摘要
为了提高温箱的温度控制精度与抗扰能力,提出了一种基于单神经元的自适应改进控制算法。分析了温箱模型及其时延性与传导惯性,将单神经元与经典PID算法相结合,利用单神经元的非线性逼近和自学习能力,实现参数的在线调整,同时利用二次型性能指标对连接权值进行约束调整,提高参数调整稳定性和收敛速度。仿真结果表明,与经典PID算法相比,该改进控制算法具有更快的响应速度和更小的超调量,系统自适应和鲁棒性能得到明显提升。
In order to improve the temperature control accuracy and anti-interference ability of the incubator,an adaptive improved control algorithm based on single neuron is proposed. The incubator model and its time delay and conduction inertia are analyzed. The single neuron is combined with the classical PID algorithm. The on-line adjustment of parameters is realized by using the nonlinear approximation and self-learning ability of single neuron. At the same time,the quadratic performance index is used to restrict the adjustment of connection weights to improve the stability and convergence speed of parameter adjustment. The simulation results show that compared with the classical PID algorithm,the improved control algorithm has faster response speed and smaller overshoot,and the adaptive and robust performance of the system is significantly improved.
作者
高磊
陶彦飞
刘祥言
GAO Lei;TAO Yan-fei;LIU Xiang-yan(The 27th Research Institute of China Electronics Technology Group Corporation,Zhengzhou 450047,China)
出处
《自动化与仪表》
2022年第1期26-29,共4页
Automation & Instrumentation
