Obstacle Avoidance Based On Double Closed Loop Control Of Autonomous Underwater Vehicle For Submarine Cable Laying

With the exploration of ocean resources and the development of the global economy,Submarine cables play an extremely important role in the fields of power transmission,communication,ocean engineering and new energy development.Because of the increasing demand of submarine cables and it is dangerous when laying cables in deep sea.It has become extremely necessary to use autonomous underwater vehicles(AUVs)for unmanned and fully automated laying and maintenance of submarine cables.How to balance the cable laying quality and AUV’s own safety in complex submarine environment is the key to solve the problem of AUV’s independent cable laying and maintenance operations.This dissertation proposes obstacle avoidance based on double closed loop control of autonomous underwater vehicle for submarine cable laying.The details are as follows:Firstly,the fixed coordinate and body coordinate are established to describe the relationship between the AUV and the preset path.The observation coordinate system is defined to describe the field of view of the AUV sensor.According to the shape characteristics of the obstacles,the obstacle is classified into bulge and depress terrain and the mathematical model of irregular obstacles is established by using piecewise functions.These characteristics could provide theoretical basis for the design of obstacle avoidance control.Secondly,an outer loop controller based on progressive optimization is designed.the different objective optimization functions are designed based on model predictive control to ensuring the safety of AUV itself and improving the quality of cable laying.To address the problem of multiple feasible solutions when AUV avoids bulge obstacles,the closest feature points between the AUV and the obstacle is defined to find the shortest avoid path.Substitute the optimal solution back into the model of the AUV to obtain the optimal control input.This method not only ensures the safety of AUV itself,but also improves the laying quality of submarine cables and effectively reduces obstacle avoidance energy consumption.Thirdly,an inner loop controller based on Gaussian process regression is designed to eliminate the interference of currents.The dissertation designs the probability model of the constant flow base on Gaussian process.The mean value and covariance of Gaussian process are used to represent the current and random white noise in the environment.Train the probability model with the error between the actual position and expected position.Using the mean value to compensate the control input of the outer loop controller for the next moment,which reduces the error of position,speed,and posture and improving path tracking quality.Finally,this dissertation compares the double closed-loop control with sphere obstacle avoidance predictive control method and the traditional double closed-loop control method based on model prediction by simulation.Compare the method designed in this dissertation with the traditional double closed-loop control method based on model prediction through experiments.Both simulation and experimental results indicate that the method proposed in this dissertation not only ensuring the safety of the AUV,but also performance well in cable laying quality and obstacle avoidance energy consumption.