毕业设计（论文）-二次型性能指标单神经元控制.docVIP

摘 要 PID控制由于其算法简单、鲁棒性好及可靠性高，被广泛应用于过程控制和运动控制中，尤其适用于可建立精确数学模型的确定性系统。然而实际工业生产过程往往具有非线性、时变不确定性，难以建立精确的数学模型，应用常规PID控制器不能达到理想的控制效果。 为了克服传统PID控制的弱点,控制界已经提出了大量的对PID控制的改进方案。但这些方案一般是针对某些具体问题，缺乏通用性，附加的结构或算法也增加了控制器的复杂性，使它们的广泛应用受到限制。 近年来，随着神经元网络的研究和应用，人们开始采用神经元网络和PID控制相结合，以便改进传统PID控制的性能，这种将神经元网络和PID控制相结合的研究已经得到了一些结果。 由具有自学习和自适应能力的单神经元构成的单神经元自适应智能PID控制器，不但结构简单，而且能适应环境变化，有较强的鲁棒性。借助最优控制中二次型性能指标的思想，在加权系数的调整中引入二次型性能指标，使输出误差和控制增量加权平方和为最小来调整加权系数，从而间接实现对输出误差和控制增量加权的约束控制。 关键词：二次型，单神经元，PID控制，自适应 Quadratic Performance Control via Single Neuron ABSTRACT PID control is widely used in motion control and process control because of its simple algorithm, good robust and reliability, particularly applicable to the certain systems of establish a precise mathematical model. Nevertheless, the actual production process are often nonlinear, uncertain time-varying, it is difficult to establish a precise mathematical model, the conventional PID control can not achieve the desired effect. To overcome the weaknesses of traditional PID control, a lot of improve projects of the PID control have been put forward in the control industry. However, these projects are generally targeted at certain specific issues, the lack of universality. Additional structural or algorithm increased the complexity of the controller, restrict their wide use. Recently, with the research and application of neural networks, neural networks and PID controller have been combined to improve the performance of conventional PID control, the research of combination of neural networks and PID control has some results. Single neuron adaptive intelligent PID controller combined with self-learning and adaptive capacity of single neurons is not only simple structure and adapt to changing environments, has good robust. By dint of the idea of quadratic performance indicators in optimization control, the quadratic performance indicators are imported in the adjustment coefficient to decrease the square sum of the error of output and increment