Research On Tracking And Mapping Of Oil Spill Plumes With Autonomous Underwater Vehicles

With the prosperity of marine oil industry,oil spill accidents happen frequently.The spilled oil are both harmful to marine environments and human lives and activities,and rapid response to and disposal of spilled oil after oil spill accidents is of significance,which requires to rapidly and efficiently obtain information on the distribution of the spilled oil.With the development of autonomous underwater vehicles(AUVs),their applications in observation of the plumes formed by spilled oil has been directed a lot of interests.By employing AUVs and their capabilities especially of the on-line mission planning based on real-time sensor data,high resolution data about the spatial and temporal distribution of a dynamic undersea oil spill plume could be collected with a rapid and efficient way.For the future applications of AUVs in collecting data of distribution of oil spill plumes,strategies and algorithms which employ AUVs' on-line planning capability should be developed,to fully fulfill AUV's capabilities and potentials and cope with oil spill plumes'dynamic features in these plume observing missions.Partly supported by a project of Chinese National Natural Science Foundation "Tracking and mapping of oceanic oil spill plumes with autonomous underwater vehicles based on bio-inspired behaviors(Grant No.41376110)",in this thesis,we mainly study the strategies and on-line planning algorithms which aim to enable an AUV to autonomously and efficiently track and map an oil spill plume based on real-time sensor information.The main research efforts and results are as follows:(1)The features and models of oil spill plumes are analyzed.Then based on oil spill plume models and the Lagrangian particle random walk approach to simulating turbulent plumes,a model for simulation of oil spill plumes is presented,which could be used to generate a proper numerical plume with turbulent oil spill plume features complicating the study in this thesis.And based on the model,a computer simulation environment is developed using C++ programming language,which implements plume simulation,flow field simulation,AUV dynamics simulation,etc.The developed plume model and the computer simulation environment provide supports to the study on strategies and algorithms of tracking and mapping oil spill plumes with AUVs in this thesis.(2)In order to guaranty the correctness of plume detection with multiple sensors mounted on an AUV and improve the efficiency of AUV plume tracking and mapping missions,a sensor fusion algorithm is developed to fuse information multiple AUV sensors and output reliable judgment that whether the AUV detects the oil spill plume or not,which is the basis for effective implementation of the AUV's on-line mission planning algorithms studied in this thesis.(3)For efficient mapping of an oil spill plume which is developed by continuous release of oil spills from the oil spill source,a strategy of adaptive lawnmower type mapping is presented,which adapts the AUV tracklines to enable an AUV to effectively track and collect data of an oil spill plume in two dimensional horizontal plane when the plume is intermittent and meandrous.A behavior-based mission planning algorithm is then designed and implemented in the developed simulation environment,and the simulation results demonstrate the effectiveness of the designed strategy and algorithm.(4)In addition to oil spill plumes with continuous source,an oil spill plume that is formed when the source stops releasing is also considered in this thesis.To effectively track and map the dynamic oil spill plume transported by ocean currents,a boundary tracking strategy which enables an AUV to track and map the plume along the plume boundary is designed.And also the behavior-based planning algorithm is developed and implemented in the simulation environment.With this strategy,potentially the movement and diffusion of the plume could be calculated.