Blockchain enabled optimal Hopfield Chaotic Neural network based secure encryption technique for industrial internet of things environment

Abstract

Industrial Internet of Things (IIoT) denotes a network of interlinked sensors, instruments, and other devices for industrial applications in the domains of manufacturing, logistics, transportation, etc. IIoT security is a major crucial research area for several applications. Image encryption techniques gained popularity in the recent years, thanks to increasing requirements for secure image transmission in IIoT environments. At the same time, conventional security solutions built for sensitive data protection are getting outdated in IIoT environment due to the participation of third party. Blockchain (BC) is one of the recent solutions used for security purpose which eliminates the involvement of a third party. With this motivation, the current research article presents a new BC-Enabled Shark Smell Optimization with Hopfield Chaotic Neural Network (SSO-HCNN) for secure encryption in IoT environment. The proposed SSO-HCNN model exploits a composite Chaotic Map (CM) which is integrated into staged logistic and tent maps to initially process the images and develop the variables needed for Arnold mapping. In addition, the SSO algorithm is developed with maximum PSNR and coefficient fitness function to select the optimum secret and public keys of the system amongst the random numbers. Besides, the diffusion phase utilizes HCNN to create a self-diffusion chaotic matrix whereas the jumbled image performs XOR operation using the keys to obtain the cipher image. In SSO-HCNN model, the cryptographic pixel value in the image is saved on BC thus guaranteeing the security and privacy of the images. To examine the superior performance of SSO-HCNN model over state-of-the-art methods, a set of simulations was conducted on benchmark test images. The simulation results of the proposed SSO-HCNN model were promising under different evaluation parameters.