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

This paper proposes a robust speech encryption algorithm based on fractional-order chaos to address the issue that existing encryption schemes do not take into account the security of frequency-domain features and encryption efficiency. The fractional-order robust chaotic equation proposed in this paper solves the current problem that fractional-order chaos lacks robustness and can resist parameter jitter. To address 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 fundamental frequency genetic algorithm. The speech features are extracted and smoothed, and finally the genetic algorithm is used to find the key with the greatest differentiation and the lowest distribution density. Facing the current security issues of speech audio domain features, this paper proposes a high-dimensional and efficient speech encryption scheme based on dual domains. Process the original speech to obtain high and low frequencies. The high-frequency and low-frequency data are respectively mapped into three-dimensional distributions, and the elements of the three-dimensional distribution are diffused and confused. Finally, they are dimensionally reduced and linearly combined. This method encrypts the frequency domain information of the original speech from the bottom layer, ensuring the feature security of the original speech. A series of numerical tests on chaotic sequences show that the proposed fractional-order chaos has good performance. A series of attacks and numerical tests on the generated ciphertext revealed that the encryption algorithm is secure, robust, and has the ability to successfully obfuscate and scramble speech points.