In wireless body sensor network (WBSN), the set of electrocardiogram (ECG) data which is collected from sensor nodes and transmitted to the server remotely supports the experts to monitor the health of a patient. While transmitting these collected data some adversaries may capture and misuse it due to the compromise of security. So, the major aim of this work is to enhance secure transmission of ECG signal in WBSN. To attain this goal, we present Pity Beetle Swarm Optimization Algorithm (PBOA) based Elliptic Galois Cryptography (EGC) with Chaotic Neural Network. To optimize the key generation process in Elliptic Curve Cryptography (ECC) over Galois field or EGC, private key is chosen optimally using PBOA algorithm. Then the encryption process is enhanced by presenting chaotic neural network which is used to generate chaotic sequences or cipher data. Results of this work show that the proposed cryptography algorithm attains better encryption time, decryption time, throughput and SNR than the conventional cryptography algorithms.

WBSN alludes to networking technology that interconnects many sensor nodes in or on the human body. It very well may be utilized in the use of medical care for persistent monitoring of patients [

WBSN can sense numerous physiological signals like electromyogram (EMG), electroencephalogram (EEG), electrocardiogram (ECG), internal heat level, pulse and surprisingly breathing or movement of the patients. Among all, ECG is essential as a result of its capacity to analyze cardiovascular illnesses that is the significant reason for deaths according to the report of WHO [

To attain this goal, the following contributions are presented in this paper.

In this approach, the sensed ECG signals are transmitted securely with the proposed optimized elliptic Galois cryptography with chaotic neural network.

For public key generation, Elliptic Curve Cryptography (ECC) over Galois field is presented, where prime numbers are considered as private key. For optimal private key selection, pity beetle swarm optimization algorithm (PBSOA) is presented.

With the generated public key, encrypted data is created using chaotic neural network.

The performance of the proposed scheme is analyzed in terms of encryption time, decryption time, throughput and network lifetime.

Upcoming sections of the manuscript are sorted as follows. Some recent literatures which focused research on secured ECG signal transmission in WBSN are survived in Section 2. Section 3 presents PBOA algorithm based EGC with chaotic neural network for secure ECG data transmission in WBSN. Results of the proposed method are analyzed in Section 4. Finally, the conclusion of the research work is discussed in Section 5.

Some recent articles which focused research on secured ECG signal transmission in WBSN are reviewed in this section. Qiu et al. [

Karthikeyan et al. [

Sanivarapu et al. [

For this work, MIT-BIH Normal Sinus Rhythm dataset is used. This dataset incorporates 18 long haul ECG recordings of subjects alluded to the Arrhythmia Laboratory at Boston’s Beth Israel Hospital (presently the Beth Israel Deaconess Medical Center). Subjects added for this data set were found to have had no critical arrhythmias; they incorporate 5 men, matured 26 to 45, and 13 ladies, matured 20 to 50.

The EGC algorithm is the extension of Elliptic Curve Cryptography (ECC) i.e., ECC over Galois field (GF). EGC is the public key cryptography where public key or secret key is generated for the user’s authentication. For authorization, public and private keys are more significant. In this approach, elliptic curve over Galois field (_{G}

For effective calculation, the value of Galois field is higher than one. Galois field consists of the infinite field range that defined as follows,

Galois field is defined as it is an infinite field with the number of integers as well as it is the integer with modulo prime field denoted as ‘_{G}

According to the

As the private key (R) is considered as random prime number for secret key generation in EGC, throughput of the algorithm may decrease. Thus, to enhance the performance of key generation in EGC, private key is selected optimally. For optimal private key selection, rain optimization algorithm (ROA) is presented in this paper. Following section explains the private key selection.

PBOA is presented for choosing the optimal private key or prime number. As this algorithm can search large areas of suitable solutions for the best global solution while solving local optima, it is chosen for selecting the private key optimally.

PBOA is executed reliant upon the behaviour of bark beetle. It shows the behaviour of polygamous mating, with the male mating with 3 to 6 females. The male beetles drill into the phloem of successfully incapacitated trees uncovering there a material chamber. While reinforcing, they change have terpenes into pheromones, pulling in females, with which they mate in the wedding chamber. From this chamber, in a star-like game-plan, females store 40–70 eggs in egg claims to fame. In this calculation, a populace contains males and females; a some of folks go about as pioneer particles that mission for the most sensible host. A numerical definition of this PBOA is depicted as follows.

Here,

^{th} dimensional. Besides, it is depicted as,

Here, RST [] represents an acronym as random sampling method and is applied to solve the problem of premature convergence. Using this RST, the positions of search space are defined into N_{pop} samples. U and L denote the global bounds.

^{th}

According to

_{p}) and used as a parameter of PBOA. According to every example, N_{pop} new pioneer particles are haphazardly situated into this inquiry territory by methods for RST:^{th} birth solution vector. ^{th} ^{th} generation step respectively.

As the beetle’s behaviour depicted already, five kinds of new hypervolume determination examples are executed into the algorithm. These determination examples are depicted as pursues:

_{ne}) that is utilized to characterize the region size is a parameter of the PBOA, the worth scope of f_{ne} is equivalent to [0.01, 0.20]. As indicated by this example, the new N_{pop} pioneer particles are arbitrarily situated into this search space dependent on the declaration of _{p} is equivalent to f_{ne}.

These generated new Npop solutions are contrasted with one another to characterize the best pioneer particle. The best is then contrasted with the beginning position, and if the rest are better drawn in, new populations will be made as follows:^{th} population, N_{broods} denotes the maximum count of broods to terminate.

_{ms}) is used to characterize this search area size, the worth scope of f_{ms} is equivalent to [0.10, 1.00]. As per this example, the new N_{pop} pioneer particles are haphazardly situated to this search space dependent on the outflow of _{p} is equivalent to f_{ms}.

Like the neighboring search hypervolume, these as of late portrayed Npop populations are appeared differently in relation to each other to describe the best pioneer particle. The best can compare the starting position, and it attracts others better and new people are generated.

_{ls}) that is utilized to characterize the size of this territory, the worth scope of f_{ls} is equivalent to [1, 100]. As indicated by this example, the new N_{pop} pioneer particles are haphazardly situated into this search space dependent on the outflow of _{p} is equivalent to f_{ls}.

Similar to the previous search hypervolume patterns these newly defined N_{pop} solutions are compared with each other for defining the best pioneer particle.

_{un}), which dictates the use of the global search hypervolume method, are defined with reference to the total function evaluations (FE_{total}) using the algorithm’s multiplier factor (f_{FE}). The worth scope of f_{FE} is equivalent to [0.05, 0.25]. As indicated by this example, the new N_{pop} pioneer particles are haphazardly situated inside this search space with the utilization of RST:

For the situation that a superior position is distinguished, as per memory consideration search hypervolume design, a limited neighbourhood search is performed by a fine-tuning factor (f_{tn}) utilized to characterize the size of this space. The worth scope of f_{tn} is equivalent to [0.005, 0.05].

(Mid-scale search hypervolume)

Where,

For encryption phase in EGC, chaotic neural network is presented. Chaotic neural network is used to enhance the security and performance of the algorithm. In this network, the plain text is converted into the cipher text using the ECC over GF secret key.

Initially, the input plain signal

Here, m = 1, 2,…., m

Using this binary sequence, the function of weight factor is defined. This weight factor can be varied depend on the input functions. The generation of weight factor is defined as follows,

Here,

Third, weight factor generation is followed by the bias function generation. For each chaos, bias function is generated using the weight factor. This bias function solves the problem of singularity. It is defined as follows,

At final, cipher signal is generated using the input function and weight factor. The generation of cipher is defined as follows,

After the completion of encryption, decryption phase will be processed at the receiver. The following section describes the phase of decryption of input signal.

The encrypted data will be decrypted at the receiver or medical server. In the phase of decryption, binary sequences are generated for the chaotic sequences and these binary sequences are given as input to the neural network. Also, the neural network generates the weight factor and bias function for the input function. Finally, decipher signal is generated with the ECC over GF secret key.

The proposed scheme is simulated in the platform of MATLAB with the system has the operating system of windows ‘10 with 64 bit and with 4GB main memory at 2 GHz dual-core PC. In this work, MIT-BIH Normal Sinus Rhythm dataset is used. For ECG signal compression, compressive sensing technique is used as well as Sparsity Adaptive Matching Pursuit algorithm (SAMP) is used for decompression. For secure transmission, PBOA based ECC over Galois field (EGC) with chaotic neural network (CNN) is presented.

The performance of the proposed EGC-PBOA-CNN is analysed in terms of encryption time, decryption time, energy consumption, throughput and network lifetime. Besides, the performance of the proposed scheme is compared with that of the EGC-CNN, EGC and ECC. The following sections describe the performance analysis of the different cryptography algorithms in terms of different metrics.

Encryption time is the total time to encrypt the ECG data.

Decryption time is the total time to decrypt the encrypted ECG data. The comparison of decryption time of different cryptography algorithms is shown in

Energy consumption defines that the total amount of energy consumed by the body sensors in the network.

Throughput defines that the number of ECG data received at the receiver to the delay of data transmission. The comparison of throughput of different cryptography algorithms is shown in

To solve the problem of transmission of ECG data against adversaries, an enhanced Elliptic Galois Cryptography (EGC) is presented in this paper. In the proposed EGC, private key has been selected optimally using PBOA algorithm. Using this optimal private key, public key has been generated. Then the encryption process of EGC has been done using chaotic neural network. As well as, the encrypted data has been decrypted using chaotic neural network. To evaluate the performance of the proposed cryptography algorithm, MIT-BIH Normal Sinus Rhythm dataset is used. The performance of the EGC-PBOA-CNN has been compared with that of the EGC-CNN, EGC and ECC. Simulation results showed that the proposed EGC-PBOA-CNN decreased the encryption time and decryption time as well as it increased the throughput of the network. In future, deep learning models will be presented for secure and efficient ECG signal transmission.

The author with a deep sense of gratitude would thank the supervisor for his guidance and constant support rendered during this research.