Path Planning And Cooperative Control Of A Fleet Of Saucer-type Underwater Gliders

The 21st century is a century of oceans.China is a vast maritime country with 3 million square kilometers of sea area under its jurisdiction and 18,000 kilometers of continental coastline.It is related to national strategic security and long-term development to care about ocean,to understand and manage ocean,to strengthen ocean observation and exploration.Underwater glider(UG)is an important tool for early warning detection,anti submarine reconnaissance,ocean observation,resource detection,search and rescue,which has a wide range of applications in military and civil fields.In many key technologies of UG,cluster path planning and cooperative path following are one of the hot research topics.In this thesis,a saucer-type UG,which is funded by the National Key Research and Development Program of China,is taken as the research object.The path planning and cooperative control problem of underactuated saucer-type UG cluster is studied,and the main work is stated as follows:1.Considering the influence of current interference and underwater obstacles,the three-dimensional path planning problem of underactuated saucer-type UG cluster is studied.The existing path planning for ocean vehicles is mostly suitable for single underwater vehicle,which can not meet the needs of path planning for ocean vehicle cluster.Moreover,most of the current path planning methods are suitable for two-dimensional horizontal plane,which can not meet the actual needs of three-dimensional ocean observation by using UG cluster.To solve the above problems,this thesis proposes a two-layer cooperative three-dimensional path planning structure based on the combination of global path planning and local path planning.Firstly,a multi UG path planning structure based on global and local coordination is proposed,which ensures that UG can flexibly adjust the local path when gliding along the global optimal path.Secondly,according to the navigation characteristics and equipment configuration of saucer-type UG,the energy consumption model and navigation time model are established,and the optimization objective of three-dimensional path planning is designed.Thirdly,based on the proposed two-layer coordination structure,the quantum-behaved adaptive particle swarm optimization and the biased min-consensus method are used to complete the three-dimensional path planning of multi UG under the optimal energy consumption and time consumption,and to avoide the collision among UG and the collision among obstacles and UG.Simulation results verify the effectiveness of the proposed multi UG three-dimensional path planning method.2.Aiming at the multiple saucer-type UG system with comprehensive complexity such as nonlinearity,strong coupling,multi-input multi-output,uncertainty,strong disturbance and input constraints,the vertical plane cooperative path following control problem of the under-actuated saucer-type UG cluster is studied.Since the length of the slideway of the internal attitude adjustment sliders of the saucer-type UG and the volume of the buoyancy adjustment oil sac are physically limited,the motion capabilities of UG actuator are limited.Therefore,the input constraints of the actuator needs to be considered.In addition,the saucer-type UG is only equipped with navigation equipment such as the electronic compass and depth gauge.Due to the small number and low accuracy of navigation equipment,and the influence of inherent measurement error,accumulated deviation and ocean current,it is necessary to consider the difficulty of obtaining UG speed information and leading to the actual situation of status information incomplete.In order to solve the above problems,this thesis proposes a cooperative diving controller suitable for saucer-type UG.Firstly,according to the established saucer-type UG mathematical model,considering the limited position of the UG attitude adjustment sliders and the limited mass of the buoyancy adjustment actuator,an adaptive bounded cooperative path following fuzzy controller is constructed in combination with the projection operator and the saturation function.Secondly,a fuzzy state observer based on low-frequency learning mechanism is proposed for the situation where the speed is unmeasurable,which not only realizes the simultaneous observation of unknown dynamics and speed information,but also ensures the low-frequency characteristics of the learning signal,and realizes the cooperative diving of multiple saucer-type UG in the vertical plane.The input-to-state stalility of the closed-loop systems are established via cascade theory.Simulation results verify the effectiveness of the proposed state and output feedback cooperative path following controllers.3.Aiming at the multiple saucer-type UG system with comprehensive complexity in the weak communication environment of underwater narrow bandwidth and low carrier frequency,the three-dimensional cooperative path following control problem of the under-actuated saucer-type UG cluster is studied.Since the saucer-type UG cluster gliding in the underwater three-dimensional space,the existing cooperative control methods of marine vehicles are only applicable to the two-dimensional plane,and cannot meet the actual requirements of the three-dimensional operation of the UG cluster.In addition,in order to maintain the formation,the UG needs to continuously exchange path information,such that the requirements for communication resources are relatively high,which requires a lot of energy consumption and is difficult to implement in practical applications.In order to solve the above problems,this thesis proposes an aperiodic triggered three-dimensional cooperative path following controller for saucer-type UG.Firstly,a three-dimensional cooperative path following fuzzy controller is designed by establishing an air path frame,and a path update law based on an event-triggered mechanism is proposed,which effectively reduces the communication times.Secondly,a three-dimensional cooperative path following fuzzy controller based on a self-triggered mechanism is designed,which not only reduces the communication times,realizes aperiodic monitoring,greatly reduces the communication burden and energy consumption,but also realizes the cooperative following of cluster planning path.The input-to-state stalility of the closed-loop systems are established via cascade theory.Simulation results verify the effectiveness of the proposed event-triggered and self-triggered three-dimensional cooperative controllers.