CAJ | 학술논문

在地面进行模拟的空间微重力育种平台是一个随机时变、强非线性的复杂系统，无法精确建模，使得系统在较大工作范围内很难实现精确控制。为此，采用基于径向基神经网络 PID 的具有自调整能力的、稳定的自适应控制器应用于本系统研究：首先通过PID控制器快速调节参数值，使其恢复到理想的期望值附近，以此来初始化RBF 神经网络；然后再用RBF神经网络在线动态调整 PID 参数的控制方法，实现系统在完成三维空间微重力模拟育种试验时所需的垂直地面Z向上的完全重力补偿；最后，通过MatLab 对系统的控制算法进行仿真研究。
재지면진행모의적공간미중력육충평태시일개수궤시변、강비선성적복잡계통，무법정학건모，사득계통재교대공작범위내흔난실현정학공제。위차，채용기우경향기신경망락 PID 적구유자조정능력적、은정적자괄응공제기응용우본계통연구：수선통과PID공제기쾌속조절삼수치，사기회복도이상적기망치부근，이차래초시화RBF 신경망락；연후재용RBF신경망락재선동태조정 PID 삼수적공제방법，실현계통재완성삼유공간미중력모의육충시험시소수적수직지면Z향상적완전중력보상；최후，통과MatLab 대계통적공제산법진행방진연구。
Studied in this paper on the ground to simulate the microgravity of space breeding platform is a random time-varying, strongly nonlinear complex system, so the system can not accurately model, which makes the system working in a wide range of difficult to achieve precise control.In this paper, the use of RBF neural network self-tuning PID has the ability, based on a stable adaptive controller is applied to study this system, first near the expected value of the parameter to quickly adjust it back to the desired value by PID controllers to this initializes the RBF neural network, and then use the RBF neural network control method online dynamic adjustment of PID parameters to achieve full gravity compensation system at the completion of the three-dimensional space microgravity simulation breeding experiments, the required Z di-rection perpendicular to the ground.Finally, study of the system control algorithms by matlab simulation.