Design and Analysis of Compressive Sensing-Based Lightweight Encryption Scheme for Multimedia IoT

This brief presents a compressive sensing (CS)-based lightweight cryptosystem for Internet of Things (IoT) applications. In a recent study, Pudi et al. presented two efficient CS-based cryptosystems: CS using stream cipher (CSSC) and CSSC with encryption of MSB-bits (CSSC-EM). The proposed work aims to improve the energy efficiency of CSSC-EM further, without compromising security. While the random Bernoulli sensing matrix in CSSC-EM offers greater flexibility for embedding security, it suffers from a high computational cost. In contrast, the deterministic nature of the structurally random matrix (SRM) rectifies the computational burden, making the CS encoder energy-conserving. In this brief, we first quantify the security of CSSC and CSSC-EM against a chosen plaintext attack (CPA). Based on this cryptanalysis, we propose a CS-based energy efficient encryption scheme using an SRM for multimedia IoT. The artificial noise is injected to quantized CS measurements to resist CPA, based on the rate-distortion analysis of CS using SRM. Further, bitstream resulting from entropy coding of quantized CS measurements is randomly corrupted through an intentional error vector to achieve uniform probability distribution. We develop an efficient hardware implementation of the overall cryptosystem based on a linear feedback shift register. Analytical and simulation results are presented to validate resistance of the proposed system against various attacks.