【作者】 高静；

【导师】 王洪斌； 丁栋；

【作者基本信息】 燕山大学 ， 工程硕士（专业学位）， 2021， 硕士

【摘要】 自主水下航行器(Autonomous underwater vehicle,AUV)在各个领域应用广泛。对多AUV编队控制的研究弥补了AUV在复杂水下作业需求上的不足,提高了作业效率,适应性强,资源利用率高。因此,本文基于领航-跟随编队策略、事件触发机制和固定时间控制理论等对多AUV编队控制问题展开研究。主要研究工作有:首先,为减少多AUV编队系统信息交换与控制器更新的能耗并加快编队收敛速度,提出了基于分布式静态事件触发机制(Distributed static event-triggered mechanism,DSETM)的固定时间编队控制策略。为了降低控制器设计的复杂性,基于反步-动态面算法,避免了“计算爆炸”问题。并将DSETM和固定时间理论同时引入多AUV编队策略中设计出事件触发固定时间编队控制器,实现系统在固定时间内稳定,稳定时间不受初始状态影响,并证明无Zeno行为。仿真验证了.所提算法.的有效性。其次,针对含未知外部扰动与模型参数不确定性的多AUV编队控制问题,设计基于固定时间的干扰观测器,对干扰进行精确估计。基于动态面控制算法,构造出基于事件触发的固定时间抗干扰控制器,并在含固定阈值的DSETM上引入动态变量,设计了分布式动态事件触发机制(Digustributed dynamic event-triggered mechanism,DDETM),结合Lyapunov稳定性理论,在降低通信能耗的同时实现系统固定时间收敛。通过仿真验证了算法的合理性。最后,在含未知外部扰动的多AUV系统编队控制中,考虑输入饱和的影响,基于双曲正切函数,结合自适应径向基神经网络(Adaptive radial basis function neural network,ARBFNN)对非线性项和未知外部扰动进行估计。并基于反步-命令滤波算法,构造出补偿机制降低滤波误差的影响,并避免“计算爆炸”问题。且结合固定时间理论、DSETM、DDETM,设计出基于ARBFNN的分布式事件触发固定时间抗饱和编队控制器,实现了系统的所有信号在固定时间内有界收敛。且DDETM在保持与DSETM相同控制性能的同时进一步降低能耗,通过仿真验证.了算法的.合理性。更多还原

【Abstract】 Autonomous underwater vehicle(AUV)has been applied in various fields.The research on multi-AUVs not only makes up for the shortcomings of the single AUV in the complex underwater operation requirements,but also improves the efficiency of the operation.At the same time,formation control has the advantages of strong adaptability and high resource utilization.Therefore,based on the leader-follower formation strategy,distributed event-triggered mechanism and fixed-time control theory,the formation control problems in multi-AUVs are studied in this paper.The detailed research work is as follows:First of all,in order to reduce the energy consumption of information exchange and controller update of multi-AUV formation systems,and to accelerate the convergence speed of formation,a fixed-time formation control strategy based on distributed static event-triggered mechanism was introduced.For the purpose of reducing the complexity of controller design,the problem of "computational explosion" is avoided based on the backstepping-dynamic surface control algorithm.And the distributed static event-triggered mechanism and the fixed-time control theory are introduced into the multi-AUVs formation strategy simultaneously.Thus,the controller is designed to achieve the system fixed-time convergence.The convergence time is independent of the initial state and it is proved that there is no Zeno behavior.The effectiveness of the proposed algorithm is verified by simulation.Secondly,aiming at the formation control problem of multi-AUVs with unknown external disturbance and uncertain model parameters,a disturbance observer based on fixed-time is designed to estimate the disturbance accurately.Based on the dynamic surface control algorithm,a fixed-time anti-interference controller based on eventtriggered is designed.As well as introducing dynamic variables on the basis of distributed static event-triggered mechanism with fixed thresholds,such as dynamic event-triggered mechanism is designed,and combined with Lyapunov stability theory,the system can achieve fixed time convergence.The rationality of the algorithm is verified by simulation.Finally,in formation control of multi-AUV systems with unknown external disturbances,the influence of input saturation is considered.Based on the hyperbolic tangent function and the adaptive radial basis function neural network,the nonlinear term and disturbance of the system are estimated.Based on the backstepping-command filtering control algorithm,a compensation mechanism is constructed to reduce the influence of filtering error and avoid the problem of "computational explosion".Then,combined with the fixed-time theory,distributed static event-triggered mechanism and digustributed dynamic event-triggered mechanism,an adaptive radial basis function neural network based distributed event-triggered fixed time anti-saturation formation controller is designed,so that all signals of the system are guaranteed to achieve bounded convergence in a fixed time.Moreover,digustributed dynamic event-triggered mechanism can maintain the same control performance as distributed static event-triggered mechanism while further reducing energy consumption.The rationality of the algorithm is verified by simulation.更多还原