simulation values. The simulation world replicated an area measurement of 1200 × 900 m using 100 mobile network nodes that shared a radio transmission distance of 250 m. Reliable wireless communication functioned through the implementation of the IEEE 802.11 MAC layer protocol. The network received steady packet transmission through Constant Bit Rate (CBR) traffic generation method. A simulation duration of 30 milliseconds provided optimal evaluation of the dynamic conditions that the proposed methodology would encounter. The Dynamic Source Routing (DSR) protocol was utilized as the base routing mechanism due to its reactive nature, which adapts to topology changes effectively. Each packet had a size of 150 bytes, balancing data transmission and overhead management. Node mobility followed the Random Waypoint mobility model, which simulates realistic movement patterns in MANETs. This setup facilitated a comprehensive analysis of network performance, including metrics such as throughput, energy consumption, routing overhead and packet delivery ratio, under varying network conditions. The simulation results validated the effectiveness of the proposed Sec-BIO methodology in enhancing security, efficiency and scalability in MANET environments.

Table 3 Simulation setup.

The simulation setup described in the Table 3 can thus considered as a rather elaborate framework for the assessment of the proposed Sec-BIO methodology. The simulation made with an overall of one hundred mobile nodes dispersed over the region of 1200 by 900 m because the nodes require enough space for operation and interaction. This makes the transmission range of the node to be 250 m so that high-level data can be transmitted with direction maintaining the formation of the Mobile Ad-Hoc Network (MANET). The MAC layer protocol adapted is the IEEE 802.11, which is widely accepted because of its high reliability in wireless communication. Traffic generation based on Constant Bit Rate (CBR) to guarantee a constant data flow through the whole network and the simulation time is set to 30 milliseconds to have a fast assessment of the proposed solution. For this purpose, the Dynamic Source Routing (DSR) protocol is used which provides reactive routing for MANETs because of their dynamic nature. Data payload fixed at 150 bytes for each packet in order to provide an efficient load of data on the packet while at the same time having small packet sizes. The movement of nodes is according to a random waypoint model, which gives a more realistic picture of mobile nodes in an ad-hoc network. Such configuration provides standardized testing environment to measure the performance parameters including throughput, energy usage, delay and routing overhead in order to establish the efficiency of Sec-BIO under different network scenarios.

The collaboration between DSA and ECC in conjunction with K-Means clustering forms a flexible security system for Sec-BIO that delivers superior performance than conventional MANET security models. The security methods currently in use focus individually on trust-based routing, cryptographic authentication or anomaly detection which separately cannot handle multiple aspects of MANET security threats. Trust-based routing operates independently of malicious activities performed by trusted insiders which transits to hostile behavior. The separate security measures of cryptographic authentication and anomaly detection without encryption fail to block selective packet drops by attackers and packet tampering respectively. Sec-BIO unites these three security methods to strike an equilibrium among network resistance, operational speed and operational flexibility. The DSA authentication functionality enables the system to restrict network routing only to legitimate nodes while protecting the setup from unauthorized access and route manipulation. ECC encryption creates a lightweight yet powerful encryption system that makes data confidentiality impossible to compromise by attackers who succeed in gaining network access. The K-Means clustering system performs automated packet classification thus delivering real-time breach detection which rapidly locates and removes harmful routing nodes. Sec-BIO demonstrates both scalability and energy-efficiency features which existing approaches lack. The adoption of ECC encryption protocols instead of RSA produces less burden on computational resources thereby making Sec-BIO practical for scarce-resource MANET devices. The computational lightness of K-Means clustering makes it well-suited for real-time attack prevention in networks since it avoids excessive strain on available resources. Through its ability to automatically update trust metrics and shift traffic to different nodes when security breaches occur, Sec-BIO operates as a secure system that preserves residual operation alongside sustained energy efficiency and packet delivery efficiency.

Table 4 Packet delivery ratio (PDR).

Fig. 4

Packet delivery ratio (PDR) vs. number of nodes.

Table 4; Fig. 4 featuring the Packet Delivery Ratio (PDR) also indicate that the proposed Sec-BIO methodology is much more efficient than E-AODV and AODV. It is clear that Sec-BIO shows higher PDR values than the others at all the nodes, meaning the model’s effectiveness with large numbers of nodes to ensure that data being transmitted dependably. For instance, with 10 nodes, the PDR obtained by Sec-BIO is 99% while E-AODV, AODV gives 96% & 93% respectively. For number of nodes greater than 50, the PDR declines very little due to traffic and routing issues. However, Sec-BIO stands with a qualitative advantage achieving 90% PDR with 100 nodes against just 85% by E-AODV and 80% by AODV. This can be explained by the ability of Sec-BIO to perform trust-based route selection and energy optimizing procedures that guarantee reliable end-to-end packet transmission even when the network topology is constantly changing. Unlike AODV, E-AODV and AODV have a steeper curve, which show that they are unable to handle congestion and routing failures properly. Therefore, the result confirms that Sec-BIO is able to provide better packet delivery rates in MANET and can be considered reliable for secure and efficient MANET operations in resource-poor or high mobility environments.

Table 5 End-to-end delay.

Fig. 5

End-to-end delay vs. number of nodes.

The evaluation of the end-to-end delay of the Sec-BIO with respect to E-AODV and AODV protocols is illustrated in Table 5; Fig. 5 in case of a changing number of nodes in a MANET. These outcomes indicate that low delay realized by Sec-BIO throughout the evaluation, which proves its capability of finding the shortest path for data forwarding with minimal delay. For instance, when using 10 nodes, the delay that is experienced by Sec-BIO is only 40ms while E-AODV experiences 60ms and AODV by 80ms. However, the delay for all protocols increases as the network size expands because of enhanced routing difficulty and competition. However, the overall improvement rate for Sec-BIO is much slower, rising gradually from 90ms with 100 nodes for Sec BIO reaching 125 ms whereas it is 160ms for AODV. This reduction in delay by Sec-BIO is quite considerable compared with the base Intelligent Information System because of BIO EHO selectivity in choosing routes and its concern with trust, energy, and congestion optimization. The results further depict that both E-AODV and AODV indicate higher delay because of their poor routing algorithms and their layout’s inability to change dynamically. These outcomes imply that Sec-BIO is superior in achieving lower latency and hence very relevant for delay sensitive applications such as Real-Time communications and IoT.

Table 6 Throughput comparison vs. no. of nodes.

Fig. 6

Throughput comparison vs. number of nodes.

Table 6; Fig. 6 shows the throughput of both the complex protocols, which are Sec-BIO, E-AODV & the basic protocol AODV for a number of nodes in a MANET cluster. Transmission rate in Kbps, is the measure of efficiency of data transfer in a network protocol. From the performance results, it is clear that Sec-BIO provides higher throughput compared to E-AODV and AODV throughout the entire matrix of node density. For instance, with 10 nodes and using Sec-BIO achieves 940 Kbps of throughput compared to 930 Kbps of throughput for E-AODV and 920 Kbps of throughputs for AODV. Typically, as the number of nodes rises, the throughput gradually tends to decline because of overcrowded nodes compounded by routing complications that affect all protocols. Nevertheless, in the case of Sec-BIO a generalized gain is observed, with a throughput of 915 Kbps at 100 nodes, against 905 Kbps of the E-AODV and 895 Kbps of the basic AODV. The reason for better performance of Sec-BIO is due to the integration of trust-based routing and energy aware optimization to minimize packet drop rate and efficient network resource utilization. Still, E–AODV and AODV experiences packet loss and inefficient routing than other protocols, thus decreasing the throughput. These results demonstrate that Sec-BIO is well suited for throughput sensitive applications and large data intensive networks such as multimedia streaming and real-time data transfer.

Table 7 Network lifetime vs. No. of nodes.

Fig. 7

Network lifetime vs. number of nodes.

Table 7; Fig. 7 represent the network lifetime of the protocols versus the number of nodes namely for Sec-BIO, E-AODV and AODV protocols. Network lifetime is in terms of seconds that forms a most useful parameter to determine the feasibility of routing protocols to small networks with limited resources. The analysis also reveals that Sec-BIO performs better compared to E-AODV and AODV in terms of the network lifetime in all node instances. For illustration, it is observed that the network lifetime is of 520 s for the Sec-BIO with 10 nodes while it is 500 and 480 s for E-AODV and AODV respectively. When the number of nodes is high then the duration for which the network will be running will be short due to high energy consumption for data transmission and route establishment. However, Sec-BIO still has high advantage where the network accomplished the lifetime of 420 s for 100 nodes as compared to E-AODV having lifetime of 400 s and traditional AODV with 380 s. This performance attributed to energy aware routing implemented in Sec-BIO to select the best path that has high residual energy and low congestion. Furthermore, in Sec-BIO, the Elephant Herding Optimization (EHO) algorithm avoids energy drain on important nodes that limits the network’s lifetime by distributing the energy being used fairly across all nodes. These results show the superiority of Sec-BIO in increased lifetime of MANETs in critical situations that need prolongation of the network lifetime in case of disasters and remote monitoring.

Table 8 Energy consumption.

Fig. 8

Energy consumption vs. number of nodes.

Table 8; Fig. 8 show the energy utilization of the Sec-BIO, E-AODV and AODV protocols as the number of nodes in the MANET increases. Energy (energy consumed in joules, J) is one of the most important criteria in the evaluation of sustainability and efficiency of routing protocols, particularly in energy-limited environments such as MANET. From the result analysis, Sec-BIO has less energy consumption compared with E-AODV and AODV in all node levels. For instance, with 10 nodes, energy consumed by Sec-BIO is 1340 J while that of E-AODV and AODV are 1350 J and 1360 J respectively. This trend extends when number of nodes raises. The value is rather lower in case of Sec-BIO, which is 1390 J for 100 nodes while E-AODV and AODV takes 1400 J and 1410 J respectively. The decrease in energy consumption for Sec-BIO due to the incorporation of energy-based routing protocols that enhances routes with superior residual energy and maintains a balance of energy throughout the network. Besides, the Elephant Herding Optimization (EHO) algorithm reduces extra retransmission and congestion, thus attaining lower energy consumption level. Compared to G-AODV, two other algorithms E-AODV and AODV consume relatively more energy because they use less efficient routing algorithms that involve unnecessarily high routing overhead and multiple data transmissions. These results show that increase in energy consumption managed efficiently by Sec-BIO for energy sensitive applications such as environment monitoring and disaster recovery.

Table 9 Routing overhead.

Fig. 9

Routing overhead vs. number of nodes.

To analyse the routing overhead, Table 9; Fig. 9 shows the results of the simulation of the number of total routing messages in each of the executed protocols including Sec-BIO, E-AODV, and AODV in MANET with variable node count. Routing overhead or routing load shows the percentage of control packets out of the total packets sent and considered as one of the most important parameters used in the analysis of routing protocols. In all simulations the routing overhead of Sec-BIO is substantially lower than that of the other two algorithms, signifying that this makes it the more efficient method of reducing unnecessary control packets. For instance, when nodes are 10, the additional overhead cost of Sec-BIO is 3%, which is far much lower than that of E-AODV (5%) and AODV (7%). This trend increases inversely to the number of nodes, although all protocols studied in this work keep a low overhead of 13% with 100 nodes, E-AODV is with 16% and AODV with 18%. Sec-BIO involves a smart routing mechanism, and the choice of a path determined by the EHO that ensures stability in routing and a minimal RH overheard. This reduces the incidences of route rediscovery, and control packet exchange, particularly in dynamic conditions in the network. In this context, it is observed that overhead in E-AODV and AODV is relatively high because of the use of many control packets for route maintenance. These conclusions also reveal that Sec-BIO is an effective solution in terms of network load and resource utilisation, so it is easy to scale up to accommodate more nodes in MANET for large and dynamically changing networks.

Table 10 Comparative performance of intrusion detection methods.

Fig. 10

Comparative performance of intrusion detection methods.

Table 10; Fig. 10 showcase different intrusion detection techniques that have been assessed through precision, recall, accuracy and F1-score metrics. The security analysis using K-Means (Sec-BIO) yielded 95.69% accuracy which surpassed Random Forest with 94.31% accuracy and both SVM and Naïve Bayes achieved 92.15% and 89.76%. The K-Means (Sec-BIO) technique achieved the best precision rate of 96.44% because it successfully detects positive cases while generating a minimum number of false positives. Random Forest demonstrated slightly lower precision of 94.8% while precision measurements for SVM and Naïve Bayes accounted to 91.88% and 88.21%, respectively. The ability to correctly detect all actual intrusions showed by K-Means (Sec-BIO) that produced the highest recall value of 95.06% followed by Random Forest with 93.9% and SVM reached 92.4% and Naïve Bayes achieved 90.05%. The F1-score analysis proved K-Means (Sec-BIO) to be the most balanced method at 95.74% followed by Random Forest at 94.35% while SVM and Naïve Bayes recorded lower scores. The performance levels of SVM and Naïve Bayes came in at 92.13% and 89.12% respectively. The intrusion detection method K-Means (Sec-BIO) delivers the highest efficiency according to the results presented in Table 10. Table 11 shows the comparative analysis of AODV, EHO-based security models, and Sec-BIO.

Table 11 Comparative analysis of AODV, EHO-based security models, and Sec-BIO.

Table 12 Throughput comparison with statistical validation.

Fig. 11

Mean throughput comparison across protocols.

The comparison of throughput performance among AODV, EHO-Based, and Sec-BIO protocols emerges through Table 12; Fig. 11. The experimental data shows that Sec-BIO delivered maximum mean throughput of 715.2 kbps surpassing the other protocols followed by EHO-Based at 620.7 kbps until AODV reached 512.4 kbps. Analysis of Sec-BIO yielded the most consistent performance results based on its standard deviation of 16.8 while EHO-Based showed slightly higher variation at 21.3 and AODV demonstrated moderate results with 18.6. The 95% confidence interval analysis demonstrates that Sec-BIO maintains throughput consistently throughout the experiment with a low variation range between 710.3 kbps and 720.1 kbps. Better throughput performance was shown by the EHO-Based protocol but it had a wider confidence interval of (614.8, 626.6) that reflected higher data point variations. Experimental results showed AODV to deliver the smallest throughput of all protocols which had a confidence interval (507.3, 517.5). The throughput efficiency data shows that Sec-BIO achieves the highest rates making it an optimal solution for data-intensive applications which need reliable performance.

Scalability analysis of Sec-BIO in large-scale MANETs

Network security in mobile ad hoc networks depends highly on scalability since network dimensions directly affect routing performance alongside network congestion levels and processing burden. The scalability of Sec-BIO was tested through simulations involving different numbers of nodes ranging from 50 to 200 while tracking performance indicators that included throughput and end-to-end delay and energy efficiency. The research measured Sec-BIO performance through analysis with both AODV and EHO-based security while evaluating network adaptation for big network sizes.

Table 13 Throughput performance for different network sizes.

Fig. 12

Throughput performance across different network sizes.

A comparison of network throughput between AODV and its variants EHO-Based and Sec-BIO takes place through Table 13; Fig. 12 under different network conditions from 50 to 200 nodes. Sec-Bio achieved better throughput than EHO-based or AODV networks in each scenario. The Sec-BIO protocol reached a throughput rate of 715.2 kbps when operating with 50 nodes which surpassed both EHO-Based at 620.7 kbps and AODV at 512.4 kbps. The number of nodes directly affected the throughput for all protocols because network congestion and communication overhead increased. When using 100 nodes the protocols Sec-BIO, EHO-Based, and AODV provided 689.1 kbps, 587.3 kbps and 468.9 kbps successfully. The research revealed that Sec-BIO maintained 652.4 kbps throughput with 150 nodes followed by EHO-Based at 540.2 kbps then AODV concluded at 435.6 kbps. The performance of Sec-BIO exceeded both EHO-Based and AODV by achieving 618.9 kbps at 200 nodes in the network. The presented results demonstrate that Sec-BIO scales well by processing expanding networks efficiently while showing enhanced transmission rates in all cases. While AODV faced the greatest performance decline in extended network spaces it becomes ineffective for large network deployment.

Table 14 End-to-end delay performance for different network sizes.

Fig. 13

End-to-end delay performance for different network sizes.

The End-to-End Delay (E2E Delay) performance evaluation of AODV, EHO-Based, and Sec-BIO protocols through network sizes of varying range across different network sizes is shown in Table 14; Fig. 13. All measures are presented in milliseconds (ms). Throughout all examined test cases Sec-BIO showed the shortest delay performance which proves its ability to minimize packet transmission time. When running with 50 nodes Sec-BIO delivered an End-to-End delay of 184.5 ms that proved better than both EHO-Based at 295.2 ms and AODV being at 412.8 ms. The protocol delay increased in all cases as the network size grew bigger due to both traffic congestion and the overhead generated by routing. When operating with 100 nodes the delay metrics reached 211.7 ms with Sec-BIO while EHO-Based delivered 328.1 ms and AODV registered 486.2 ms. The delay of Sec-BIO at 150 nodes reached 246.8 ms which proved superior to EHO-Based (366.4 ms) and AODV (559.7 ms). During the evaluation with 200 nodes the maximum delays were observed for all protocols where Sec-BIO reached 279.5 ms while EHO-Based reached 402.9 ms and AODV reached 637.3 ms. Sec-BIO proves its role as the best choice for real-time applications as it achieves superior delay reduction over alternative protocols. AODV demonstrated the longest delay time since it proved unable to operate effectively in extensive network environments.

Table 15 Energy consumption per node across different network sizes.

Fig. 14

Energy consumption per node across different network sizes.

The analysis of energy usage per node per second for AODV and EHO-Based and Sec-BIO models can be found in Table 15; Fig. 14 throughout different network sizes. In all simulated conditions Sec-BIO demonstrated the most energy-efficient characteristics which outperformed EHO-Based and consumed less than AODV. The energy consumption of Sec-BIO remains at 0.392 J/node/s for 50 nodes thus outperforming both EHO-Based (0.487 J/node/s) and AODV (0.562 J/node/s). The increase in network size resulted in raised energy consumption for all protocols because it generates extra routing overhead and elevated communication workload. The Sec-BIO protocol used 0.418 J/node/s and AODV used 0.603 J/node/s while EHO-Based used 0.513 J/node/s when they operated on 100 nodes. Sec-BIO sustained lower energy usage of 0.451 J/node/s between 150 nodes compared to EHO-Based with 0.538 J/node/s and AODV which consumed 0.645 J/node/s. The consumption of Sec-BIO remained optimal at 0.478 J/node/s when running 200 nodes while EHO-Based and AODV reached 0.562 J/node/s and 0.684 J/node/s, respectively. The findings demonstrate that Sec-BIO optimizes network energy usage effectively thus making it appropriate for energy-limited network environments. The highest energy consumption levels were observed in AODV during the evaluation period which implies its non-optimal behavior in extensive network implementations.

Security analysis of Sec-BIO

Security analysis of Sec-BIO determines its capability to withstand Blackhole, Greyhole, Wormhole and Sybil attacks within MANET networks. The implementation effectiveness of the model depends on its trust-aware routing protocols together with authentication protocols and encryption schemes and intrusion detection systems. The use of Elliptic Curve Cryptography (ECC) authentication together with Elephant Herding Optimization (EHO) trust-aware routing and K-Means clustering intrusion detection in Sec-BIO deals effectively with attacks in MANETs while enabling secure and efficient transmission.

Malicious nodes initiate Blackhole Attacks by presenting false information about their path to destination which causes neighboring network nodes to mistake them as shortest route providers. The malicious node receives the traffic packets but drops them rather than transferring them which leads to network communication breakdown. The threat from this assault becomes dangerous because attackers can steal large amounts of data which often goes unnoticed by victims. Trust-based routing together with authentication and receiver-end intrusion detection creates the solution to fighting Blackhole Attacks in Sec-BIO. Route selection decisions heavily depend on the trust value calculated through the Elephant Herding Optimization (EHO) technique. The process automatically excludes routers whose trust ratings decrease after they send packets because of their malicious behavior. Blackhole nodes fail to authenticate through digital signatures because they cannot provide valid private keys for signing messages thus preventing them from sending fake route replies. Unauthenticated nodes send any route update which gets immediately rejected by the network. The K-Means clustering system functioning at receivers examines strange packet loss frequencies to identify rogue nodes which subsequently lead to their network restriction. Real-time packet analysis combined with authentication procedures and proactive trust evaluation procedures enable Sec-BIO to effectively stop Blackhole Attacks and protect data from loss.

Selectively dropping network traffic characterizes a Greyhole Attack as an advanced version of the Blackhole Attack against network systems. The behavior of greyhole nodes appears identical to unrelated nodes with temporary network issues leading to their elusiveness during detection. The network’s quality eventually decreases with such strategic packet discarding because standard detection tools become ineffective. Sec-BIO addresses Greyhole Attacks through adaptive trust evaluation and anomaly-based intrusion detection. The system maintains difference from traditional trust value methodologies by performing continuous trust score adjustments through observed packet forwarding assessment. A node with unsteady forwarding actions will receive reduced trust score which ultimately leads to its systematic exclusion from routing purposes. At the receiver port K-Means clustering performs an operation to categorize packets by assessing both drop rate patterns and source reputation. The system distinguishes real network packet drops from malicious ones by implementing this feature thus preventing incorrectly identifying innocent activities. The Sec-BIO system uses real-time trust models to detect Greyhole nodes while machine learning classifiers help isolate them which minimizes disruptions to data pathways.

The Wormhole Attack represents a complex attack mechanism at the network layer which enables two collaborating malicious nodes to create a secure passage between disparate locations to pass packets through the tunnel. The attackers use this setup to create a deceptive path impression which forces other nodes to direct their data through their network. The attackers gain abilities to intercept data and control routing decisions as well as to block packets following the redirection of substantial network traffic. Sec-BIO fights Wormhole Attacks by implementing a defense strategy which relies on hop count examination as well as transmission timing verifications together with trustworthy route selection mechanisms. The hop count and transmission time evaluation helps Sec-BIO recognize wormhole links because these links appear to have abnormally short distances. Routes that deviate from the anticipated distance-time relationship will automatically become suspicious and prevented from appearing in routing decisions. Route requests that wormhole attackers attempt to fake cannot be validated because they do not possess the needed cryptographic authentication. Security system monitors the trust scores of nodes while maintaining a record of unusually short linking decisions which cause trust deductions leading to network separation from participating nodes. Multi-layer validation procedures within Sec-BIO protect the network from wormhole attackers who attempt to manipulate routes and intercept data.

The Sybil Attack arises whenever an evil node generates several wrongful identities to steer routing choices and generate artificial traffic or initiate denial-service attacks. The decision-making process in MANETs based on neighbor discovery makes Sybil nodes able to greatly disrupt communication through their behavior of pretending to be many legitimate nodes. The use of unreliable routing protocols produces increased risks of both resource exhaustion and data breaches in the network systems. Sec-BIO counteracts Sybil Attacks through signature-based identity verification and trust-based node validation. ECC-based cryptographic identifiers tell each node that is legitimate to connect to the network. Vulnerable Sybil nodes cannot authenticate their signatures thus they do not succeed during verification which blocks their access to legitimate nodes. Sec-BIO implements a system to track past reputations which identifies new nodes that cannot prove their identity through multiple accounts due to suspicious behavior. The network reduces the trust score of nodes that try to duplicate their identities until they become completely disconnected from routing choices. Through its implementation of cryptographic authentication alongside adaptive trust monitoring Sec-BIO stops Sybil nodes from using the network because it solely permits trusted and validated nodes to take part in routing operations. The security analysis of Sec-BIO against current protocols appears in Table 16.

Table 16 Security evaluation of Sec-BIO vs. existing protocols.

A comparison of Sec-BIO’s security performance against existing routing protocols highlights its superiority in attack prevention. Traditional AODV routing lacks security measures, making it highly susceptible to all four attacks. The enhanced E-AODV protocol improves resilience through basic authentication, but it remains vulnerable to wormhole and Sybil attacks due to its lack of trust-aware routing and anomaly-based detection. In contrast, Sec-BIO integrates trust evaluation, cryptographic authentication, and intrusion detection, effectively detecting and isolating malicious nodes. This ensures that the network remains secure, stable, and resilient against advanced attacks. The Sec-BIO model effectively mitigates Blackhole, Greyhole, Wormhole, and Sybil attacks by combining trust-aware route selection, signature-based authentication, encryption, and receiver-end intrusion detection. Unlike traditional protocols that rely solely on reactive security mechanisms, Sec-BIO employs proactive anomaly detection and adaptive security responses, ensuring that malicious nodes are dynamically identified and removed from the network. By incorporating lightweight cryptographic security with intelligent machine learning-based intrusion detection, Sec-BIO maintains high performance while ensuring robust network security.

Feasibility of deploying Sec-BIO in practical scenarios

The deployment of Sec-BIO in real-world environments demands an evaluation of compatible hardware together with scalability expectations and computational viability and practical implementation obstacles. The security requirements for MANET applications encompass military tactical networks together with emergency response systems and IoT-based smart city infrastructure deployment because security solutions need to be dynamic and power-efficient for changing network conditions. Traditional MANET security models demonstrate constraints when implementing attacks in real-time because they produce excessive energy use together with high computational overheads while facing problems in detecting attacks. These limitations affect the usability of resource-constrained devices like IoT sensors and mobile ad-hoc communication units and UAVs. The Sec-BIO security solution addresses such problems through a combination of lightweight cryptographic authentication with bio-inspired trust-aware routing and anomaly-based intrusion detection thus enabling practical real deployments in situations where resources are constrained. Hardware compatibility stands as an essential factor in real-world feasibility because embedded devices used in MANET utilize minimal processing power and their batteries drain fast. The security methods based on RSA encryption prove ineffective for limited-power systems because they need big key lengths and high system performance. Security through ECC enables Sec-BIO to provide the same security benefits as RSA encryption by reducing key sizes and achieving better energy savings and faster computational speed. The secure authentication capabilities of Digital Signature Algorithm (DSA) help mobile devices and energy-sensitive setups with ability to work without repeated key exchange requirements. The EHO-based trust-aware path selection demonstrates superior routing efficiency than deep reinforcement learning approaches because it lowers power usage requirements for secured MANET operations. The wide deployment of Sec-BIO becomes possible since it works seamlessly across different hardware systems which include IoT sensor nodes together with Raspberry Pi-based communication devices as well as UAV-mounted edge computing platforms alongside military-grade radio networks.

Sec-BIO shows effectiveness when used for three specific areas which encompass military networks alongside emergency response systems and IoT-based smart city solutions. Sec-BIO offers military tactical networks three essential security features including authentication prevention mechanisms against Sybil attacks along with lightweight encryption for channel security and real-time intrusion detection measures to detect Blackhole and Greyhole attacks. Sec-BIO implements trust-aware path selection as an alternative approach to traditional MANET security models that depends on hop-based routing because it chooses communication paths that lead through high-trust nodes thus decreasing the possibility of malicious traffic redirection and data interception. The Sec-BIO system boosts emergency response network security through authentication of official first responders while preventing deceptive information from unauthorized sources. Machine learning-based K-Means clustering through real-time anomaly detection enables fast identification of compromised nodes thereby maintaining protected communication between rescue team members and emergency medical units along with their operational drones.

Conclusion and future work

Mobile Ad-Hoc Networks (MANETs) are networks that operate in a decentralized manner. Due to the absence of centralized control over communication, the total network is susceptible to a range of security concerns, including Blackhole, Wormhole, Flooding, and Unauthorized access. While several publications have examined these difficulties, the matter of security remains a tough one. In earlier studies, security has been achieved via the visible use of energy and resources. To address this issue in mobile network, a novel methodology known as Secure Bio-Inspired Optimization (Sec-BIO) algorithm has been presented. The overall routing procedure was constructed upon the conventional AODV protocol which uses route request and reply mechanism. To ensure security level and to eliminate malicious nodes from the network, Digital Signature Algorithm (DSA) based authentication procedure is executed with the help of certificate authority. For all nodes which proved their authenticity, optimum as well as secure route is selected by Elephant Herding Optimization (EHO) algorithm with the consideration of multiple metrics. Further, data security is ensured by Elliptic Curve Cryptography (ECC) algorithm. For intrusion detection, k-means algorithm-based packet classification procedure is presented. Overall, the proposed approach shows better efficiency in energy consumption, data transmission and delay metrics. In future, it has been planned to extend this work with blockchain technology to improve security level.