【作者】 苏抗；

【导师】 周建江；

【作者基本信息】 南京航空航天大学 ， 信号与信息处理， 2011， 博士

【摘要】 本文以低可观测微小卫星设计需求为依据，进行低可观测关键技术的研究。与常规飞行器不同，微小卫星低可观测设计具备以下特征：1）探测设备均分布于卫星下方，主要威胁类型为雷达；2）微小卫星轨道机动能力有限，在有限能源下难以通过变轨实现威胁规避；3）太空环境背景简单，不存在可供卫星隐藏的遮挡体；4）强烈的太空粒子辐照、有限的可用资源及严格受限的发射重量、体积，限制了吸波材料、频率选择表面及主动电子对抗设备的有效使用。针对上述特征，本文重点进行了微小卫星低RCS外形设计及低可观测外形飞行姿态规划方法的研究。全文可概括为以下几个部分：1、在无法获取国外研究详情而国内尚无公开报道的情况下，通过对国外专利及相关技术的跟踪，结合威胁设备发展情况、相似技术研究情况及我国现有低可观测微小卫星研究进展，提出了以低RCS外形设计为基础、低可观测外形飞行姿态规划为辅助的研究方法，在目标特性分析理论、特征控制技术及优化方法研究的基础上，保证了研究成果的先进性和实用性。2、针对已有飞行器低可观测性能评估模型精度有限、易受先验知识影响的特点，提出了一种基于卫星轨道根数的评估模型。模型将威胁设备分布概率与威胁水平相结合，根据卫星的轨道、散射特性及威胁设备的分布、工作特征，可精确计算出不同方位的威胁性分布情况，实现有效的卫星低可观测性能评估。同时，以设计模型为依据，在对不同轨道威胁性分布进行大量计算的基础上，获得了微小卫星低可观测轨道的优化选择策略。3、为降低微小卫星低可观测设计对星体结构的影响，提出了一种独立于卫星热控与力学结构的微小卫星低RCS外形，避免了对光学窗口、半导体太阳能电池板等复杂结构的RCS减缩处理，在降低卫星设计难度的同时，获取了更好的低可观测性能。在低RCS外形优化设计中，提出了一种卫星低RCS外形快速优化方法，通过关键参数选择及低可观测性能评估函数设计，将外形优化转换为评估函数最小值求解问题，结合基于数据拟合的快速预估方法，实现了设计外形的快速优化处理。4、针对低可观测外形飞行姿态规划需求，提出了一种适用于地面站、具备全局最优解搜索能力的静环境规划算法。建立了适用于卫星外形姿态规划的数字空间图模型，提出了基于“锥形”威胁均值计算空间及局部最优姿态的规划空间压缩方法，制定了与姿控复杂度约束无关的规划空间压缩策略，结合专门设计的启发预估代价快速计算方法、OPEN/CLOSE表结构及相应操作策略，解决了已有算法计算量大、难以应用于工程实践的问题，实现了不同约束条件下稳定、有效的规划。5、为满足计算资源有限的微小卫星自主规划需求，设计了一种基于局部最优的动环境规划模型，提出了一种基于粒子群优化的动环境规划算法。制定了基于规划时间分段的规划搜索策略，显著降低了规划算法的搜索难度，并提升了威胁代价评估的有效性；引入粒子群优化，设计了流程简单、易于执行、存储空间需求低、规划计算量小的动环境规划算法，将规划实时性提高了近4个数量级。同时，设计算法的规划计算复杂度与规划空间维度、复杂性无关，并具备了同时降低雷达、激光探测设备及高能激光武器威胁性的能力。6、以基于局部最优的动环境规划模型为基础，提出了一种基于改进遗传算法的卫星外形姿态规划算法。针对规划适应度函数计算复杂度较高的特点，设计了每轮迭代只需进行1次适应度计算的自适应交换、变异策略；定义了基于实数编码的个体基因结构，并将粒子群优化中粒子追踪的思想引入遗传算法的变异操作，提升了算法的搜索效率，降低了算法执行的复杂度。与相同计算量的典型自适应遗传算法相比，在迭代步长有限的情况下，设计算法收敛速度优势明显，满足了计算资源有限的微小卫星自主规划需求。7、结合静环境规划算法及动环境规划算法的设计特点，以无需规划空间计算的全局最优解模型为基础，提出了一种具备全局近似最优解快速搜索能力的卫星外形姿态规划算法。建立了基于规划姿态性能评估函数的规划模型，定义了扩展链表式个体结构，设计了基于个体节点信息的适应度快速计算方法，制定了具备威胁分布预感知能力的初始种群生成方法、保留优势无效个体的进化搜索策略、基于进化收敛情况的自适应规划精度及计算量控制策略，在确保全局最优解搜寻性能的基础上，充分降低了规划算法的计算复杂度。仿真结果表明，在计算资源有限的条件下，设计算法的低可观测规划性能优于已有规划算法。更多还原

【Abstract】 The key technologies on micro-satellite low observability are explored, to meet low observablemicro-satellite design needs. Compared to aerocraft low observable design, micro-satellite lowobservable design has some special characteristics:1) Detecting systems are distributed under themicro-satellite, and radar is the major threat.2) Most micro-satellites have limited orbitmaneuverability. It’s hard for them to avoid threat by orbit transfer.3) The out space is clear, and thereis no object to shield the micro-satellite.4) Considering micro-satellite’s limited energy, weight,cubage and strong space particle irradiation, RAM, FSS and active electronic countermeasure areunsuitable for micro-satellite.According to the characteristics above, micro-satellite low RCS shape design and low observableshape flight attitude planning are studied in this paper, and the following subjects are presented:1) Based on relative patents, similar technologies, threat development, also, considering lowobservable micro-satellite research and manufacture ability today, a research plan is proposed, whichis aimed to enhance the on-orbit micro-satellite’s operational effectiveness and survivability by lowRCS shape design and low observable shape flight attitude planning. Plenty of existed similartechnologies and theories make the plan feasible and meaningful.2) An orbit-parameter-based threat evaluation model is proposed, to improve the precision andavoid the subjective analysis in existed models. According to micro-satellite orbit, RCS, LCS andthreat distribution, both threat existing probabilities and threat level in different directions could becalculated effectively in the model, which makes the evaluation result effective. Based on theorbit-parameter-based model, different micro-satellite orbits are calculated. According to the statisticalresults, a low observable micro-satellite orbit selection strategy is presented.3) A micro-satellite low RCS shape is proposed. The shape is separated from thermal andmechanics shape. Compared to the existed ones, the shape is simpler in structure and better inperformance. At the same time, a micro-satellite low observable shape optimization method isproposed. The optimization method includes three steps, which are key parameter selection, lowobservable performance evaluation and evaluation result fitting. Based on the optimization method,the low RCS shape optimization design is carried out quickly and efficiently.4) A static-environment shape attitude planning algorithm is proposed for micro-satellite earthstation, to find the global optimal flight attitude. A digital space model is established. Special planningspace compression methods are presented. Heuristic estimating function and refined OPEN/CLOSE tables are defined. Compared to the existed global optimal search algorithms, the static-environmentalgorithm has limited computational complexity, and is able to accomplish shape attitude planningunder different constrains stably and effectively.5) A dynamic-environment shape attitude planning algorithm based on particle swarm optimization(PSO) is proposed for micro-satellite application. The algorithm divides planning time into severalsections and processes the planning in each section separated. Taking advantage of the search strategy,the PSO-based dynamic-environment planning algorithm has very limited computational complexity,and the real-time performance is enhanced by thousands of times. In addition, the algorithm’scomputational complexity is independent of the planning space dimension and complexity, and thealgorithm is able to reduce the threat level of radar, laser detecting system and laser weapon at thesame time.6) Based on planning time dividing model, a dynamic-environment shape attitude planningalgorithm based on refined adaptive genetic algorithm (AGA) is proposed. A novel search strategy ispresented, special individual structure is defined, and refined crossover and mutation strategies aredesigned, to improve the algorithm’s search ability and reduce the algorithm’s computationalcomplexity. Compared to the classic AGA, within limited computational load, the algorithm has fasterconvergence speed, and meet the needs of micro-satellite with limited signal processing ability.7) Combined the static-environment planning algorithm’s search ability and dynamic-environmentplanning algorithm’s low computational complexity, a novel global planning algorithm is proposedfor shape attitude planning. A planning model based on threat evaluation function is established,extended link-list individual structure is defined, a simple fitness calculation method is introduced,and special individual production, reproduction, crossover and mutation strategies are presented, toenhance the global optimal flight atittude search ability and reduce the algorithm’s computationalcomplexity. Accoding to the simulation results, within limited computational load, the algorithmworks better than existed static-environment planning algorithms and dynamic-environment planningalgorithms in low observable performance.更多还原