【作者】 李昕；

【导师】 王磊；

【作者基本信息】 大连理工大学 ， 运筹学与控制论， 2024， 博士

【摘要】 航母是一类大型水面综合作战舰艇,搭载固定翼舰载机、直升机、无人机、牵引车等多种舰载装备。舰载装备在出动、回收、保障阶段都需要依托安全、可行的移动路径开展任务,任一环节发生事故都会严重削弱航母的综合战斗力。在出动阶段,需要为无人机编队规划航迹,使其在最短时间抵达目标点。在回收阶段,需要控制舰载机稳定地跟踪参考轨迹,安全降落在航母甲板上。在保障阶段,需要使牵引车安全移动到舰载机附近,为其提供调运保障服务。这三类问题均可基于最优控制问题的框架进行求解,根据具体任务场景进行运动规划与控制,高效躲避各类障碍物,对于提升各类舰载装备的任务执行效率有着重要的理论意义和应用价值。本文针对不同任务执行阶段舰载装备的运动规划和控制问题,建立了相应的最优控制模型,并基于最优控制问题的数值求解方法提出了有效的解决方案,主要内容如下:1.考虑舰载无人机协同飞行的任务需求,结合动力学方程、避障需求、编队构型、通讯距离等约束条件,构建了时间最优控制问题。由于初解可能对数值求解最优控制问题的效果影响很大,低质量的初解甚至会使问题无法求解,我们设计了一种基于改进人工势场法的热启动策略提供高质量的最优控制问题初解。此外,针对严格编队约束设计了一种松弛策略处理编队构型,增加编队飞行过程中的灵活性,处理狭窄区域编队构型无法时时保持的情形。最后,通过直接法将最优控制问题离散为非线性规划问题,通过最优控制求解器启动解算,得到满足动力学约束的高质量协同轨迹。数值仿真验证了所提求解方法的高效性和鲁棒性,并且该方法可以扩展到多个编队的轨迹规划问题中。2.由于舰载机在航母甲板上降落风险高、易发生事故,我们提出了一种基于自适应动态规划的自动着舰控制求解框架,辅助飞行员实现安全且高效的着舰过程。首先,为了时刻给舰载机提供理想着舰点位置的参考依据,设计了基于回声状态网络的运动预测模块,预测海洋环境下的甲板运动。其次,以舰载机的扩展纵向线性小扰动模型为状态方程,以扰动输入的变化率为控制输入,以最小化跟踪误差为性能指标函数,建立纵向着舰最优控制模型。最后,通过自适应动态规划对该问题进行求解,能够根据舰载机状态实时产生最优控制输入,实现不同海况下舰载机的平稳着舰控制。仿真算例表明,在低、中、中高和高海况下所提方法均能够控制舰载机自动降落在理想着舰点附近。3.航母甲板环境复杂且调运任务密集,为了实现甲板环境上的自主调运,基于全自动牵引车,我们设计了一种在线轨迹规划方法。首先,为使移动距离最短并获得避障性能,设计了一种结合移动距离、航向角、到目标点距离和人工势场法合力指标的性能指标函数。其次,以牵引车的位置和航向角为状态量,以左右轮的线速度为控制输入,综合牵引车的运动方程、物理限制、避障需求等约束条件和新设计的性能指标函数,建立最优控制问题模型。然后,由于自适应动态规划具有在线求解能力,可通过该方法近似控制策略和性能指标函数,以寻求问题的最优控制。最后,通过收敛性分析为该问题的求解提供了理论保证。不同场景下的仿真算例表明该方法在较拥挤的空间中仍能获得较好的规划效果,并且在算例场景中可以实时规划牵引车的轨迹以避让其他运动系统。综上所述,本文从舰载装备出发,综合考虑任务需求和各类约束条件,从建立最优控制问题模型和设计相应数值求解算法两方面展开运动规划与控制相关的研究工作。这对于提升舰载装备的任务执行效率和航母的综合战斗力具有重要意义,具备进一步开发应用的潜能,同时所提出的求解方法也可用于其他运动规划与控制问题的研究。更多还原

【Abstract】 Aircraft carriers are a class of large surface-integrated combat platform developed in modern warfare,carrying fixed-wing aircrafts,helicopters,unmanned aerial vehicles,tractors,and other carrier-based vehicles.Carrier-based vehicles rely on safe and feasible paths to perform missions,any part of the accident will seriously weaken the comprehensive combat capability.The unmanned aerial vehicle formation needs to plan flight trajectories to fly to the target point in the shortest possible time when carrying out a mission.The fixed-wing carrier aircraft needs to track the reference sliding trajectory steadily to land safely on the carrier deck.The automatic tractor needs to plan a real-time trajectory for the carrier-based aircraft traction system on the flight deck.Using automation and intelligent technology to plan and track trajectories,efficiently avoid obstacles,and improve mission execution efficiency for various types of carrier-based vehicles can greatly enhance the combat capability of aircraft carriers and reduce the probability of accidents.Therefore,trajectory planning and control is a key technical problem for the efficient operation of carrier-based vehicles.This thesis constructs the corresponding optimal control problem model for trajectory planning and tracking problems in different mission scenarios and proposes an efficient and robust solution framework based on the ideas of optimal control direct method and intelligent method.The main contents of this study are as follows:1.Considering the conflict between the formation configuration maintenance and obstacle avoidance,an offline solution method based on the artificial potential field method is proposed for the fixed-wing unmanned aerial vehicle formation to efficiently solve the complex constraints.First,the time-optimal trajectory planning problem is transformed into an optimal control problem to by combining the dynamic equations,collision-free requirement,formation configurations,communication distance,and other constraints.Second,to provide high-quality initial guesses for the transformed problem,a warm-started method based on an improved artificial potential field algorithm is developed.Then,a loosen strategy is designed to extend the feasible domain of the problem and to facilitate the convergence of the solving process.Last,the optimal control problem is discretized into a nonlinear programming problem and solved by the direct method.This study can generate high-quality cooperative trajectories that satisfy the dynamics constraints.Numerical simulations demonstrate the efficiency and robustness of the proposed solution framework,which can flexibly fly through narrow spaces while guaranteeing the maneuverability of unmanned aerial vehicles,and the proposed planning framework can be extended to solve the problem with multiple formations.2.Considering the problems of the high risk and high accident rate of the landing process on the flight deck,an automatic landing control method based on adaptive dynamic programming is proposed for the carrier-based aircraft.First,the longitudinal landing control problem is transformed into an optimal control problem by combining the linear small disturbance equation and the state and control variables,so that the carrier-based aircraft can track the reference sliding trajectory in real time.Then,to provide the accurate position of the ideal landing point,a data-driven forecasting module for the irregular deck motion is designed.Last,based on the adaptive dynamic programming algorithm,a fast and robust landing control method that can track reference sliding trajectory steadily under different sea conditions is developed.Numerical simulation results demonstrate the performance of the proposed method,the proposed method can control the carrier-based aircraft to land at the ideal landing point,thus assisting the pilot to achieve a safe and efficient landing process.In addition,the designed forecasting module can reflect the evolutionary information of the dynamic system,which facilitates the modeling and analysis of time series and improves the forecasting performance.3.Considering the complex environment with dense obstacles of the flight deck,an online trajectory planning method based on adaptive dynamic programming is proposed for the automatic tractor.First,the real-time trajectory planning problem is transformed into an optimal control problem by integrating the motion equation of the tractor model,physical constraints,collision-free constraints,and the performance index function.Second,the total distance traveled,the effect of the heading angle,the distance from the target to the tractor,and the resultant force of the artificial potential field are included in the novel designed performance index function.Then,the problem is solved by the adaptive dynamic programming algorithm,which iteratively approximates the control strategy and the designed performance index function to explore the optimal control.Last,the convergence analysis provides the theoretical guarantee for the proposed real-time planning method,and this study can generate a feasible and safe trajectory for the automatic tractor.Simulations with different scenarios show that the proposed method achieves better planning results in more congested spaces and can generate the real time trajectory for the tractor to avoid obstacles in both the small-scale scenarios and the flight deck scenario.In summary,this thesis focuses on the research of carrier-based vehicles and carries out the study related to trajectory planning and control in terms of establishing the optimal control problem model and designing the corresponding solution methods.This is of great significance for improving the mission execution efficiency of carrier-based vehicles,and has the potential for further development and application,while the proposed solution method can also be used for the study of other trajectory-related problems.更多还原