基于混沌优化的板形信号模式识别的研究

By the analysis of shape signal and mathematical model of shape recognition, the traditional least squares shape pattern recognition method was of poor performance and low precision, the neural network recognition method also had poor results in practical application. To improve the accuracy and speed of shape pattern recognition, we make it to be a function optimization problem and use Chaos theory and fuzzy identification optimization method to solve the problem. Fuzzy recognition based on Legendre polynomial as a preliminary identification of the optimal solution to reduce the dimension of chaos and narrow the search space. The gradient descent algorithm is used to improve the chaos optimization algorithm. And the chaos optimization results are further improved. Simulation results show good performance