Study On Hydrothermal Plume Searching And Localization Via An Autonomous Underwater Vehicle

The discovery of seafloor hydrothermal activity was an important event of marine science and earth science research.The area around hydrothermal vents has created a unique ecological environment,and supplied exploitable rich hydrothermal sulfide minerals resources.The plume sprayed form hydrothermal vent was called chemical plume.The traditional hydrothermal detection method included two main kings of method:ship-board towed system and using HO Vs into deep sea for searching with umbilical cable from research vessel.With the continuous development of science and technology and in contrast to traditional detection methods,AUV has enormous advantages in autonomous detection,using AUV for searching hydrothermal vents will become necessary method in this field.Making a kind of high-efficiency and accurate plan for AUV tracing plume source was important subject in hydrothermal vents detecting research.We attempted to study on tracking strategy based on source likelihood map,and activity-based on-line mission planning.The method above will provide necessary theoretical basis for future efficient location research.The main points of this paper were as follow:1.A modeling method based on filament random motion was set in this paper through traditional plume model and computational fluid mechanics.Non-buoyant plume simulation model was set based on the method for verifying chemical plume tracing and localization algorithm.Through display module of dynamic simulation platform,we can see the spare shape of non-buoyant plume was winding in the simulation,the simulation results showed the characters of plume were evenly distributed and discontinuous.The non-buoyant plume simulation model can catch up the description lack of existing chemical plume simulation,also can simulate the character of real flow environment very well.2.Study on chemical plume tracing and localization based on intelligence science and proposed a draw method for mapping the likelihood map based on improved partially observable markov decision process.Using six-tuple to build chemical plume source likelihood map.Online updating the chemical plume source likelihood map through belief states updating mechanism.Compared with mapping method based on Bayesian theory,the results verified the method proposed in this paper could draw a more accurate source likelihood map for searching and localization hydrothermal plume.3.For the time-dependent flow environment,a set of plume tracing and localization method based behavior-based was proposed in this paper.There were changing Z-shape searching behavior,tracing behavior based improved artificial potential field theory,multi-triangle shape rediscovery behavior and source confirming behavior based on the source likelihood map.In the 4 behavior process above,plume tracing behavior have been studied in detail.Artificial potential field theory was taken into plume tracing problem,the source probability value in grid was set as a kind of gravitation.The autonomous vehicle can trace plume follow the direction of calculated resultant force.4.Considered the complexity and high cost of the experiment in water,for increasing the safety and success rates of water experiment,firstly we did the experiment on ground using the ground mobile robot to trace and localization the source position.In this paper ground simulation platform was built and a lot of simulation experiments were did based the platform,and also did the practicality experiment was did on land environment.The simulation results and practicality experiment results were both can verify the feasibility and robustness of tracing strategy and algorithm.5.For water simulation platform,a mass of simulations results compared with tradition method in water simulation experiment summed up by predecessors.The results showed that the method was purposed in this paper had more advantages in search time and success rate.We also did physical experiment in real water environmental platform,AUV traced the chemical plume followed the planed path calculated by tracing and localization method proposed in this paper,and the results verified the validity and efficiency of tracing algorithm.