| In recent years,the exploration and development of marine resources have gradually attracted great attention from countries around the world.Autonomous underwater vehicle(AUV),as an important tool for exploration of the ocean,has been widely applied in the military and civilian fields due to its obvious advantages such as small size,good flexibility and strong endurance.Among the key technical fields of autonomous underwater vehicles,the problems of target tracking and formation control have attracted increasing attention of researchers.Therefore,this thesis aims to address the target tracking,formation control in horizontal plane and three-dimensional space of underactuated autonomous underwater vehicles by using nonlinear control system theory,backstepping control method and adaptive neural network(NN)technique.The main contents and contributions of this thesis are summarized as follows:According to the AUV geometry and distribution of appendages including propeller and rudders,the six-degree-of-freedom kinematic and dynamic models are derived via the coordinate conversion and multi-body dynamics theory.Then,the control model of AUV is constructed,and this model considers the effects of the unknown environmental disturbances,AUV nonlinear hydrodynamic damping and coupling terms,which provides a theoretical foundation for the motion control of underactuated AUVs.The problems of path following and trajectory tracking for underactuated AUV in three-dimensional space have been considered.Both tracking control strategies are designed based on command-filter backstepping method,and the adaptive NN technology is employed to online estimate and compensate the uncertain dynamics.It is necessary to emphasize that the proposed controllers do not require any AUV hydrodynamic damping coefficient,which are easily implemented in the actual applications.In addition,both control schemes obtain the derivative value of virtual control variable through a second-order integral filter instead of directly analysing the derivative,which can not only avoid the shortcomings of calculation explosion in traditional backstepping method,but also reduce the effects of high-frequency measurement noise for the control system,and a filter error compensation loop is also developed to guarantee the tracking accuracy of the filtered signals.Then,the Lyapunov’s stability analysis proves that the whole closed-loop control system is semi-globally bounded,and simulation results demonstrates the effectiveness and superiority of the proposed control strategies.To deal with the formation control problems of multiple underactuated AUVs subject with uncertain parameters and unknown environmental disturbances in the horizontal plane,two robust formation control schemes only relying on the position information of leader are proposed for underactuated AUVs.Specifically,within leader-follower formation structure,the first formation controller is presented by employing backstepping sliding mode control technique,and a neural estimator is designed to online approximate the uncertain dynamics of leader and followers.In addition,to weaken the inherent chattering in conventional sliding mode control approach,a continuous proportional-integral(PI)function is designed to replace the signum function.Stability proof of the closed-loop control system is provided based on Lyapunov’s stability theorem.Next,within point-to-point tracking mechanism of vehicles,a formation tracking controller is designed based on backstepping control technology.Meanwhile,a predictor based NN system is designed to fast compensate the model uncertainties and unknown environmental disturbances,where the update laws of NN weight are developed based on the estimation errors instead of tracking errors,which can effectively suppress high-frequency oscillation of the control signals and improve the transient characteristics of the control system.The input-tostate stability of the closed-loop control system is proved via a cascade theory.Ultimately,contrastive simulation results are provided to verify the effectiveness and superiority of the proposed algorithms.To handle the formation control problems of underactuated AUVs in threedimensional space,two formation control strategies were proposed.Specifically,within leader-follower formation structure,a filter-backstepping based formation controller is developed,where the neural network is employed to online learn the uncertain dynamics of followers,and the adaptive control technique is applied to compensate the unknown external diaturbances and approximation errors of the NN.Then,within point-to-point tracking mechanism of vehicles,an observer-based formation controller is proposed,where a nonlinear observer is provided to online obtain the velocity information of followers,which effectively reduces the amount of communications.Moreover,a generalized saturation function is developed to guarantee the boundness of control signals,which reduces the risk of the actuator saturation.Finally,global stability of the two proposed formation controllers are proved via Lyapunov’s direct method and cascade theory,and the effectiveness and superiority of the presented controllers are verified by contrastive simulations. |
