A Control Method for Unknown Chaotic Systems

In chaos control proposed until today, the target system is known in many cases. However, since the generating mechanism of chaos is not strange in few case, it is required to control only by the time series data observed from the system. Then, as the method of stabilizing the state of the unknown system, we propose the technique made combining PFC and DFC using a parameter which indicates the balance of both methods. The prediction values at the PFC input portion are determined from the known output using RNN. The input is impressed only near UPO calculated from the output using the concept of UPR.