摘要
为完成对微型直升机航向的控制,利用机载测控系统获得人工操纵数据,采用神经网络和参数拟合的方法提取人工控制指令u和对应的航向变化ψ,从而得出人工控制策略,并建立相应的控制策略模型。在比较两种不同的控制策略模型时发现,神经网络模型适用于控制指令实时输出的参考,而参数化模型更易于在微型直升机的控制系统中使用。以参数化控制策略模型和PID控制律为基础,可规划控制指令轨迹并设计航向控制器。实验结果表明,该控制器能够稳定而准确地改变和保持微型直升机航向。
A human operator's ability to control the yaw of a micro helicopter was studied to develop an automated yaw control system. The human response was recorded, through an onboard measurement and control system during tests with a micro helicopter. The experimental data was used to build a neural network and a parameterization method to abstract the human yaw angle control strategy. The neural network model was used as the reference output of the control command, while the parameterization model was used in the onboard control system. The parameterization model was combined with a PID controller to control the yaw movement. Flight experiments show that the combination can stably and quickly complete yaw movements.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2003年第5期621-624,631,共5页
Journal of Tsinghua University(Science and Technology)
基金
国家教育振兴计划资助项目(202012-013)