PID补偿的完全在线序贯极限学习机控制器在输入扰动系统自适应控制中的应用

针对输入受外界扰动的系统在实现自适应控制难的问题，提出一种比例-积分-微分（PID）补偿的完全在线序贯极限学习机（FOS-ELM）控制器设计方法。首先，建立系统的动态线性模型，采用FOS-ELM算法设计控制器并学习其参数；其次，计算系统的实际输出误差，结合系统的控制误差，设计所需补偿的PID增量参数；最后，对PID补偿的FOS-ELM控制器参数在线调整并用于系统控制。在发动机空气燃油比（AFR）控制系统模型上进行实验，实验结果表明上述方法在实现自适应控制的同时降低了系统扰动输入带来的干扰，提高了系统有效控制率，在正负干扰系数为0.2时，其有效控制率从不足53%提高到93%以上。同时该方法易于实现，具有很强的鲁棒性和实用价值。

Application of fully online sequential extreme learning machine controller with PID compensation in input-disturbance system adaptive control

To deal with the difficulty of input disturbance system in achieving adaptive control, a design method for the Fully Online Sequential Extreme Learning Machine (FOS-ELM) controller with Proportion-Integral-Derivative (PID) compensation was proposed. Firstly, a dynamic linear model of the system was established, then the FOS-ELM algorithm was used to design the controller and learn its parameters. Secondly, by calculating the output error of the system and combining with the system control error, the PID parameters of the system compensation were designed. Finally, the FOS-ELM controller parameters for PID compensation were adjusted online and used for system control. The experiment was carried out on engine Air Fuel Ratio (AFR) control system. The results show that the proposed method can achieve the adaptive control, reduce the disturbance caused by system disturbance input, and obviously improve the effective control rate of the system at the same time. When the positive and negative interference coefficients are 0.2, the effective control rate is increased from less than 53% to over 93%. In addtion, the proposed method is easy to implement and has strong robustness and practical value.