【作者】 傅阳光；

【导师】 胡汉平； 周成平；

【作者基本信息】 华中科技大学 ， 控制科学与工程， 2011， 博士

【摘要】 无人飞行器航迹规划作为任务规划的重要组成部分,无论是在军用还是民用方面都越来越受到人们的重视,成为当前研究的热点。本论文对粒子群优化算法及其在无人飞行器航迹规划中的应用进行了研究。针对无人飞行器航迹规划的特点,围绕基于粒子群优化算法及其改进算法的航迹规划方法展开了深入的研究。主要研究内容包括：(1)粒子群优化算法收敛性分析；(2)基于带繁殖策略的量子粒子群优化算法的航迹规划方法研究；(3)基于相位角编码的量子粒子群优化算法的航迹规划方法研究；(4)基于差分进化量子粒子群优化算法的航迹规划方法研究。通过将标准粒子群优化算法参数的经验区域划分为四个子区域,研究了标准粒子群优化算法在经验区域的各个子区域内的收敛和发散行为,分析了系统特征根与算法参数的关系,得到了一系列结论。数值仿真实验展示了不同子区域内的算法参数对粒子位置和粒子速度运动轨迹的不同影响,进一步验证了所得结论的正确性。同时,给出了粒子群优化算法参数的选择策略,为后面章节仿真实验中粒子群优化算法参数的选取提供了理论依据。针对粒子群优化算法(PSO)存在的早熟收敛问题,通过将种群的繁殖机制引入量子粒子群优化算法(QPSO),提出了一种带繁殖机制的量子粒子群优化算法(HQPSO),并成功地将该算法应用于无人飞行器的三维航迹规划。同时,运用统计学方法,通过仿真实验比较了HQPSO算法与QPSO算法以及带动态变化惯性权系数的PSO算法的性能。仿真实验结果表明,与基于QPSO算法和基于PSO算法的三维航迹规划方法相比,基于HQPSO算法的三维航迹规划方法能以更快的收敛速度找到更优的航迹,且该方法生成的三维航迹能够有效地实现威胁回避和更好地进行地形跟随。通过对QPSO算法的编码方式进行改进,本文提出了相位角编码量子粒子群优化算法(θ-QPSO算法)。选用了几个常用的标准测试函数对θ-QPSO算法的性能进行了测试,并将其与GA算法、DE算法、PSO算法、θ-pso算法和QPSO算法的性能作了比较。标准测试函数仿真实验结果表明,θ-QPSO算法在选用的标准测试函数集上均表现出比其他五种算法更好的性能。此外,本文还将θ-QPSO算法成功地应用于无人飞行器航迹规划问题。以航迹规划这一实际工程应用问题为背景,在不同战场环境下比较了θ-QPSO算法和GA算法、DE算法、PSO算法、θ-PSO算法、QPSO算法五种算法的性能。仿真实验结果表明,基于θ-QPSO算法航迹规划方法比基于其他几种算法的航迹规划方法更好。最后,本文提出了一种DE算法和QPSO算法的混合算法,称为DEQPSO算法,并将其成功地应用于海洋环境下的无人飞行器航迹规划。为了降低问题的复杂度和提高算法的求解效率,本文提出了一种海洋环境下的地形预处理方法。该方法首先提取出岛屿的轮廓,然后用椭圆对岛屿轮廓进行拟合。鉴于传统的基于采样的威胁代价计算方法非常耗时且难于确定采样率,本文提出一种基于航迹段与威胁源位置关系的威胁代价计算方法。在不同的战场环境下(不同的发射点,不同的目标点,不同的雷达威胁),DEQPSO算法均显示出了其高效性和优越性。仿真实验结果表明,DEQPSO算法不但比GA算法、DE算法、PSO算法、DEPSO算法和QPSO算法具有更强的寻优能力,而且具有更快的收敛速度和更好的算法稳定性。更多还原

【Abstract】 As an important component of mission planning, route planning for Unmanned Aerial Vehicles (UAV) in both military and civilian applications is getting more and more attention for researchers and has become a reasearch hopspot in recent years. The particle swarm optimization algorithm and its application in route planning for UAV is studied in this thesis. Considering the characteristics of route planning for UAV, the route planning methods based on particle swarm optimization algorithm and its improved algorithms are investigated deeply in this paper. The main contents of this paper include:(1) the convergence analysis of particle swarm optimization algorithm; (2) route planning for UAV based on quantum-behaved particle swarm optimization (QPSO) with breeding strategy; (3) route planning for UAV based on phased angle-encoed and quantum-behaved particle swarm optimization; (4) route planning for UAV based on hybrid differential evolution with QPSO.The empirical region of algorithm parameters was divided into four small regions, where the convergence and divergence properties of standard particle swarm optimization (PSO) algorithm were studied. At the same time, the relationship between the characteristic roots and algorithm parameters were analyzed. A series of conclusions were deduced through rigorous mathematics derivation. Finally, numerical simulation demonstrated the different effects of different algorithm parameters on the loca of particle position and particle velocity, and further illustrated the validity of conclusions given in this paper. At the same time, the parameters selection strategy for PSO was given, which provided the theoretical basis for the selection of parameters of PSO in the simulation experiments of the following chapters.In view of the premature convergence problem of PSO, a hybrid quantum-behaved particle swarm optimization algorithm (HQPSO) was presented by introducing the breeding strategy into QPSO in this paper. A method of 3-D route planning for UAV was set up base on the proposed HQPSO algorithm. The performance of HQPSO was compared against the QPSO and PSO with inertial weight through using statistical method. Simulation results demonstrate that HQPSO not only has stronger global searching ability, but also achieves a faster convergence speed compared with QPSO and PSO. The path planner based on HQPSO can find better path with faster convergence speed. In addition, the 3-D path generated by HQPSO algorithm can fulfill the threat avoidance and terrain following effectively.Through changing the coding mode, the phase angle-encoded and quantum-behaved particle swarm optimization (0-QPSO) is proposed in this paper. Several representative benchmark functions are selected as testing functions. The real-valued genetic algorithm (GA), differential evolution (DE), PSO, phase angle-encoded particle swarm optimization (θ-PSO), QPSO, andθ-QPSO are tested and compared with each other on the selected unimodal and multimodal functions. To corroborate the results obtained on the benchmark functions, a new route planner for UAV is designed to generate a safe and flyable path in the presence of different threat environments based on theθ-QPSO algorithm. The PSO,θ-PSO, and QPSO are presented and compared with theθ-QPSO algorithm as well as GA and DE through the UAV path planning application. Experimental results demonstrated good performance of theθ-QPSO in planning a safe and flyable path for UAV when compared with the GA, DE, and three other PSO-based algorithms.Finally, this paper presents a hybrid DE with QPSO for the UAV route planning on the sea. A simple method of pretreatment to the terrain environment is proposed to reduce the complexity of problem and improve the computational efficiency. The terrain pretreatment includes extracting the contour of islands and fitting them by ellipses. In consideration of the conflicet between sample accuracy and computation efficiency, a new method based on the location relation of path segement and threat region is proposed. To show the high performance of the proposed method, the DEQPSO algorithm is compared with the GA, DE, PSO, hybrid particle swarm with differential evolution operator (DEPSO), and QPSO, in the presence of different threat environments, which mean different start points, endpoints, and radar threats. Experimental results demonstrated that the proposed method is capable of generating higher quality paths stably and efficiently for UAV than any other tested optimization algorithms.更多还原