Research On Path Planning Method For Underwater Robot’s Efficient Cooperative Target Grasping In Complex Environment

With the rapid development of related technologies of Remotely Operated Vehicles(ROVs),ROVs show extremely high application value in underwater target grasping tasks.In order to ensure that the ROV can effectively grasp the target within a certain range,it is necessary to carry out efficient grasping path planning for multiple ROVs.This paper conducts in-depth research on this topic.At the same time,the dynamic changes of the underwater environment,ocean currents,efficient grasping,and multi-robot collaboration are considered to carry out the design and analysis of the path planning algorithm.The main research contents are as follows:Firstly,the seabed grasping environment is analyzed,and the ROV’s identification of biological targets in different states of operation is studied.Considering the impact of seabed ocean currents on ROV operations,a ROV target grasping motion cost model is designed for the seabed environment.The cost and fishing consumption cost are represented and applied to the global path planning of multiple ROVs to plan the global path with the shortest catching cost.Secondly,aiming at the problem of underwater environment change,based on deep reinforcement learning Value Iteration Networks(VIN),the existing problems are analyzed,and the underwater environment is predicted by Long Short Term Memory(LSTM).,improve and design a dual-value iterative network,in order to test the effectiveness of the above algorithm,it is verified in a variety of scenarios,and applied to multiple ROV global path planning,the simulation results prove the effectiveness of the algorithm.Finally,for the task assignment problem of multiple ROV cooperative grasping,a task assignment model is designed,combined with the motion cost model to optimize and improve the particle swarm optimization algorithm,and the fitness function is designed to improve the grasping ability of multiple robots within a certain range.take efficiency.The grasping tasks of multiple ROVs are allocated through simulation experiments,and the dual-value iterative network is used to plan the allocation results in the environment before and after the prediction.The simulation results are analyzed,and the simulation results verify the effectiveness of the allocation algorithm.This paper takes the multi-robot path planning for grasping tasks in complex environments as the main research line,analyzes the submarine grasping environment and designs the ROV target grasping motion cost model,and improves and designs a dual-value iterative network.The corresponding results of this paper It is expected to play a paving and reference role for subsequent research.