Speech encryption algorithm based on robust fractional order chaos with maximum difference key optimization algorithm

Abstract This paper proposes a robust speech encryption algorithm based on fractional-order chaos to address the issue that existing encryption schemes fail to account for the security of frequency-domain features and encryption efficiency. The fractional-order robust chaotic equation proposed in this paper resolves the current problem where fractional-order chaos lacks robustness and is vulnerable to parameter jitter. To tackle the issue that current key generation algorithms do not consider the security of the keys themselves, this paper develops a key search scheme based on a fundamental frequency genetic algorithm. Speech features are extracted and smoothed, and the genetic algorithm is then used to identify the key with the highest differentiation and the lowest distribution density. To address the security challenges of speech audio domain features, this paper proposes a high-dimensional and efficient speech encryption scheme based on dual domains. The original speech is processed to extract high and low frequencies, which are then mapped into three-dimensional distributions. The elements of these three-dimensional distributions are diffused and confused, followed by dimensional reduction and linear combination. This method encrypts the frequency-domain information of the original speech at the foundational level, ensuring the security of the original speech features. A series of numerical tests on chaotic sequences demonstrate that the proposed fractional-order chaos exhibits strong performance. Furthermore, a series of attacks and numerical evaluations on the generated ciphertext reveal that the encryption algorithm is secure, robust, and capable of effectively obfuscating and scrambling speech samples.