Improved trajectory data encryption method for internet of vehicles using GAN-based chaotic logistic algorithm

As vehicle networking technology advances, securing the transmission of vehicular trajectory data has become crucial. Trajectory encryption might be considered an image encryption issue. However, traditional image encryption algorithms need more critical space and a low correlation between the key and the original data. We introduce a trajectory data encryption method based on Generative Adversarial Networks (GANs) to address these challenges. The proposed method transforms IoV (Internet of Vehicles) trajectory data into image data. It utilises an optimised chaotic logistic algorithm enhanced by adding M-semi sensor product operations with six original parameters. A loss function that gauges the difficulty of cracking the encrypted data is incorporated into the GAN framework to improve key generation. Our experimental results indicate that the method can withstand common attacks like noise attacks, DoS, and MIM and outperforms existing methods in encryption metrics and computational speed during parallel processing. The proposed work is validated with findings that confirm this approach's suitability for regular encryption transmission of IoV data. The proposed work outperformance in the following experimentation: critical space analysis, statistical attack analysis, critical sensitivity analysis and system run time. Future work will focus on enriching the loss functions used in GAN training, expanding training objectives, exploring a semi-fixed critical pool, and developing a key data-matching algorithm to optimise the method further.