基于鲁棒多目标优化方法的UCAV武器投放规划

摘要

This paper studies the problem of generating robust optimal air-to-ground weapon delivery mission plan for unmanned combat aerial vehicles (UCAV). Aiming at the deficiency of existing approaches in processing some uncertainties, such as disturbing in battlefield, model imprecision and operating errors, a strategy based on a robust multi-objective optimization (RMO) approach is proposed. Firstly, some constraints which include the flight capability constraint, weapon constraint and battlefield constraint, are considered. Secondly, several robust optimal cost functions are defined by using the Monte Carlo method simulating criteria of weapon delivery, and then the weapon delivery planning (WDP) problem is transformed into a RMO problem. Thirdly, an improved fast non-dominated sorting genetic algorithm (NSGA-Ⅱ) combining Monte Carlo simulation is proposed, and then a trajectory generating strategy based on basic trajectory segments (BTS) is presented to generate weapon delivery trajectories. Finally, a demonstrated air-to-ground weapon delivery planning considering uncertainties is proposed and solved, and the simulated results show that the proposed approach can deal with some uncertainties in WDP efficiently.%研究了无人作战飞机(unmanned combat aerial vehicles,UCAV)对地攻击阶段的武器投放鲁棒性规划问题.针对现有规划方法在处理战场环境扰动、模型不准确、操作偏差等不确定性因素方面存在的不足,提出了一种鲁棒多目标优化求解策略.首先,建立了飞机机动性能、武器装备性能和战场环境等约束条件模型；其次,使用仿真近似法,建立了优化指标模型,并将武器投放规划问题转化为鲁棒多目标优化问题；然后,设计了一种结合蒙特卡罗方法的快速非支配排序遗传算法对问题进行求解,并采用基于基本轨迹片元的机动轨迹生成策略生成武器投放轨迹.仿真结果表明,该方法能够有效提高武器投放规划的鲁棒性.