考虑脱靶量约束的超声速导弹三维微分博弈突防制导OA
Three-dimensional differential game penetration guidance for supersonic missiles considering miss distance constraints
针对助推滑翔超声速导弹末段无动力滑翔机动能力有限,多拦截协同合围下主动突防与终端高精度打击难以兼顾的问题,提出一种考虑脱靶量约束的超声速导弹三维微分博弈反拦截制导方法.该方法结合导弹末段滑翔特性采用"纵向弹道优化+水平面博弈突防"的解耦架构,将三维攻防运动拆分为纵向落角约束比例导引跟踪和水平面多玩家追逃博弈机动;并重构微分博弈性能指标,以全域三维脱靶量积分项替代传统终端单点脱靶代价,同步约束突防生存性与终端命中精度;同时,基于自适应动态规划设计思想,通过评价-动作神经网络在线逼近最优值函数与博弈策略,推导网络权值梯度自适应更新律,并严格证明了博弈均衡解的收敛性.仿真结果表明,在"一对多"典型场景下,对比传统纯比例导引制导方案,该文方法能使得超声速导弹成功规避拦截并精确命中目标.该方法在非线性、强对抗、多拦截场景下具有显著突防效能优势,为超声速导弹反拦截制导律设计提供了新的有效途径.
In view of the challenges faced by boost-glide supersonic missiles,such as limited maneuverability during the terminal phase of unpowered glide and the difficulty in balancing active penetration and high-precision terminal strike under multiple interceptor cooperative encirclement,a three-dimensional differential game anti-interception guidance method for supersonic missiles considering miss distance constraints is proposed.Combined with the terminal gliding characteristics of missiles,this method adopts a decoupled framework of"vertical trajectory optimization and horizontal game-based penetration",which decomposes the three-dimensional attack-defense motion into vertical trajectory tracking with proportional navigation constrained by impact angle,and multi-player pursuit-evasion game maneuvering in the horizontal plane.Moreover,the performance index of the differential game is reconstructed,where the integral term of full-domain three-dimensional miss distance replaces the traditional terminal single-point miss distance cost to simultaneously restrict penetration survivability and terminal hit accuracy.Meanwhile,based on the design idea of adaptive dynamic programming,an actor-critic neural network is utilized to online approximate the optimal value function and game strategies,the adaptive gradient update law for network weights is derived,and the convergence of the game equilibrium solution is rigorously proven.Simulation results demonstrate that,in typical"one-to-many"combat scenarios,compared with the traditional pure proportional guidance scheme,the proposed method enables supersonic missiles to evade interception successfully and ensures high-precision target strikes after effective penetration.This method exhibits remarkable advantages in penetration performance under nonlinear,intense-confrontation and multi-interception conditions,and provides a novel and effective approach for designing anti-interception guidance laws of supersonic missiles.
赵日;陆宇;孙瑞胜;游勇安
南京理工大学 能源与动力工程学院,江苏 南京 210094南京理工大学 能源与动力工程学院,江苏 南京 210094南京理工大学 能源与动力工程学院,江苏 南京 210094南京理工大学 能源与动力工程学院,江苏 南京 210094
军事科技
反拦截突防微分博弈脱靶量约束自适应动态规划弹道制导神经网络
anti-intercept penetrationdifferential gamemiss distance constraintsadaptive dynamic programmingtrajectory guidanceneural network
《南京理工大学学报(自然科学版)》 2026 (4)
406-419,14
国家自然科学基金(1257240362303225)
评论