Enhancing Continuous Chaos Communication Using Machine Learning in Resource-Limited Devices

Machine learning is rapidly finding its way into the solving of everyday complex problems. One such application is in the area of chaotic encryption, where machine learning techniques can be used to improve the security and synchronization of encryption algorithms. Chaotic encryption is a technique that uses chaos theory to encrypt messages communicated between a transmitter and a receiver, making them extremely difficult to decipher without the correct decryption key. Here, we first discuss error correction for chaotic synchronization using conventional methods with an accuracy of 86%. We then use machine learning algorithms to reduce the error of the decrypted message extracted by learning patterns in the encrypted message and adjusting the encryption parameters accordingly. Using linear regression, k-mean, and DB-Scan, We present an increase in the original accuracy achieved by the decrypted message. Additionally, we use machine learning algorithms to detect anomalies in encrypted messages. The use of machine learning in chaotic encryption has the potential to greatly improve the security of encryption algorithms.