【作者】 李成凤；

【导师】 田彦涛；

【作者基本信息】 吉林大学 ， 控制理论与控制工程， 2012， 硕士

【摘要】 群体自主机器人系统因其具有鲁棒性、可扩展性、适应性和经济性等方面的优势,成为近年来机器人领域的研究热点之一。本文以群体自主机器人系统一致性控制为切入点,以国家自然科学基金项目和研究生创新项目为依托,着重研究了群体系统对复杂环境的适应性、群体收敛速度的可控性以及群体系统性能的优化。具体如下：1)设计了一种分布式协同控制策略来实现群体一致性行为对非线性环境的适应。利用具备有界梯度的特定势能函数对外部非线性环境信息进行描述,结合虚拟力和最近邻居原则的思想设计了基于局部信息交互的分布式协同控制策略,实现非线性环境下的群体一致性行为,并对群体收敛时间进行了估计。仿真结果表明了该控制策略的有效性和可行性,同时也体现出了群体系统的适应性和可扩展性。2)针对障碍物环境设计了群体一致性行为的避障控制器,并进行了稳定性分析。考虑群体机器人系统中个体数量众多的优势,提出了一种基于改进传统人工势场法的群体机器人避障控制算法,并利用群体中虚拟领导者的引导作用使群体在感知到障碍物后,能够实现对障碍物的安全避碰,并同时完成协同一致任务。理论分析和仿真结果表明该算法克服了传统人工势场法中容易出现死锁现象的缺陷。3)解决了在不破坏群体机器人系统局部交互规则的情况下群体速度的可控问题。利用“软控制”思想在群体中引入可控智能体,通过对可控智能体施加控制来实现整个群体的速度可控,对具有恒定期望速度的群体自治系统和具有时变期望速度的群体非自治系统,分别利用LaSalle不变性原理和Barbalat引理从群体速度可控性、群体结构稳定性以及群体收敛快速性的角度证明所设计控制器的有效性。4)针对群体自主机器人系统一致性控制的综合性能优化问题进行了分析和研究。以群体速度误差、群体结构误差和通信量的优化为目标,以群体速度收敛时间最短为约束条件,利用粒子群优化算法得出了群体的最佳通信频率和通信半径,从而提高了群体系统可控状态的准确性、速度收敛的快速性以及通信成本的经济性。综上所述,本文就群体自主机器人系统的一致性行为进行了一系列的理论研究,主要目的是实现群体系统的适应性、可控性和优化性,并通过仿真实验进行了相应的验证与分析。更多还原

【Abstract】 For significant advantages on the robustness, adaptability, extendibility and economy of the swarm robots system, it has been an active research in the robot area in resent years. In this thesis, the main research work is the adaptability to the complex environment of swarm robots system, the controllability of swarm velocity and the optimization of system performance, which is based on the flocking control of swarm robots system. This work is supported by the National Nature Science Fund Project and Graduate Innovative Project. The main work in this paper is as follows:Firstly, the distributed control strategy is designed to make the flocking behavior of swarm robots system adapt to nonlinear environment. And the nonlinear environment is described by a given potential function which is assumed to have finite bounded slopes. And based on the local information interaction distributed control algorithm is presented by using the idea of virtual force and nearest neighborhood law, which can realize the swarm flocking behavior in the nonlinear environment, and the finish time of flocking can be estimated. Simulation results are included to verify the distributed controller, and also show the adaptability and extendibility of swarm system.Secondly, the controller is designed for swarm robot system in the environment with unknown obstacles and the stability of system is proved. The traditional artificial potential field method is improved to achieve obstacle avoidance based on the number advantage of the swarm. After perceiving obstacles the swarm can achieve safe avoidance with the obstacles and then finish flocking task by using the heading effect of virtual leader in swarm. To validate the proposed algorithm, the theory analysis and experimental simulation are performed. And the results show that the algorithm can make swarm avoid obstacles effectively, and overcome the defect that traditional artificial potential field method easy to deadlock.Then, the velocity controllability problem of swarm flocking is solved without breaking the local interaction rules of swarm robots system. Controllable agents (one or more) are introduced to the swarm based on the idea of "soft control", and the velocity controllability of the swarm is realized through controlling the controllable agents. For the swarm autonomous system with constant expected speed and the swarm non-autonomous system with time-varying expected velocity, LaSalle invariance principle and Barbalat Lemma is adopted in analysis of system stability, so that we proof speed vector of all robots can converge to the desired value, the swarm will form a stable structure, and the more controllable agents are introduced, the faster the swarm convergence. So the results show that the designed control strategy is effective.Finally, the synthesis performance optimization problem of consistency control for groups of autonomous robots system is analyzed and studied. Make the velocity error of swarm, the structure error of swarm and the traffic as optimization goal, and treat the minimum convergence time of swarm velocity as constraint, particle swarm optimization algorithm is used to obtain the best communication frequency and communication radius of swarm, in order to improve the accuracy of controlled state, the speed of swarm convergence, and the economy of communication costs.In summary, a series of theoretical studies about the flocking behavior of swarm robots system are conducted in this thesis. The main purpose of this work is to realize the robustness, adaptability and extendibility of swarm flocking system. And then, Simulation experiments are performed for the purpose of related verification and analysis.更多还原