Energy-aware cluster head optimization and secure blockchain integration for heterogeneous 6G-enabled IoMT networks

Abstract

The integration of blockchain into 6G-enabled Internet of Medical Things (IoMT) networks promises secure and decentralized communication but introduces challenges related to energy efficiency, latency, and authentication overhead. Existing clustering and security schemes fail to balance these aspects effectively in heterogeneous networks. This paper proposes a novel energy-aware cluster head (CH) selection framework using Artificial Democratic Cuckoo Glowworm Remora Optimization (ADCGRO), integrated with a lightweight blockchain layer for secure authentication and data integrity. The system optimizes task allocation across advanced, intermediate, and normal IoMT devices to minimize energy depletion while meeting ultra-reliable low-latency communication (URLLC) requirements. Simulation results demonstrate that the proposed approach enhances network lifetime by 27%, reduces average latency by 35%, and achieves 99% authentication accuracy, surpassing baseline protocols such as LEACH and HEED. These results highlight the effectiveness of combining ADCGRO-based optimization with blockchain to enhance performance and security in 6G wireless networks.

Introduction

Significant developments in wireless communication and computer vision¹ complement newly developed 6G technology. Modern computer vision systems, combined with robust communication networks, could drastically alter our perspective and interaction with the physical world. With data speeds exceeding 100 GB/s and latency of less than one millisecond, 6G is predicted to be the next significant advancement. It aims to build on the successes of 5G. Once unthinkable, real-time computer vision applications are now realistic due to their unparalleled speed and connectivity. The promise of global connectivity offered by 6G enables computer vision systems to operate in far-off, previously unreachable regions. This will generate various opportunities in sectors such as augmented reality, transportation, agriculture, and others². SixG technology marks a new era defined by ultra-fast data transmission, creative computer vision, and remedies to long-standing problems that will fundamentally alter our digital lives. Our actual experiences will pale compared to our dreams. Blockchain technologies could enable the integration of computer vision systems with 6G networks. This aligns perfectly with the distributed ledger and immutable transaction records of blockchain technology³. Three fundamental needs—data privacy, integrity, and trust—have a solution in blockchain technology.

Blockchain technology is crucial for computer vision and 6G networks, as it ensures data confidentiality and integrity. Protecting the privacy and security of this data is crucial, as 6G networks enable the rapid transmission of large data volumes required for computer vision operations. Based on cryptographic concepts, the immutable ledger of blockchain technology protects data from unauthorized access or manipulation. This is particularly important in contexts like driverless automobiles, smart cities, and remote medical diagnostics, where data confidentiality and accuracy are vitally critical⁶. Blockchain technology also encourages individuals to trust the system⁴. Using computer vision technology means relying on the accuracy of the visual data and its source. The distributed ledger properties of blockchain technology, along with its immutability, enable users to authenticate data sources and verify the correctness of computations. People’s cooperation on projects, including augmented reality systems and multi-agent robotics, depends on personal confidence to run regularly and consistently⁵. Blockchain technologies also enable the simplicity of data monetization and provenance tracking.

Smart contracts and tokenization enable participants in the computer vision and 6G ecosystems to transact services or data in a fair, transparent, and efficient manner. The expansion of data marketplaces has increased the chances for computer vision data providers to receive fair compensation. Using blockchain technology, tracking data origin provides responsibility and traceability qualities extremely essential for audit purposes and regulatory compliance⁶. Six G networks are rather close to more decentralization in the current digital information era. Still, if we are to further this process, several significant obstacles have to be overcome. Among these challenges are building confidence, pooling resources, ensuring secure data transfer, implementing access controls, protecting privacy, and establishing effective governance^7,8. Typically, conventional network security solutions, which rely on encryption algorithms and centralized authority, ensure data protection, transparency, and dependability. These techniques might not be sufficient, though, in the highly dispersed systems usual of a 6G environment, where scalability, decentralization, and real-time operation are paramount.

The decentralization, immutability, and traceability of blockchain technology make it a perfect candidate for the optimum moment to emerge; it solves present-day challenges and might greatly improve the security and dependability of 6G networks⁹. Blockchain technology has proven to be quite valuable in enhancing the security of wireless networks, particularly in 6G communication networks. Its high power consumption is one of its major downsides, particularly for nodes with short battery lives¹⁰. Network scalability and stability are compromised when nodes fail at a rapid pace, which has a significant impact on the energy consumption associated with the sophisticated consensus process¹¹. Energy usage is drastically reduced through innovative solutions, such as the Green Sharding (GS) system, which decreases the involvement of wireless nodes in global communication. These systems use node sharding technologies and are based on geolocation. This addresses the primary issue. There have been reports of promising results using this method for millimeter-wave (mmWave) and terahertz (THz) communications¹². Several consensus mechanisms optimized for 6G scenarios are necessary for integrating blockchain with 6G networks.

While GS and other optimization methods help to solve energy consumption problems, this flexibility directly influences the general efficiency and performance of the system¹³. Researchers have previously attempted to integrate blockchain technology with wireless communication; however, the protocols they created have not satisfied the requirements for low-latency and high-concurrency transactions¹⁴. Table 1 reveals that traditional methods of attaining agreement typically exhibit less scalability, greater delay, and inferior throughput. This is the outcome of their greater attention to maintaining constant quality.

Critical healthcare services are increasingly dependent on IoMT systems, which have motivated this study because of the potential catastrophic effects of network outages, energy shortages, or security breaches. While some studies have examined blockchain and metaheuristic clustering (e.g., Glowworm Swarm Optimization and Cuckoo Search) independently for secure communication, very few have combined the two into a single framework optimized for 6G networks. Additionally, previous research has failed to adequately account for the diverse energy consumption patterns of IoMT devices, leading to wasteful resource utilization and premature network failures.

Blockchain, paired with 6G networks, could meet contemporary information security needs and inspire new digital economic and social paradigms, potentially having significant societal implications^15,16. Many technological-related difficulties and concerns preclude this convergence¹⁷. Among the problems are low latency, effective data transport, security and privacy concerns, as well as the scalability and performance of consensus techniques. Considering this backdrop, we aimed to develop a strategy that would maximize efficacy while implementing security policies simultaneously. This method seeks to effectively integrate blockchain technologies within a 6G network, thereby integrating and protecting them. The main objectives of this research are these:

• Introducing a hybrid metaheuristic optimization approach (ADCGRO) that balances the strengths of Cuckoo Search, Glowworm Swarm Optimization, and Remora Learning to optimally select cluster heads, considering device heterogeneity and residual energy.

• Integrating a lightweight blockchain consensus mechanism with the clustering process to ensure secure, transparent, and tamper-resistant data transmission without imposing excessive computational overhead.

• Developing a comprehensive system model that evaluates the framework across multiple performance metrics, including throughput, packet delivery ratio, energy efficiency, latency, accuracy, and network lifetime, all of which are validated through extensive simulations.

Simulations are necessary to assess the impact of the proposed security and optimization strategies on the network’s overall performance, thereby determining the viability of integrating blockchain-6G in practical applications.

This project aims to support the development of intelligent, safe, and high-performance communication systems by leveraging the complementary capabilities of blockchain and 6G networks. The following sets up the next work with shadows: "Related works" section provides the literature review; "System model" section presents a comprehensive overview of the proposed approach; "Simulation settings" section examines the results; "Conclusions and future work" section ends with a last summary.

Related works

Begum et al.¹⁸ presented an HLDAG-BC—a mix of blockchain technology and augmented recurrent neural networks to guarantee safe and effective routing in MANET-IoT systems. Identity-based conditional privacy-preserving authentication (ICPA) is a lightweight, secure node authentication system offered by the HLDAG-BC, which provides tamper-proof communication. The kernel neutrosophic c-means (KNCM) method helps the network’s nodes to be generally classified. Using the red piranha optimization (RPO) approach helps one to choose the best CH. Finding the most effective routing path to improve dependability and lower latency falls to RERNN. Intrusion detection using an HDLNN helps enhance overall network security. By achieving a 31.35% improvement in packet delivery ratio, a 34.56% increase in throughput, a 30.29% reduction in latency, and a 28.67% increase in network lifespan, the recommended HLDAG-BC-RERNN strategy outperforms current approaches. By providing a safe and scalable key for MANET-IoT networks, the proposed framework seems to be a practical option for potential 6G applications. FedChain, a novel federated blockchain-based clustering system designed to enhance connection and network security in cell-free massive MIMO (CF-mMIMO) FANETs, was first introduced by Pramitarianini et al.¹⁹.

Blockchain and federated learning (FL) help protect the cluster against Sybil attacks. As such, a safe cluster can be created without increasing the amount of control messages. Maximizing stable cluster formation with minimal control overhead depends on formulating the cost function maximizing issue using a cross-layer architecture under minimal influence overhead. Data from the physical layer (position, mobility, channel capacity, and residual energy) is combined in this design with data from the network layer (connectivity). Furthermore, selecting the best CHs based on parameters such as maximum residual energy and velocity restrictions ensures long-term stability. Inter-node transactions are authenticated using blockchain technology, which ensures secure development by separating Sybil attack nodes from real users.

This action helps to reduce the security concern. Through a unique FL framework, which can predict and identify node conditions in real-time without requiring additional control packets, security performance can be enhanced, and control overhead during cluster formation can be reduced. Regarding connection, control overhead, and security performance, the FedChain-based clustering method beats the lowest ID (LI), maximum connectivity degree (HCD), and standard blockchain-based clustering (CBC) protocols according to simulation results. The results show that the FedChain-based clustering system provides significant security and connectivity, making it a great option for dynamic settings. Using 6G technology and convolutional neural networks (CNNs) to identify Chinese characters in their natural surroundings, Li and Li²⁰ developed SE2ERS, a safe end-to-end recognition system. Using the Weighted Hyperbolic Curve Cryptograph (WHCC), the proposed SE2ERS idea combines blockchain technology with the secure data transmission capabilities of the 6G network design.

For character estimation, the computer vision system utilizes a data flow, specifically a 6G gradient directional histogram (GDH). By utilizing WHCC and GDH, the SE2ERS concept facilitates the safe transmission of data within the 6G network. The suggested SE2ERS analyzes conventional Chinese text recognition techniques and data transmission parameters to achieve 6G communication. With simple Chinese characters, the testing results show that SE2ERS achieves an accuracy of 88% and beats current approaches. Unlike prior approaches, which achieved a 72.8% success rate, our method achieves an average success rate of 84.4% using challenging Chinese characters. Furthermore, the WHCC model exhibits a twelve times higher data encryption rate complexity than more traditional methods, thereby significantly improving data security. By utilizing blockchain technology, Chinnasamy et al.²¹ propose a novel approach for enhancing security in 6G wireless networks. This project uses blockchain technology in a user sensor network. The second phase involves utilizing artificial democratic optimization to complete the network optimization. The results of the simulation have been affected by several network parameters, including accuracy, packet delivery ratio, end-to-end latency, throughput, and energy efficiency.

Furthermore, by selecting the optimal node and path for data flow, it can reduce network traffic, a valuable additional benefit. The suggested method achieved a 95% energy efficiency, 97% throughput, 50% end-to-end latency, 96% accuracy, and a 94% packet delivery ratio. Liu et al.⁷ investigate the integration of blockchain technology with 6G networks from the perspectives of security and performance, thereby proposing a technique that focuses on building secure mechanisms and improving performance. The potential applications of 6G networks and how blockchain technology can enhance them must be highlighted to underscore the need to incorporate blockchain technology into 6G networks. Therefore, it is essential to assess security needs in the construction of 6G networks to ensure the security of data flow and communication using blockchain technology. Furthermore, to address ways to maximize communication and storage to meet the needs of 6G networks, a Directed Acyclic Graph (DAG) solution is proposed for the asynchronous consensus mechanism used by blockchains. This will be implemented to maximize the integration of blockchain and 6G networks. Ultimately, it is equally important to provide valuable suggestions for future studies that will advance blockchain technology and facilitate the integration of 6G networks.

System model

This study employs various IoMT devices with distinct initial energy levels in monitoring scenarios utilizing 6G-enabled Internet of Things (IoT) capabilities. N 6G-enabled IoMT devices constitute the network, which comprises three levels of energy heterogeneity. Furthermore, Fig. 1 illustrates the methodology of the proposed prototype.

The 6G-enabled IoMT devices that comprise the network range from highly sophisticated to relatively simple, with varying degrees of initial energy. As a result of the energy disparity, advanced 6G-enabled IoMT devices start with more energy than intermediate 6G-enabled IoMT devices, which in turn have more energy than regular nodes $\:{E}_{nrm}<{E}_{int}<\:{E}_{adv}$. The overall number of every type of 6G-enabled IoMT devices is cutting-edge ($\:{n}_{\text{a}\text{d}\text{v}}$), intermediate ($\:{n}_{\text{i}\text{n}\text{t}}$), and normal ($\:{n}_{\text{n}\text{r}\text{m}}$) reflects a descending orders, with standard 6G-enabled IoMT devices being the most frequent, filling the inequality $\:{n}_{nrm}>\:{n}_{int}>\:{n}_{adv}$. The first energy for every 6G-enabled IoMT device types are represented as E0, and the overall energy in the network is signified by $\:{E}_{\text{t}}$. The progressive 6G-enabled IoMT devices have their energy described as $\:{E}_{0}\times\:(1\:+\:\omega\:)\times\:{n}_{adv}$, and intermediate 6G-enabled IoMT devices as $\:{E}_{0}\times\:(1+\varPsi\:)\times\:{n}_{int}$, with normal 6G-enabled IoMT device at $\:{E}_{0}\times\:{n}_{nrm}$. Therefore, the aggregate energy of all 6G-enabled IoMT devices—advanced, intermediate, and normal—makes up the network’s total energy $\:{E}_{T}={E}_{adv}+{E}_{int}+{E}_{nrm}$. The overall energy is $\:{E}_{T}={E}_{0}\times\:N\times\:(1+\varPsi\:\times\:{k}_{o}+k\times\:\omega\:)$, combining the energy contributions from all levels. The energy ratios $\:\varPsi\:$ and $\:\omega\:$ correspond to the energy fractions of the advanced and intermediate 6G-enabled IoMT device, correspondingly. The proportion of advanced and intermediate 6G-enabled IoMT devices within the network are indicated by $\:k$ and $\:{k}_{\text{o}}$. A diverse range of 6G-enabled IoMT devices, categorized as advanced, intermediate, and basic nodes, each with its unique energy capacity, are integrated into the network architecture. To ensure that energy-intensive processes, such as CH selection, are handled by competent devices, a hierarchy is established, where advanced nodes have the most energy reserves, intermediate nodes have moderate energy, and regular nodes have the least. Normal nodes are most numerous, followed by intermediate nodes, and advanced nodes make up the smallest group; this distribution also represents practical deployments. This network architecture allows the system to use high-energy nodes for mission-critical activities and lightweight devices for routine sensing. To optimize CH selection in 6G settings, an energy-aware setup is essential. It allows balanced energy consumption, extends the total lifespan of the network, and ensures optimal integration with blockchain for safe, low-latency data transfer.

Fig. 1

Workflow of the research classical.

There are several assumptions about the network setup of the suggested work, which are as follows:

• Although 6G can facilitate considerable mobility in IoMT devices, the proposed study presumes a static network configuration. In the context of network functioning, both the 6G-enabled IoMT devices and the data-gathering sink are considered stationary.

• Despite the differing energy capacity of nodes, it is anticipated that all 6G-enabled IoMT device within the network would maintain uniform configurations and operational characteristics. This assumption facilitates the study and use of the proposed strategy.

• Unlike IoMT devices, which are constrained by energy limitations, the data- gathering sink is anticipated to operate without such restrictions.

• The network’s IoMT devices are anticipated to be sensor nodes equipped with 6G communication capabilities, improving their connectivity and data transmission effectiveness.

The purpose of this endeavor is to utilize the same radio energy utilization model as prior studies, to focus on the transmission and receipt of data between 6G-enabled IoT devices and sink connections. Three elements define the architecture: the administrative domain, the network of the 6G-enabled IoMT, and the blockchain network. Utilizing 6G technology for fast and low-latency connectivity, devices located in homes, hospitals, and ambulances comprise the IoMT network. By utilizing smart contracts, the blockchain network ensures secure and transparent data management. Using this will enhance data security and facilitate tracking. Monitoring the network, the administrator ensures the domain achieves optimal performance and resolves any newly emerging issues. Gathering data, a “sink” sends it to the “blockchain” for processing. The IoT devices send their data to the sink. Monitoring the network and ensuring regulatory compliance helps the administrator keep the system running as intended.

Fitness evaluation

The primary objective of the suggested research is to reduce energy consumption in IoMT networks by leveraging 6G technology, thereby promoting sustainable communication in these systems. Achieving this goal depends on the CH’s performance in supervising the network’s energy consumption. The ADCGRO method produces an ideal answer, which guides the choice of CH. Establishing a comprehensive fitness function with several key elements will help determine the CH precisely. Each of these factors significantly impacts the CH choosing process. This all-encompassing solution ensures effective energy management in 6G IoT systems. This work sought to integrate four different fitness metrics: residual energy, node-to-sink closeness, cluster energy, and remaining node energy levels. This comprehensive approach ensures a more thorough and efficient decision-making process.

• (1) residual energy

  Residual energy is a crucial fitness parameter for the CH since its capacity to gather and forward data to the sink consumes more energy than that of the other nodes in the cluster. Choosing a particularly active CH is therefore rather crucial. The energy levels of the nodes will progressively decrease as the network operates, leaving a varying amount of energy remaining at the end of each round. This is assessed concerning the first fitness factor,$\:{F}_{1}$

  $${F}_{1}=\sum\:_{K=1}^{N}\frac{{\epsilon\:}_{resid}\left(K\right)}{{\epsilon\:}_{init}\left(K\right)}$$

  (1)

  which assesses the rate of the node’s residual energy (${E}_{\text{r}\text{e}\text{s}\text{i}\text{d}}$) to its primary energy (${E}_{\text{i}\text{n}\text{i}\text{t}}$).

• (2) node and sink proximity

  The distance sensor nodes must travel to interact with either other nodes or the sink, which is a fundamental component of this work. Especially in determining the CH, evaluating a node’s proximity to the sink is crucial. Selecting CH, the node closest to the sink, will help to drastically lower energy use. By evaluating the distance between the sink and the node (${D}_{\text{S}\text{N}\text{o}\text{d}\text{e}-\text{S}\text{i}\text{n}\text{k}}$) in respect to the average distance between all nodes and the sink (${D}_{\text{A}\text{v}\text{g}(\text{S}\text{N}\text{o}\text{d}\text{e}\text{s}-\text{S}\text{i}\text{n}\text{k})}$), this fitness parameter guarantees the network runs efficiently and cheaply.

  $$F_{2} = \sum\limits_{{K = 1}}^{N} {\frac{{D_{{{\text{SNode}} - {\text{Sink}}}} \left( K \right)}}{{D_{{{\text{Avg}}\left( {{\text{SNodes}} - {\text{Sink}}} \right)}} \left( K \right)}}}$$

  (2)

• (3) remaining energy levels of nodes

  Our approach depends critically on selecting the CH from among the remaining live nodes, as, once network operation starts, the network’s energy starts to drop, leading to an increasing number of dead nodes over period. As seen below, to define a fitness parameter.

  $$F_{3} = \frac{1}{N}\sum\limits_{{K = 1}}^{N} {{\mathcal{E}}_{{resid}} \left( K \right)}$$

  (3)

  This determines the amount of energy that persists in the network. This threshold must be maintained at an elevated level for effective CH selection. Thus, the most energy-efficient nodes will be selected, guaranteeing the network’s operational continuity and durability.

• (4) cluster’s energy

  As the energy of the nodes within a cluster diminishes, the aggregate energy level of the cluster concurrently declines. The energy condition of the cluster housing the candidate node must be considered during the CH selection process for the network to operate effectively. This component is initially designated as 1 in the first round, as cluster formation transpires after the selection of the CH. Conversely, cluster energy level emerges as a significant determinant in CH selection beginning with the second round. To compute $\:{F}_{4}$, the fifth fitness parameter, one must adhere to the following steps.

$$F_{4} = \sum\limits_{{K = 1}}^{{N_{{CH}} }} {{\mathcal{E}}_{{resid}} \left( K \right)}$$

(4)

To provide effective energy management and balanced load distribution across the network, the node density must be high, allowing a suitable candidate node to be selected as the CH.

Our method involves combining multiple fitness criteria to convert a function into a single-objective one. To achieve this, we multiply these parameters using linear weight functions, as described in our technique. This integration makes it easier to evaluate the system’s performance while still considering all the important factors.

$$\:F=a{F}_{1}+\beta\:{F}_{2}+\lambda\:{F}_{3}+\delta\:{F}_{4}$$

(5)

where $\:\alpha\:+\beta\:+\lambda\:+\delta\:=1$. It provides a weighted aggregate of multiple separate elements, where each element is multiplied by its corresponding characteristics. The suggested method is to evaluate each node based on the value of its fitness function. The node with the greatest fitness value upon completion of this calculation is designated as the CH.

Cluster head selection using artificial democratic cuckoo glowworm remora optimisation (ADCGRO)

Three separate groups constitute an artificial bee colony: onlooker bees, employed bees, and scout bees. The bee’s inherent curiosity enables it to identify food sources, which it then shares with other nearby bees. Employing the information relayed by the foraging bee, the observer bee surveys its surroundings to locate a food source that better meets its needs. The observing bees will transition to scout bees and leave the food source when the food supply reaches a specific threshold. Expansion of the food supply will result in this occurrence. The method by which bees search for new food sources is chaotic. As per Eq. (6), scout bees are responsible for monitoring the exploration procedure, whereas paid and bystander bees are tasked with exploiting the ABC²²:

$$\:{v}_{ij}={x}_{ij}+{\varphi\:}_{ij}({x}_{ij}-{x}_{kj})$$

(6)

The new location of the i-th worker bee is represented by $\:{v}_{\text{i}}$, whereas the former position is indicated by $\:{x}_{\text{i}}$. The variable k represents an arbitrary integer ranging from 1 to the total count of bees utilized. Each issue possesses an arbitrary dimension represented by j, and the variable $\:{\varphi\:}_{ij}$ is a randomly assigned value within the range of − 1 to 1. The vigilant bee uses Eq. (7) to ascertain which items to consume, taking into account the likelihood of each worker bee’s actions:

In addition to being governed by Eq. (8), fit represents the fitness values of the worker bee’s answer.

$$\:{p}_{i}=\frac{{fit}_{i}}{\sum\:_{j=1}^{SV}{fit}_{j}}$$

(7)

fit is fitness values of the working bee’s answer, Eq. (8) is utilized to regulate it.

$$fit = \left\{ {\begin{array}{*{20}l} {\frac{1}{{1 + f\left( {x_{i} } \right)}},} \hfill & {f\left( {x_{i} } \right) \ge \:0} \hfill \\ {1 + \left| {f\left( {x_{i} } \right)} \right|,} \hfill & {ff\left( {x_{i} } \right) < 0} \hfill \\ \end{array} } \right.$$

(8)

Bees employ it to enhance their understanding of the chosen food supply, whereas worker bees utilize it for navigation purposes (9). The crucial decision in optimizing with the Cuckoo algorithm is the formulation of the goal function. Depending on the objective function, it is feasible to optimize for scheme efficiency or to reduce scheme errors. This study aims to identify the most efficient approach for mitigating errors in system or cost functions. The cost function (fc) may be computed using the following equations to calculate the integral of square error (ISE):

$$\:e\left(t\right)=Pmax-P\left(t\right)$$

(9)

$$\:ISE={\int\:}_{0}^{\infty\:}{e}^{2}\left(t\right)dt$$

(10)

Like other optimization strategies, the cuckoo technique commences with an initial population, whereby each cuckoo occupies a unique spatial location. The optimization findings will get more accurate as the population of cuckoos rises. As the density of the cuckoo population increases, the convergence rate diminishes. Nonetheless, even with offline parameter optimization, the convergence rate remains invariant. Consequently, enumerating the cuckoos that will deposit eggs is the initial essential phase. The variables of the problem are presented using an array. I will refer to these arrays as “habits” when implementing the Cuckoo approach. This issue space encompasses several habitats, each proposing a distinct viable answer. The variable values pertinent to the situation must be encapsulated within an array. The “Habitat” array, an element of the cuckoo optimization process, serves this purpose. In addressing an optimization problem, the subsequent position of the habitat, indicated by the symbol Nvar, is determined by the cuckoo’s current location, expressed as 1 * Nvar. Equation (11) defines these components.

$$\:Nvarhabitat=\left[{x}_{1},{x}_{2},{x}_{2},\dots\:,xNvar\right]$$

(11)

The parameter values ($\:{x}_{1},{x}_{2},{x}_{2},\dots\:,xNvar$) each characterises the sum points. The profit functions habitat is assessed to determine (Eq. 12).

$$\:profit={f}_{p}\left(habitat\right)={f}_{p}({x}_{1},{x}_{2},{x}_{2},\dots\:,xNvar)$$

(12)

The cuckoo strategy utilizes a notably superior profit function. To incorporate this strategy into the minimization algorithm, it is essential to apply a negative sign to the cost function. The GSO algorithm’s mechanisms were inspired by the natural behavior of glowworms, which are attracted to brighter sources of light. To maximize multimodal functions, swarm intelligence (SI) methods, such as the GSO technique, operate in a manner analogous to that of glowworms. Swarm intelligence algorithms may effectively tackle complicated optimization problems that prior methods are either incapable of addressing or unable to resolve comprehensively. The typical procedure is impractical. In the realm of statistics, several algorithms provide numerous advantages and disadvantages. You may be able to observe glowworms during a nocturnal stroll in the meadows. Throughout the night, these timeless beings shimmer exquisitely, illuminating the darkness. All glowworms possess a light substance called luciferin. Consequently, every glowworm will radiate light, irrespective of the surrounding darkness or cave environment. A decision range and a sensor range define the community, allowing any two glowworms in the area to interact with each other.

For the sake of reproduction and nutrition, glowworms relentlessly scour their environments for glowing luciferin. The function value that depends on the location of the glowworm has a substantial impact on the luciferin updating phase. The luciferin level of each glowworm is adjusted by increasing its previous value. While the glowworm is in its positionally effective state, this change occurs. A portion of its value is subtracted from its total value to mimic the steady deterioration of luciferin. The following suggestions are offered for updating glowworm luciferin using Eq. (13):

$$\:{l}_{i}\left(t+1\right)=\left(1-\rho\:\right){l}_{i}\left(t\right)+yJ\left({x}_{i}\right(x+1\left)\right)$$

(13)

where $\:\left(xi\right(t\left)\right)$ where li(t) is the luciferin level of glowworm i at time t, p is the decay constant (0 < p < 1), y is the constant, and is the value of the objective function at agent i’s position at time t. The GSO principle dictates that the agent or glowworm will approach or encircle a neighbour with a higher luciferin level than its own. During this phase, the glowworm will also use the probabilistic approach given by Eq. (14) to travel to a neighbour with a higher intensity:

$$\:{N}_{1}\left(t\right)=\left\{j:{d}_{ij}\left(t\right)<{r}_{d}^{i}\left(t\right);{l}_{j}\left(t\right)<{l}_{j}\right\}$$

(14)

The collection of time t is Eq. (14). A sensor variety that bounded $\:(0<rd\:i<\:rs)$ surrounds the capricious. Equation (15) elasticities the probability of neighbour $\:j\in\:{N}_{i}\left(t\right)$ for each glowworm $\:i$.

$$\:{P}_{ij}=\frac{{l}_{j}\left(t\right)-{l}_{i}\left(t\right)}{{\sum\:}_{k\in\:{N}_{i}\left(t\right)}{I}_{k}\left(t\right)-{I}_{i}\left(t\right)}$$

(15)

Equation (16) then characterizes the glowworm’s movement model.

$$\:x_{i} \left( {t + 1} \right) = x_{i} \left( t \right) + s\left[ {\frac{{x_{j} \left( t \right) - x_{i} \left( t \right)}}{{\left\| {x_{j} \left( t \right) - x_{i} \left( t \right)} \right\|}}} \right]$$

(16)

where || || where s is the step size and is an operator for the Euclidean norm. Then, in the m-dimensional real space R m, the position at time t is denoted by $\:{x}_{i}\left(t\right)\in\:{R}_{m}$. Equation (17) illustrates a neighbor range update phase used to identify multiple peaks in a landscape of multimodal functions. Then, set V_0 to be the initial range charge for each glowworm. The subsequent rule is used to update each glowworm’s neighbourhood variety:

$$\:{r}_{d}^{U}\left(t+1\right)=min\left\{{r}_{s},max\left\{0,{r}_{d}^{U}\left(t\right)+\beta\:\left({n}_{t}-\left|{N}_{l}\left(t\right)\right|\right)\right\}\right\}$$

(17)

If there is less than a 4-meter gap between a Grey Wolf (GW) and its nearest neighbour, the GW is allowed to roam. Additionally, the coverage rate may be slower with a smaller step size. Finding the optimal step size is thus a challenging task. This investigation will not have a constant step size; rather, it may vary with each cycle according on the GW. There will be several adjustments to GW’s step size. One of these is the introduction of a dynamic step size, which should increase computational precision in velocity stage. What follows is an explanation of the ROA, or remora optimisation method.

Unrestricted Journeys: The location update mechanism for the SFO strategy was developed using Eq. (18), which produces the algorithm’s elite idea.

$$\:{R}_{i}^{t+1}={R}_{best}^{t}-\left(rand\times\:\left(\frac{{R}_{best}^{t}-{R}_{rand}^{t}}{2}\right)-{R}_{rand}^{t}\right)$$

(18)

where $\:{R}_{rand}^{t}$ is a random location.

An experience attack works in a manner comparable to how knowledge grows: it checks in with the host regularly to determine if it needs to be updated. To represent these ideas, to use the following expression in Eq. (19):

$$\:{R}_{att}={R}_{i}^{t}-\left({R}_{i}^{t}-{R}_{pre}\right)\times\:randn$$

(19)

where $\:{R}_{att}$ is a step and $\:{R}_{pre}$ is the prior iteration’s site, distinct by Eq. (20).

$$\:f\left({R}_{i}^{t}\right)>f\left({R}_{att}\right)$$

(20)

Equation (21) designates updated formulas with alterations.

$$\:{V}_{i}\left(t+1\right)=D\times\:{e}^{a}\times\:cos\left(2\pi\:a\right)+{X}_{best}\left(t\right)$$

(21)

$$\:D=\left|{X}_{best}\left(t\right)-{X}_{i}\left(t\right)\right|$$

(22)

$$\:a=rand\times\:\left(b-1\right)+1$$

(23)

$$\:b=-\left(1+\frac{t}{T}\right)$$

(24)

where D is the distance between the remora and the meal. To further enhance the solution, the remora can utilize the encircling prey mechanism in WOA to take a small step, as illustrated in Eq. (25).:

$$\:{X}_{i}\left(t+1\right)={V}_{i}\left(t+1\right)+A\times\:D{\prime\:}$$

(25)

$$\:A=2\times\:B\times\:rand-B$$

(26)

$$\:B=2\times\:\left(1-\frac{t}{T}\right)$$

(27)

$$\:{D}^{{\prime\:}}={V}_{i}\left(t+1\right)-C\times\:{X}_{best}\left(t\right)$$

(28)

where the i-th remora’s freshly created site is characterised by $\:{X}_{i}\left(t+1\right)$. The value of the remora factor, denoted as C in ROA, is fixed at 0.1. Finding its food is something the remora is capable of. With this in mind, to enhance the remora’s search efficiency by including a novel autonomous foraging mechanism into the core algorithm. A fair balance between exploration and exploitation is achieved by the upgraded remora optimisation algorithm, which features a more adaptive mode. Note that this approach does not increase the computational complexity of the original program. Since the proposed approach is compatible with various optimisation algorithms, it has some generalisability.

Blockchain integration

The proposed approach significantly enhances data security and integrity by integrating blockchain technology with IoT environments facilitated by 6G²³. The fundamental concept of the system’s blockchain layer is a permissioned blockchain network utilizing nodes with substantial computational capacity within the network. This indicates that pre-authentication has occurred and that all nodes inside the network have developed a certain level of trust among themselves. The nodes employ Proof of Authority (PoA), a streamlined consensus mechanism. The technique depends on a consortium of trustworthy nodes, referred to as authorities, to authenticate transactions and produce blocks. At each instance when data is transmitted from the CH to the primary blockchain layer, a blockchain node should validate the transaction, authorize it, and generate a new block. All the blockchain nodes in the network receive the newly created block, making maintenance easier. There are two main components to a block on this blockchain: the header and the transaction data. This block’s version, genesis timestamp, nonce, difficulty target, Merkle root, and hash of the previous block are all essential pieces of information that are included in the header. The hash of the preceding block is included in the hash of the current block, ensuring the integrity of the blockchain and rendering it immutable. Altering the content of one block will initiate a domino effect, resulting in changes to the hashes of succeeding blocks. Secure, decentralized data exchange and automated decision-making among diverse devices and service providers are two ways in which blockchain is facilitating the adoption of intelligent, context-aware services in 6G networks. Without relying on a centralized authority, 6G apps such as smart healthcare, autonomous vehicles, or adaptive content distribution may utilize smart contracts to negotiate resources like bandwidth and processing power dynamically. They can also make choices based on real-time context. Blockchain technology ensures the integrity, immutability, and auditability of all contextual data, including user movement patterns, network load, and service priority. This enables the network’s artificial intelligence (AI) agents to make reliable predictions and adaptive improvements. For instance, in a 6G smart city network, blockchain technology can enable vehicle edge nodes to safely exchange traffic data and establish low-latency priority lanes for emergency vehicles using smart contracts, ensuring that decisions are made transparently and securely.

A variety of verifications are employed to guarantee the accuracy of each block. The verifier initiates the process by assessing if the current value exceeds its predecessors and confirming the presence and legitimacy of the preceding block. Furthermore, it has been verified that additional components, such as the transaction root and difficulty, are authentic. All transactions within the block undergo independent verification. An error notice appears if any transaction fails. If all transactions have been properly authenticated and the final state matches the current root state, the block is deemed validated. A formula exists to compute the hash of a block:

$$\:Hash\left({B}_{n}\right)=Hash(Hash\left({B}_{n-1}\right)+{B}_{nData})$$

(29)

where Bn signifies the block for which the hash is being calculated, $\:Hash\left({B}_{n-1}\right)$ is the prior block’s hash, and $\:{B}_{nData}$ represents the current block data.

The utilization of the current block in hash calculations significantly bolsters the blockchain’s durability against assaults. Implementing data modifications in any block becomes exceedingly challenging as the number of blocks in the chain increases significantly. Should a single block be altered, it would create a discrepancy in the hashes of all following blocks, jeopardizing the whole blockchain and facilitating manipulation. In this method, cluster chiefs diminish the transaction volume sent to the blockchain and alleviate the processing burden on nodes by consolidating data from IoMT sensor devices. A permissioned blockchain utilizing Proof of Authority (PoA) consensus, with a limited number of pre-approved validators, might expedite transaction validation, reduce latency, and improve throughput compared to conventional methods such as Proof of Stake (PoS) or Proof of Work (PoW). Scalability and efficacy for IoT devices are attained by mitigating network congestion and energy consumption through the restriction of the total number of validator nodes. By replacing trust in centralised authority with a distributed, immutable ledger, blockchain technology enhances 6G decentralized authentication and access management. Decentralized consensus systems, instead of a central server, manage authentication for each device, user, or service node by assigning them a unique cryptographic identity and storing it on the blockchain. By automatically enforcing role- or attribute-based access controls, smart contracts can limit access to 6G services and network slices to approved organizations exclusively. For example, a blockchain can be utilized in an IoMT system enabled by 6G to restrict access to patient data to authorized healthcare practitioners and to record all transactions for transparent audit purposes. In addition to improving reliability and security, this eliminates frequent problems associated with conventional methods, including centralized authentication methods such as identity spoofing and single points of failure.

Simulation settings

A region covering 100 square meters by 100 square meters employs between 100 and 200 sensors, connected to various “gateways,” to assess the suggested methodology. This evaluation utilizes Network Simulator-2. Initially, the points got 100 joules of energy, and the gates received 200 joules. It is recommended to incorporate a dynamic threshold generation method directly into the operation, rather than appending a predefined threshold value to the detection process²⁴. The parameters under examination are likely to evolve. This is due to their vulnerability to physical influences such as temporal variations and seasonal fluctuations. Although temperatures may increase by midday, they are frequently lower in the morning and evening²⁵. To minimize the incidence of false positives, it is recommended that the threshold be adaptable. The sliding window theory is utilized to ascertain value. Forty samples related to the sliding window were collected for analysis. A determination concerning the cutoff is made by employing the final forty samples. The information shown in Table 1 is the same. The performance of the proposed ADCGRO-blockchain framework was evaluated against six state-of-the-art methods: KNCM, WHCC, DAG, LEACH, HEED, and a PBFT-based blockchain model. Metrics including throughput, energy efficiency, packet delivery ratio, authentication accuracy, latency, and network lifetime were analyzed across 1,000 simulation rounds in heterogeneous 6G-enabled IoMT networks.

Table 1 The simulation limits of the proposed scheme.

Validation analysis of the suggested classical with existing procedures

Table 2 presents the proposed model’s probability of attacks in various metrics. The proposed model consistently outperformed all baselines, achieving average energy efficiency gains of 18%, throughput improvements of 16%, and authentication accuracy exceeding 99%, compared to 82–88% for the next best methods (DAG and WHCC). Notably, network lifetime increased by 27% over DAG and by 35% compared to LEACH, highlighting the effectiveness of ADCGRO’s energy-aware CH selection and lightweight blockchain consensus.

To confirm that these results were not due to random variations, paired t-tests were performed between the proposed framework and each baseline. All improvements were statistically significant with p-values < 0.05. For example, the network lifetime improvement over DAG (27%) was observed with a 95% confidence interval of ± 3%, while the latency reduction (35%) over HEED had a confidence interval of ± 2%. These results confirm the robustness of the proposed model.

Table 2 Experimental analysis of the suggested model’s attack.

The experimental analysis of the proposed model’s attack investigates two critical factors: Rr (Response Rate) and Rc (Response Cost). This inquiry examines the influence of varying assault probability on these metrics. As the likelihood of an assault escalates from 0.1 to 1, both Rr and Rc exhibit a steadily rising trajectory. This holds for each of these metrics. The expense of the answer is 111.21, the response rate is 394.8, and the probability is 0.1. Nevertheless, these values escalate to 1195 and 303.65, respectively, when the likelihood of an assault reaches its maximum, which is 1. The presence of this trend indicates that as the number of assaults escalates, the system is compelled to address increased reaction costs and response rates. This highlights the need for enhanced security measures to mitigate the impact of attackers.

Comparative analysis of the proposed model

Table 3 presents the experimental investigation of the proposed model in comparison to existing procedures, evaluated using four metrics: throughput and accuracy. Figure 2 shows the analysis of proposed model on number of clusters. Figure 3 displays the visual analysis of various models on number of users. Figure 4 exhibits the graphical Comparison of various models on number of base stations. One important statistic is latency, as 6G URLLC services require end-to-end delays to be less than milliseconds, even with blockchain’s additional consensus procedures. Both throughput and energy efficiency are critical. Throughput measures the system’s capacity to maintain multi-gigabit data rates throughout huge IoT traffic, while energy efficiency guarantees that blockchain validation operations do not unnecessarily drain the limited power of edge devices. For security reasons, authentication and access accuracy measure the system’s capacity to reject valid nodes without allowing unwanted access, while transaction finality time shows how fast the blockchain commits and validates data, which affects real-time service delivery. Balanced traffic loads, as a function of network resource consumption, assess the efficacy of spectrum allocation and smart contract-driven routing. Data integrity score, attack detection ratio, and other security robustness measures evaluate the 6G system’s resistance to spoofing, Sybil, and double-spending attacks.

Table 3 Comparative analysis of various algorithms.

Fig. 2

Analysis of proposed model on number of clusters.

Fig. 3

Visual analysis of various models on number of users.

Fig. 4

Graphical comparison of various models on number of base stations.

The Comparative Analysis of Several Algorithms aims to evaluate the performance of several methods—specifically KNCM, WHCC, DAG, and the Proposed model—regarding accuracy and packet delivery ratio (PDR) across three distinct scenarios: varying the total number of stations. The suggested model achieves a throughput of 86%, an energy efficiency of 90%, a packet delivery ratio of 82%, and an accuracy of 92%, surpassing all previous methodologies. In total, there are ninety-two clusters. The suggested approach achieves values of 88, 87, 83, and 94, hence sustaining a competitive edge and surpassing competing strategies. A further edge over competitors is the size of the user base. The proposed model outperforms DAG, WHCC, and KNCM in terms of accuracy, throughput, and energy efficiency, yielding superior results across all metrics. It provides 97% accuracy, 95% PDR, 98% throughput, and 96% energy economy. Ultimately, the proposed model consistently yields superior outcomes across various experimental settings, validating its efficacy in enhancing network performance.

Fig. 5

Performance comparison.

Figure 5 presents a comparative analysis of the proposed ADCGRO-blockchain framework against KNCM, WHCC, DAG, LEACH, HEED, and PBFT-based blockchains in terms of throughput, energy efficiency, packet delivery ratio (PDR), and accuracy. The proposed model consistently outperforms all baselines, achieving 95% throughput, 92% energy efficiency, 90% PDR, and 99% accuracy, which represent improvements of 16–30% over conventional clustering protocols like LEACH and HEED, and 8–20% over advanced schemes like DAG and PBFT. These gains result from the energy-aware CH selection of ADCGRO, which balances network load, and the lightweight blockchain integration, which secures communications without introducing significant overhead. Collectively, these results validate the model’s ability to deliver reliable, energy-efficient, and secure performance in heterogeneous 6G-enabled IoMT networks.

Fig. 6

Confusion matrix.

Figure 6 shows the confusion matrix. Out of 2000 authentication attempts, the system correctly classifies 99.3% of legitimate nodes and maintains a False Positive Rate (FPR) of less than 1%, confirming the reliability of the security layer. This result demonstrates the proposed scheme’s ability to minimize unauthorized access while maintaining seamless network operations.

Fig. 7

Time-series trends—energy depletion and packet delivery ratio (PDR).

Figure 7 illustrates the time–series trends for energy depletion and Packet Delivery Ratio (PDR). The proposed scheme shows a gradual, balanced energy consumption pattern, retaining operational energy beyond 1000 rounds, whereas LEACH and HEED networks deplete energy significantly by round 600. This trend validates the effectiveness of ADCGRO’s load-aware CH rotation strategy in prolonging network lifetime. The proposed framework maintains a PDR above 85% throughout the simulation, whereas LEACH and HEED degrade to below 50% by round 800 due to unstable clustering and higher packet losses. This demonstrates the proposed system’s robustness and reliability in extended operations, which is critical for real-time IoMT applications in 6G environments.

Thousands of sensor nodes with diverse mobility patterns, traffic volumes, and energy capacities will be utilized in future assessments to evaluate the scalability and usability of the proposed IDS architecture within various Mobile Sensor Networks. It remains to be seen how well the framework performs in extremely large-scale and dynamic situations, even if the present research demonstrates its advantages in accuracy, energy efficiency, and network longevity within modestly scaled 6G-enabled IoMT scenarios. The IDS will be tested in various scenarios to assess its performance in cluster formation, intrusion detection, and blockchain-secured communication, with minimal energy overhead and low latency. These scenarios include smart city infrastructures, disaster recovery environments, and vehicular sensor networks. To achieve this, the system will incorporate hierarchical clustering techniques and adaptive decision-making capabilities powered by machine learning. This will allow for the dynamic allocation of resources, prediction of intrusion behaviors, and resilience in the face of rapidly changing network topologies.

Conclusions and future work

This study presented a novel ADCGRO-based CH selection framework integrated with blockchain to enhance the energy efficiency, security, and reliability of heterogeneous 6G-enabled IoMT networks. By combining metaheuristic optimization for balanced clustering with lightweight blockchain consensus, the proposed model achieved substantial improvements, including a 27% increase in network lifetime, a 35% reduction in latency, and 99% authentication accuracy compared to LEACH, HEED, DAG, and PBFT-based models. While the results confirm the framework’s ability to enable safe, eco-friendly, and scalable communication for next-generation IoMT systems, certain limitations remain. The current evaluation is limited to static IoMT deployments, utilizing a simplified consensus protocol that overlooks scalability and security trade-offs under high transaction loads. Furthermore, the metaheuristic clustering approach does not yet leverage adaptive deep learning (DL) methods, which could enhance responsiveness to dynamic network conditions. Future research will focus on scaling the framework to large heterogeneous IoT deployments, integrating advanced neural network models such as reinforcement learning and federated DL for adaptive optimization, exploring blockchain techniques like sharding and hybrid consensus to balance security and efficiency, and validating the framework’s extensibility across other domains, including smart transportation and industrial IoT. Addressing these directions will enable the framework to evolve into a fully adaptive, secure, and scalable solution for managing diverse 6G-enabled IoT environments.