基于协同进化的平面追逃对策研究

建立了二维平面内动力学约束下追逃运动的数学模型.首先为追捕者设计了基于比例制导算法和进化算法的混合追捕策略,以提高其追捕能力;然后利用协同进化算法对追捕者和逃跑者的追逃策略进行进化.仿真结果表明,进化后的逃跑策略能有效规避比例制导的追捕者,逃跑者在协同进化过程中涌现出众多复杂多变的规避策略.

     

基于博弈论及Q学习的多Agent协作追捕算法

多Agent协作追捕问题是多Agent协调与协作研究中的一个典型问题。针对具有学习能力的单逃跑者追捕问题，提出了一种基于博弈论及Q学习的多Agent协作追捕算法。首先,建立协作追捕团队，并构建协作追捕的博弈模型；其次,通过对逃跑者策略选择的学习，建立逃跑者有限的Step-T累积奖赏的运动轨迹，并把运动轨迹调整到追捕者的策略集中；最后,求解协作追捕博弈得到Nash均衡解，每个Agent执行均衡策略完成追捕任务。同时,针对在求解中可能存在多个均衡解的问题，加入了虚拟行动行为选择算法来选择最优的均衡策略。C#仿真实验表明，所提算法能够有效地解决障碍环境中单个具有学习能力的逃跑者的追捕问题，实验数据对比分析表明该算法在同等条件下的追捕效率要优于纯博弈或纯学习的追捕算法。     

多追捕者-单-逃跑者追逃问题实现成功捕获的约束条件

针对包含有n个追捕者及1个逃跑者的2维平面多机器人追逃问题,对实现成功捕获的约束条件进行了研究.经过理论分析得出:在机器人拥有全局视野的情况下,即使单一逃跑者性能优于每个追捕者,只要满足追捕者与逃跑者的速率比大于sin(π/n),逃跑机器人落在追捕机器人所构成的凸多边形内部且逃跑者和追捕者构成的相邻追-逃阿波罗尼奥斯圆满足两两相交(相切)这2个约束条件,则追捕者通过选择合适的追捕策略就一定可以实现成功抓捕.此外,还给出了在此约束条件下的追捕者和逃跑者的追逃策略.多组仿真实验同样证明了本文提出的约束条件是正确的.     

基于量子少数者博弈的多机器人追捕

多个带有自利因素的追捕机器人在追捕一个逃跑者的过程中,自身利益与整体利益之间会产生冲突,导致系统付出更多代价.若调整收益分配机制,并引入量子少数者博弈,则可将经典策略空间扩展到量子策略空间.在该空间下,追捕者追求自身利益最大化时,也能达到整体最优.通过对追捕过程中量子少数者博弈进行实验分析表明,采用量子策略的机器人,其自身利益与整体利益得到统一,追捕效率大幅度提高.     

一种多机器人协作追捕-逃跑问题策略研究

提出一种多机器人协作追捕多个移动目标策略,基于主从式协作结构提出了分区主从式协作方法,在确定追捕目标点后通过引入追捕意向、追捕耗时,心智态度等三个指标概念选择最优合作追捕团队成员,并根据逃跑者状况以及协作效用判断追捕结果,评价追捕效率,仿真试验结果证明该策略的可行性及有效性.     

一种多机器入协作追捕-逃跑问题策略研究

提出一种多机器人协作追捕多个移动目标策略．基于主从式协作结构提出了分区主从式协作方法,在确定追捕目标点后通过引入追捕意向、追捕耗时、心智态度等三个指标概念选择最优合作追摘团队成员,并根据逃跑者状况以及协作效用判断追捕结果,评价追捕效率．仿真试验结果证明该策略的可行性及有效性．     

碰撞自规避多弹分布式协同制导与控制

针对多枚导弹三维空间协同攻击机动目标情形,提出一种碰撞自规避多弹分布式协同制导律及其实现方案.基于可以测得的目标信息,将目标视作领弹,与参与协同攻击的多枚导弹形成"领弹–从弹"拓扑结构,基于网络同步算法实现导弹对目标的协同攻击.为了实现碰撞自规避,当导弹与目标相距较远时,采用带安全距离的同步算法.当导弹与目标接近时,取消该安全距离.基于运动学关系,将协同制导指令转化为速度、弹道倾角和弹道偏角指令.针对指令的跟踪控制问题,提出了一种基于改进微分器的抗干扰动态面控制方案.将参数不确定和外部扰动均视为系统干扰,采用改进微分器实现对该干扰的精确估计,从而保证了闭环系统的跟踪性能.仿真结果表明本文提出的控制器实现了对协同制导指令的精确跟踪,整个方案实现了碰撞自规避协同攻击.     

融合协同进化人工鱼群算法和SVM的雾霾预测方法

针对日益严重的雾霾污染问题,提出融合协同进化人工鱼群算法和支持向量机的雾霾预测方法.首先,运用佳点集构造均匀分布的种群,并引入自适应视野范围策略、自适应步长策略、种群间协同策略,提出协同进化人工鱼群算法.然后,使用协同进化人工鱼群算法,优化支持向量机的主要参数.最后,构建基于支持向量机的雾霾预测模型,预测雾霾天气.在10个测试函数上的实验证明协同进化人工鱼群算法的性能,在6个UCI数据集上的实验验证预测模型的稳定性和有效性.     

自适应两阶段分组求解大规模全局优化问题

协同进化是解决大规模全局优化问题的一种有效策略，但是该策略不能对存在相关性变量的大规模问题进行有效分组，最终导致算法性能下降.针对上述问题，提出一种基于自适应两阶段分组的差分协同进化算法.首先，在第1阶段分组中，根据决策变量贡献度，将其分为正促进组和负抑制组；然后，在第2阶段分组中，分别对两组内的变量进行相关性识别，根据相关变量所占比例进行自适应分组；最后，采用差分协同进化算法对分组后的组件进行优化.实验结果表明本文所提方法能够实现对大规模全局优化问题中相关变量的有效分组，提高了算法的收敛性，通过标准大规模优化测试函数集验证了算法的有效性和适用性.     

针对动态目标的多无人机协同组合差分进化搜索方法

针对多无人机动态目标协同搜索问题,提出一种组合差分进化无人机协同搜索航迹规划方法.建立动态目标协同搜索环境信息图模型及无人机运动模型.基于改进差分蝙蝠算法和自适应差分进化算法,设计基于种群数量自适应分配的组合框架,将差分进化算法中的变异、交叉和选择机制引入蝙蝠算法,构建组合差分进化算法的协同搜索算法,并对无人机动态目标协同搜索的航迹进行优化.针对待搜索目标轨迹随机多变且具有规避侦察特性的现实场景,建立可回访数字信息图和自适应目标搜索增益函数,从而提高无人机对动态目标的捕获能力.最后,通过仿真结果验证所提出的无人机动态目标协同搜索算法的有效性.     

无人机的追逃对抗仿真研究

代数图论方法较之微分对策建立的无人机追逃对抗模型更易仿真求解,利用梯度方法改变Laplacian矩阵的非零特征值给出了对抗双方的控制输入,仿真证明了该方法的可行性.利用边界权值函数考虑无人机通信或感知能力的变化.然后,针对仿真过程中遇到的问题提出根据实际情况构造不同的性能指标.仿真表明逃逸者为迷惑追击者不时交叉逃逸;而追击者也会更换追击目标:相对距离不断变化.     

Policy Improvements for Probabilistic Pursuit-Evasion Game

This paper focuses on a pursuit-evasion game (PEG) which involves two teams: one side consists of pursuers trying to minimize the time required to capture evaders, and the other side consists of evaders trying to maximize the capture time by escaping the pursuers. In this paper, we propose a hybrid pursuit policy for a probabilistic PEG, which possesses the combined merits of local-max and global-max pursuit policies proposed in previous literature. A method to find optimal pursuit and evasion polices for two competitive parties of the pursuers and evaders is also proposed. For this, we employ an episodic parameter optimization (EPO) algorithm to learn good values for the weighting parameters of a hybrid pursuit policy and an intelligent evasion policy. The EPO algorithm is performed during the numerous repeated simulation runs of the PEG and the reward of each episode is updated using reinforcement learning, and the optimal weighting parameters are selected by using particle swarm optimization. We analyze the trend of the optimal parameter values with respect to the number of the pursuers and evaders. The proposed strategy is validated both in simulations and experiments with small ground robots.     

Coordinated pursuer control using particle filters for autonomous search-and-capture

In a real-world pursuit-evasion (PE) game, the pursuers often have a limited field-of-view of the evaders and thus are required to search for and detect the evaders before capturing them. This paper presents a unified framework and control algorithm using particle filters (PFs) for the coordination of multiple pursuers to search for and capture multiple evaders given the ability of PF to estimate highly non-Gaussian densities prevalent in search problems. The pursuer control problem is formulated as a stochastic control problem where global objectives function of both searching and capturing are common. To take the evaders’ actions into account, an action measure (AM) is defined over the evaders’ PDs is used to represent the probability that the evader may transit each state in the PD. The global objective functions for search and capture are then decomposed into local objective functions for unification through objective priority weights. Coordination between the pursuers takes place through the multi-sensor update where the observation likelihoods of all pursuers are used in the PF update stage. The control actions of each pursuer are then determined individually, based on the updated PDs given the objective weights, action measures as well as evader importance weights in the case of multiple evaders. The proposed algorithm is tested in three scenarios for its effectiveness. In addition, a parametric study on the average capture time against the initial variances of the target state uncertainty is conducted to test for robustness. Results show that the pursuers are able to capture all the evaders in each case with the capture time for the second and last scenario differing by only 2.9% implying firstly that under the proposed algorithm, the capture time is not proportional to the increase in the number of evaders and also suggested robustness and potential scalability of the proposed algorithm.     

A non-stationary problem of group pursuit

A linear non-stationary conflict-interaction problem for controlled objects with n pursuers and m evaders with equal dynamical potency of each participant is considered. The objective of the pursuers is to capture all the evaders; the objective of the evaders is to avoid overtaking of at least one of them. Sufficient conditions for the global evasion problem solvability are stated.     

Scalable and practical pursuit-evasion with networked robots

In this paper, we consider the design and implementation of practical pursuit-evasion games with networked robots, where a communication network provides sensing-at-a-distance as well as a communication backbone that enables tighter coordination between pursuers. We first develop, using the theory of zero-sum games, an algorithm that computes the minimal completion time strategy for pursuit-evasion when pursuers and evaders have same speed, and when all players make optimal decisions based on complete knowledge. Then, we extend this algorithm to when evader are significantly faster than pursuers. Unfortunately, these algorithms do not scale beyond a small number of robots. To overcome this problem, we design and implement a partition algorithm where pursuers capture evaders by decomposing the game into multiple multi-pursuer single-evader games. We show that the partition algorithm terminates, has bounded capture time, is robust, and is scalable in the number of robots. We then describe the design of a real-world mobile robot-based pursuit evasion game. We validate our algorithms by experiments in a moderate-scale testbed in a challenging office environment. Overall, our work illustrates an innovative interplay between robotics and communication.     

Soft capture of two coordinated evaders

Sufficient conditions for the soft capture of two evaders are obtained for a linear problem of pursuit of two evaders by a group of pursuers. It is assumed that the capabilities of all the participants are equal, and the same control is used by the evaders.     

Problem of pursuit of a group of coordinated evaders in linear recurrent differential games

For a linear nonstationary problem of pursuit of a group of evaders by a group of pursuers under the condition that the fundamental matrix of the homogeneous system is a recurrent function and all the evaders use the same control, sufficient conditions for capturing at least one evader are found.     

UAVs in Urban Operations: Target Interception and Containment

In future urban warfare scenarios, the expectations placed on unmanned aerial vehicles (UAVs) in terms of autonomy, reliability and cooperation will be significantly increased. In this paper, a novel algorithm is developed which enables a swarm UAVs to contain and intercept multiple evading targets. In particular, the algorithm allows a swarm of UAVs to intercept significantly faster evading targets if there are more UAVs than targets. Several numerical experiments are conducted, in which four pursuers attempt to intercept one or two faster evaders. The algorithm is very effective at containing the targets; however, as the number of evading targets increases the speed of the evaders must be reduced to ensure capture. Finally, a scenario in which there are more evading targets than pursuers is considered, and, given that the evaders were slower than the pursers, containment and capture was successfully accomplished.     

A cooperative Homicidal Chauffeur game

We address a pursuit-evasion problem involving an unbounded planar environment, a single evader and multiple pursuers moving along curves of bounded curvature. The problem amounts to a multi-agent version of the classic Homicidal Chauffeur problem; we identify parameter ranges in which a single pursuer is not sufficient to guarantee evader capture. We propose a novel multi-phase cooperative strategy in which the pursuers move in specific formations and confine the evader to a bounded region. The proposed strategy is inspired by the hunting and foraging behaviors of various fish species. We characterize the required number of pursuers for which our strategy is guaranteed to lead to confinement.     

复杂三维多面体环境中空地协作追逃问题

结合无人机(UAV)的空中移动和无人车(UGV)的地面移动特点,本文提出了一种UAV/UGV空地协作系统,并且针对其在复杂地形中的追逃问题,提出了一种复杂三维多面体环境中UAV/UGV空地协作追逃策略.首先介绍了UAV/UGV空地协作系统的结构与协作追逃问题描述.接着将边界值问题(BVP)改进并离散化作为博弈走法生成器.然后,针对逃方已知追方位置,而追方只具备直线视野(LOS)的不利条件,分析了最坏情况.逃方策略在保证最大生存条件下尽可能获得博弈胜利.追方策略根据逃方状态分成3种情况进行讨论:逃方处于追方的视野范围内、逃方刚刚消失于追方视野以及追方完全丢失逃方的情况.最后,对比仿真结果说明了本文算法的有效性,并分析了追逃结果的影响因素.由于地形是非凸的并且充满障碍,因此该策略虽不能保证追方一定能够胜利,但在最坏情况下是最优的.