Reinforcement Learning-Based Chaotic Communication System for Secure Transmission of Encrypted State Information in the Smart Grid

This paper proposes a system for transmitting and receiving encrypted state information via a communication channel in the presence of noise and interference. The proposed system uses a chaotic signal generated by a Lorentz oscillator to encrypt the state information, which is then transmitted through the communication channel. At the receiver end, the received signal is decrypted using a similar key generated by a forced Lorentz oscillator. The accurate determination of the force signal is essential for synchronizing chaotic signals, and this paper proposes the use of reinforcement deep learning agents to train and determine the force signal. The proposed communication scheme involves the use of a state estimator, a master chaotic oscillator, two slave oscillators, and two RL agents. The proposed system was simulated using MATLAB Simulink, and the results show that the errors exhibit a repetitive nature, with low and high values corresponding to the input signal. The proposed system provides a reliable and secure system for transmitting sensitive information over communication channels.