【作者】 姜大鹏；

【导师】 庞永杰；

【作者基本信息】 哈尔滨工程大学 ， 流体力学， 2011， 博士

【摘要】 近几十年来,智能水下机器人(Autonomous Underwater Vehicle/AUV)技术取得了长足的发展。其在海洋环境数据采集、区域搜索、海底目标探测等方面发挥着越来越重要的作用。随着作业任务复杂度的增加,单个水下机器人有时无法满足工作的需要。多水下机器人系统(Multiple Autonomous Underwater Vehicles/MAUV)凭借其空间分布性、任务执行中所体现出来的高效性、鲁棒性以及灵活性方面都具有单个水下机器人无法比拟的优势。关于多水下机器人系统的理论与应用研究在近十几年间逐渐引起国内外相关机构的重视。本文针对多水下机器人协同作业研究中需要解决的几项关键技术进行了研究：1.开展了模块化、分布式多水下机器人协调控制软件体系结构研究。在所建立体系结构中,每个AUV都被抽象为一组功能相对独立、信息相互依赖的功能模块。基于网络协议的信息传递机制为单个AUV内部功能模块之间的信息交互提供了稳定高效的支持,同时为功能模块的分布式运行提供了基础的保证。系统内水下机器人之间的信息交互手段被抽象为相对独立的通讯模块,保证了AUV间信息交互格式的统一与规范。2.开展了基于行为的多水下机器人协调控制方法研究。针对水下作业环境复杂多变的特点,从提高水下机器人自主性、自适应能力的需求出发,开展了基于行为的水下机器人智能控制方法研究。多水下机器人协同作业过程中的运动控制指令由-系列的行为(包括协作行为)共同作用来实现。对于行为融合(行为选择)问题论文中采用了基于多目标值优化的方法进行研究。论文中所使用的IvP计算模型通过对行为决策空间的离散化保证了行为融合算法的实时性,以及行为融合结果的优化性。3.考虑到行为融合结果跳变对AUV运动控制稳定性可能产生影响,论文结合非线性控制理论和混合系统理论的相关方法开展了基于行为的AUV运动控制稳定性研究。进行了基于指令滤波反演设计方法的AUV运动控制器设计,开展了基于多Lyapunov函数的行为融合输出跳变条件下的AUV运动控制稳定性分析,得到了行为输出跳跃条件下AUV运动控制稳定的条件。4.开展了基于市场拍卖机制的多水下机器人任务分配方法研究。针对单任务拍卖模型容易陷入局部极大值的问题,通过考虑任务间的“协同”作用,提出了基于“聚类”模型的多水下机器人任务分配算法,提高了多水下机器人任务分配的优化程度。通过仿真算例验证了方法的有效性。5.针对多水下机器人协调编队控制问题,开展了多水下机器人协同编队控制方法研究。通过对多AUV协同编队运动模型进行简化,利用图论的方法描述系统中AUV之间的信息交互拓扑结构,基于多智能体一致性理论开展了不同通讯拓扑结构下分布式控制律的设计,使得AUV利用局部感知信息进行运动指令解算最终驱动其艏向趋于一致。通过建立相应的Lyapunov函数证明了系统的稳定性。6.建立了多水下机器人协调控制仿真环境,针对未来多水下机器人系统的典型应用设计了不同的试验用例。通过仿真试验初步验证了本论文相关研究方法的可行性、有效性。更多还原

【Abstract】 The technology of autonomous underwater vehicle(AUV) has achieved an impressive de-velopment during the last decades. AUV is playing an important role in such tasks as oceano-graphic data collection, resource exploration and underwater object inspection. As tasks get more complex, some times the single AUV can not full fill the requirements of tasks. Robust-ness, efficiency and flexibility are the advantages offered by multiple autonomous underwater vehicles(MAUV) over a single autonomous underwater vehicle system and hence has drawn more attention from research institutions all over the world than ever before.Although fast developed, some of the problems concerned with MAUV were not well solved. In this thesis, several key techniques concerned with cooperation of multiple au-tonomous underwater vehicle system were studied:1. A modularized, distributed software architecture was established for coordinated control of multiple autonomous underwater vehicle system. In this architecture,each AUV was abstracted as a group of software modules which has different functionality of their own while depends on the information provided by each other. The network protocol based information exchange mechanism provides a fast and stable means for exchange of information between software modules belong to the same AUV. The unified information exchange mechanism also makes it possible for modules of the same AUV to run in distributed fashion. Functionality of communication between different AUVs was realized by an relatively independent communi-cation module(which provides the software abstraction of the communication devices such as underwater modem),which provides a unified format for communication between AUVs.2. "Behavior-based" control theory was applied to the coordinated control of multiple autonomous vehicles. In order to improve the autonomy and adaptability of the AUVs in dy-namical undersea environments, behavior-based intelligent control strategy was studied. In the execution of cooperative tasks, motion command was generated by a set of interacting behav-iors.The multiple objective optimization based behavior fusion method was used in behavior fusion. IvP model was applied to ensure the realtime demand and optimized output of behavior fusion.3. As the jump of behavior fusion results will have influences on stability of AUV motion control,Nonlinear control theory and hybrid system theory was applied to the problem of "stability of behavior based control of AUV". A nonlinear motion controller of AUV based on command filtered backstepping(CFBS) was established. Control stability of AUV under jump of behavior output was studied and the conditions under which the AUV would remain stable was derived based on theory of multiple Lyapunov function.4. The problem of task allocation of MAUV was studied and "market-based" allo-cation strategy was applied to solve the problem.By inspecting the "Synergy effect" between tasks, an improved auction-based task allocation algorithm(cluster based task bidding algo-rithm) was proposed to improve the optimality of task allocation results.5. The problem of coordinated formation control was studied. Simplification was made on the motion model of MAUV coordinated formation control, graph theory was used to describe communication topologies between AUVs in the system. Consensus theory of multi-agent system was used to design distributed control laws of AUVs in formation control. The control law can be calculated by local information collected by the AUV and the direction of all the AUVs can be driven to the same value under the control law. Lyapunov function was design to prove stability of the system.6. Simulation environment for coordinated control of multiple autonomous under-water vehicle system was established, some typical task cases were designed to verify the feasibility and effectiveness of the methods proposed in this thesis.更多还原