提高无人机舵机电动加载系统载荷跟踪性能的关键技术研究

Along with the performance development of UAV's（Unmanned Aerial Vehicle）actuators, it is urgently necessary to develop the actuator load simulator platform with high performance. The traditional structure-invariance control strategy cannot satisfy the control requirements on the condition of inevitable nonlinearities such as mechanical clearance, friction and strong disturbance like surplus torque etc. in the electric loading system. Based on the hinge moments of certain UAV's actuators, the new, low-inertia and high-performance sine-driven brushless motor is selected as the loading device. The mechanical structure and control algorithm are synthetically optimized and the best rotary inertia and the best coupling stiffness of system are analyzed. A CMAC hybrid controller is used to realize the loading control of the brushless motor. The key technologies, like structure parameter optimization, stability, output chattering restraint, network training method, weights updating rules, dynamic learning rate adjustment and memory space reduction etc. , are intensively investigated to take full advantages of CMAC's strong nonlinear approximation ability, quick learning rate, simple and real-time control ability. Through the theoretical research, simulation analysis and experiment validation on DSP+FPGA platform to improve the intelligent control strategy, the load tracking performance will be improved thoroughly and the research achievements have a very wide prospect for application.

随着无人机舵机性能的日益提高，研制高性能的舵机负载模拟实验仿真平台成为迫切需要。针对加载系统中不可避免的机械间隙、摩擦等非线性和多余力矩等强扰动问题使传统结构不变性控制策略无法满足需求的现状，以某型无人机舵机的铰链力矩为基础，选取新型低惯量、高性能正弦波型无刷力矩电机作为负载模拟的执行机构，综合优化系统的机械结构和控制算法，分析系统的最佳转动惯量和连接刚度，采用小脑模型神经网络（CMAC）复合控制方法，进行直流力矩电机力矩加载的控制研究，深入探索CMAC复合控制的结构参数优化、稳定性、输出抖动抑制、网络训练方法、权值更新规则、学习率动态调整以及存储空间压缩等关键技术，充分发挥CMAC智能控制非线性逼近能力强、学习速度快、简单、适于实时控制等优势，通过理论研究、仿真分析以及基于DSP+FPGA平台的实验验证完善智能控制策略，以全面提高加载系统载荷跟踪性能，具有广阔的应用前景。

随着无人机舵机性能指标的日益提高，研制高性能的舵机负载模拟实验仿真平台成为迫切需要。无人机舵机所承受的气动载荷相对较小，主要采用以直流力矩电机为加载执行机构的电动加载系统，无笨重的油源，污染小、易维护、响应速度快、结构简单且可靠性高，非常适于实验室环境条件进行无人机舵机性能测试。 .针对无人机舵机电动加载系统中不可避免的机械间隙、摩擦等非线性因素和多余力矩等强扰动问题，使传统结构不变性控制策略无法满足需求的现状，以某型无人机的舵机铰链力矩为基础，选取新型低惯量、高性能正弦波型无刷力矩电机作为负载模拟的执行机构，综合分析系统的机械结构和优化力矩控制算法，研究系统的转动惯量和连接刚度对加载系统载荷跟踪性能的影响，并采用小脑模型神经网络（CMAC）和比例微分（PD）复合控制策略法，进行直流力矩电机力矩加载的控制研究，深入探索CMAC复合控制的结构参数优化、稳定性、输出抖动抑制及平滑、网络训练方法、权值更新规则、学习率动态调整、非均匀量化、激活区域的模糊化处理、网络映射以及存储空间压缩等关键技术，充分发挥CMAC智能控制算法非线性逼近能力强、学习速度快、简单、适于实时控制等优势，通过理论研究、仿真分析以及基于DSP+FPGA平台的实验验证完善智能控制策略，以全面提高加载系统载荷跟踪性能，具有广阔的应用前景。