Chaotic Signal Processing Methods And Applied Research And Its Hardware Implementation

Until recently, the notions of determinism and randomness were seen as opposites and were studied as separate subjects with little or no overlap. In a deterministic system, chaotic dynamics can amplify small differences, which in the long run produces effectively unpredictable behavior. Chaos implies that not all random-looking behavior is the product of complicated physics. The character which chaos behaves makes it possible to predict chaos time series. Short-time prediction of chaos has developed in more than twenty years. Based on Takens embedding theory and delay reconstruction in phase space theory, many methods had been proposed. Among these methods, nonlinear adaptive predictor for adaptive prediction has widely attentions for it can realize real-time processing conveniently.In this paper we make a deeply research on the prediction algorithm of low dimensional chaotic time series. And we also make discuss on the real-time realization. The contents of the paper are summarized as follows,1) Some basic concepts in chaos including the theory of delay reconstruction in phase space, several characters in chaotic dynamics. Phase space reconstruction is the creation of a multidimensional, deterministic state space from a lower dimensional time series. It is a prerequisite step for analyzing a time series for making predictive state space models.2) Deeply study on several algorithms used for prediction, including RPSOVF, SOVF, S-M, and nonlinear adaptive prediction model based on several closet neighbors.3) Research some real hopping frequency codes by predicting them with RPSOVF, SOVF, S-M, and nonlinear adaptive prediction model based on several closet neighbors.4) The paper is concerned with the property and application of the new float DSP chip TMS320C6701, the structure of TMS320C6701 EVM board and the software Code Composer Studio (CCStudio).5) Run these algorithms on the EVM experiment board. By measure and compare, RPSOVF is the most efficiency algorithm among them. In order to make it has better capability and predict one data within 10μs, we propose two ways to optimize the process of the program.