流线型无人飞艇MPC姿态优化控制研究

Unmanned airships were in a wide range of applications either in military missionsor civilian missions, as they were providing improved apability such as convenient take-off and landing, floating and hovering, long endurance and high flexibility, and so on. However, it was a challenging task to design an autonomous flight control system, since airship models were charactered as multivariable, strongly coupled, time-varying, ime delay, nonlinear. Directional control system of airship was an important part of the flight control system, which was used to track the desired flight path. It is of great theoretical and engineering value to apply predictive control to solve the problems of multi-input and multi-output,nonlinear coupling, model uncertainty and saturation limitation in unmanned airship control system. On the basis of existing research, In order to solve the problems encountered on current controller design of unmanned airship directional system, such as multi-variety, nonlinear, model uncertainty, actuator saturation, and so on, this project is to aim at study model predictive control method for directional control system of unmanned airship, and the effectiveness of the proposed method will be verified by flight simulation. The research results will provide the theoretical basis and design basis for the actual development of unmanned airship control system.

无人飞艇凭借其在起飞降落方便、可长时间滞空、续航能力强、机动性好等方面的优势，无论在军事和民事领域都有广泛的应用。然而由于无人飞艇本身是一个多变量、强耦合、时变、大迟延的非线性系统，其自主飞行控制系统的设计是一个颇具挑战性的任务。预测控制凭借其在处理约束、多变量和复杂耦合系统问题上的优势，受到了极大的关注。将预测控制应用于解决无人飞艇控制系统中的多输入多输出、非线性耦合、模型不确定、饱和受限等问题，具有重要的理论意义和工程价值。本项目拟在已有研究基础上，针对当前无人飞艇控制系统研究中存在的问题，如多变量、非线性、模型不确定、执行机构饱和，设计无人飞艇的预测控制系统，并通过飞行仿真验证所设计方法的有效性。研究结果将为无人飞艇控制系统的实际研制提供理论基础和设计依据。

本项目首先基于无人飞艇的线性化模型，考虑多执行机构复合控制问题，设计了基于多变量预测控制的复合控制系统，并且通过设计权值的性能指标，获得反映执行机构实际特性的控制分配效果，而不需要改变控制器本身的求解过程。然而，基于简单线性化模型所设计的控制器又有其难以克服的缺点，即当飞艇的工作点偏离所选定的平衡点很远时，因模型相差较大，控制效果难以保证。进而，基于解析非线性模型预测控制方法，设计无人飞艇的非线性姿态控制器，该方法可以直接考虑系统的非线性，在有限时域内获得控制量的解析解。无人飞艇模型的强非线性和强耦合性，使得非线性控制器设计过程十分复杂，不适用于实际应用。线性变参数（LPV）系统相比于线性定常系统，对非线性系统有着更强的描述作用，并且基于LPV系统设计的控制器与常用的非线性控制器相比，具有实用性强，设计相对简单等优点。因此，考虑将无人飞艇模型转化为LPV模型来描述飞艇模型的不确定性，并且考虑执行机构饱和受限问题，设计鲁棒及增益调度模型预测控制器，保证模型参数不确定情况下的闭环鲁棒性能，降低了控制器设计的保守性，提高了无人飞艇航向的跟踪性能及扰动抑制能力。