带攻击角度约束的深度强化元学习制导律

针对执行机构部分失效的速度时变导弹机动目标拦截问题,本文提出一种基于深度强化元学习和剩余飞行时间感知逻辑函数的攻击角度约束三维制导律.首先,采用基于模型的深度强化元学习方法,建立深度神经网络动力学模型;引入模型预测路径积分控制,该深度神经网络动力学模型作为预测模型;采用元学习方法,在线学习执行机构部分失效及目标机动等环...     

Three-dimensional cooperative guidance strategy and guidance law for intercepting highly maneuvering target

Cooperative interception of the target with strong maneuverability by multiple missiles with weak maneuverability in the three-dimensional nonlinear model is studied. Firstly, the three-dimensional nonlinear model of cooperative guidance is established. The three-dimensional reachable region is represented composed of lateral acceleration and longitudinal acceleration in the two-dimensional coordinate system. Secondly, the problem of the multiple missile's reachable coverage area is transformed into the problem of cooperative coverage. A cooperative coverage strategy is proposed and an algorithm for quickly calculating the number of required missiles is designed. Then, the guidance law based on the cooperative coverage strategy is proposed, and it is proved that cooperative interception of the target can be achieved under the acceleration limit. Moreover, the relations among the number of missiles, the initial array position of terminal guidance and the coverage area of the target’s large maneuver are analyzed. The dynamic adjustment strategy of guidance parameters is proposed to reduce the guidance error. Finally, simulation results show that multiple missiles with low maneuverability can achieve effective interception of target with strong maneuverability through the proposed cooperative strategy and cooperative guidance method.     

Impact point prediction guidance based on iterative process for dual-spin projectile with fixed canards

The fixed canards configuration of a dual-spin projectile makes it difficult to apply the traditional guidance law. In this study, a modified impact point prediction guidance strategy based on an iterative process was developed for a class of dual-spin projectiles with fixed canards, to reduce the impact point dispersion. The guidance strategy is dependent on the modified projectile linear theory to rapidly predict the flight states and the impact point. For projectiles with control applied to the trajectory, the modified projectile linear theory method is known to achieve poor impact point prediction. To improve the prediction accuracy, improvements were made to the modified projectile linear theory by considering the products of the yaw rate and other small quantities. The guidance strategy is based on the iterative process for the continuous adjustment of the expected output of the roll angle of the course correction fuze, to minimize the direction error between the predicted impact point and target location. Studies were conducted on a model dual-spin projectile configuration to demonstrate the guidance details. The numerical simulations indicate that the proposed guidance strategy can effectively reduce the projectile impact point dispersion.     

Fuel-optimal and Energy-optimal guidance schemes for lunar soft landing at a desired location

Two guidance schemes (i) fuel-optimal (ii) energy-optimal to realize soft landing at a desired location on the moon are developed using the optimal control laws. The optimal control laws are obtained by solving a two-point boundary value problem formulated based on Pontryagin’s principle. The guidance laws, adapted from the optimal control laws, are obtained as a function of unknown co-state variables. Differential Transformation (DT) technique is employed to determine the unknown co-states at each time instant of landing trajectory using the information on the current vehicle state, target landing site (loaded on-board apriori) and the time-to-go. The numerical integration of co-state dynamics is avoided due to the DT based approach. The time-to-go, a critical parameter for any guidance scheme, is computed and updated real time using a simple strategy which uses the current and end states. The simple strategy for time-to-go works well even when the shape of the trajectory is nonlinear. Extensive analysis is carried out to evaluate and compare the proposed guidance schemes. Further, the proposed schemes are compared with other popular guidance schemes. The DT based proposed schemes help quantify the landing masses for fuel-optimal and energy-optimal objectives. Other features of the proposed schemes are that they do not assume constant gravity field and independent of reference trajectory.     

基于奇异摄动与反馈线性化的滑翔制导律

针对高超声速再入飞行器最远距离滑翔制导关键技术,设计了一种新型制导律.首先,利用奇异摄动理论对再入飞行器状态变量进行两个时间尺度的划分,将系统状态方程分解成两个低维子系统.然后,应用最优控制理论得出慢时间尺度上的解,并将其作为降维后的最优参考轨迹;在此基础上,利用非线性微分几何反馈线性化方法,设计快时间尺度制导指令,实...     

多仓库下物流无人机协同配送方法

无人机以其快捷、低成本优势,在物流配送中可以实现高效的包裹配送,但也存在着运行时间短、载重不足等缺点。针对当前配送建模考虑因素不够全面的问题,构建了基于能耗变化、混合时间窗和同时取送货的多仓库物流无人机配送模型,以实现配送经济成本最低。与经典的多基地车辆路径问题相比,文中研究的问题没有限制无人机出发和返回的仓库,旨在最大限度地减少无人机的数量和所有无人机行驶的总距离。为进一步优化物流无人机配送成本,针对遗传算法(GeneticAlgorithm,GA)寻优能力较差的问题,引入大规模领域搜索算法(LargeNeighbor-hoodSearchAlgorithm,LNS)作为局部搜索算子,进而提出基于改进 GA(ImprovedGA,IGA)的物流无人机协同配送算法。经仿真测试以及 Solomn标准数据验证,该算法较传统 GA在降低配送成本方面成效明显。     

Modified rolling guidance law for single moving mass controlled reentry vehicle against maneuvering target with impact angle constraints

This paper deals with the problem of guidance law design for the single moving mass controlled reentry vehicle when impact angle constraints and maneuvering target are taken into consideration. More specifically, a modified rolling guidance law is proposed with the interactive virtual target and the landing point prediction strategy. First, considering the fact that the roll channel can be controlled directly, the relative motion between the single moving mass controlled reentry vehicle and the ...     

Q-learning强化学习协同拦截制导律CSCD

为实现多枚导弹协同拦截机动目标,提升拦截效能,提出了一种Q-learning强化学习协同拦截制导律。首先,基于逃逸域覆盖理论,建立了非线性多弹协同拦截模型。其次,以视线角速率为状态,依据脱靶量构造奖励函数,通过离线训练生成强化学习智能体,并结合传统比例制导控制方法,构建基于强化学习的变导引系数制导律,实时生成实现协同拦截的制导指令。最终,通过数值仿真验证了所提算法的有效性和优越性。     

LYM-3量具研磨机装夹定位的改进

叙述了量具研磨机的结构工作原理（不含千分尺部分）以及各种参数变化对研磨的影响。对存在的问题进行了系统归类和对比分析，提出了改进方法，如单独配研磨器减少研磨行程，利用钳口铜块画刻线确定卡尺位置，找正主尺与直角尺两侧面重合而确保外量爪与滑块平行。     

基于改进人工蜂群算法的无人机实时航迹规划

为了适应现代战争战场环境的复杂性,针对所建立的突发威胁模型,提出了一种基于改进人工蜂群算法的无人机航迹规划方法,对无人机进行动态航迹寻优,并通过MATLAB进行了仿真验证。仿真结果表明,所设计的方法能够根据实时出现的威胁状况进行动态航迹规划、评价和选优,同时满足实时性要求。