Research On Monitoring System Design And Path Planning Method Of Flaw Detection ROV

At present,ships and underwater structures serving in the environment of waters have serious corrosion conditions which will cause many safety hazards in the long run because of structural stress and human factors.Therefore,a safe and effective underwater detection method is needed to complete the detection task.Compared with other underwater detection methods,the flaw detection ROV has lower cost,higher safety and unparalleled advantages.This article was funded by the 2017 scientific research project of the Jiangsu High-Tech Ship Collaborative Innovation Center.The aim of this article was to develop a flaw detection ROV,and carry out research on underwater positioning and path planning of the flaw detection ROV.Firstly,the design scheme of flaw detection ROV was formulated based on the current research status of ROV at home and abroad.The selection of sensor modules was determined based on the existing design experience of medium and small robots.The structural design of ROV was carried out too.The design and production of the information collection board of the water surface control box was accomplished in terms of hardware.The software of the water surface monitoring system depended on Java was accomplished.The preparation of the main control board program on the basis of the STM32F4 chip and the slave control board program based on the STM32F103RBT6 chip was developed.Secondly,the underwater positioning method of ROV using multi-sensor data fusion was studied and the ROV motion coordinate system was established.An adaptive unscented Kalman filter algorithm(AUKF)using fusing ultra-short baseline positioning data and inertial navigation data was proposed in order to solve the problem of inaccurate positioning accuracy when using inertial navigation alone.It combined with unscented Kalman filter algorithm(UKF)and Sage-Husa adaptive filter algorithm.The other noise could dynamically update when one of system noise or observation noise was known.The simulation experiment of the fusion filter algorithm and the UKF algorithm proved that AUKF had better positioning accuracy and stability.Thirdly,the traversal path planning algorithm of ROV was studied.A standard underwater raster map was established based on the underwater environment.Traditional biologically inspired neural network traversal path planning algorithm(BINN)has the problems that the repeat coverage rate is too high and the path between sub-regions is not optimal.Therefore the traversal path planning method of a biological incentive based on internal spiral search was proposed.After the region was divided,the inner spiral algorithm was used to complete the traversal of the sub-region.The optimal path planning was achieved between the sub-regions through the A* algorithm.It was proved that the path generated by this method had significant improvements in the repeat coverage rate,running time,and path length indicators when it compared with the original method through simulation experiments.Finally,the developed robot was tested on land.The water surface monitoring program and related sensors were debugged along with network communication and the underwater main control board.Serial communication with the control box information acquisition board,video connection test of the robot camera and graphical display of the inertial navigation data were debugged too.Experiments showed that the expected goals could be achieved.Soon after the semi-physical simulation of path planning and underwater positioning methods was carried out,and the robot could better complete the tasks of positioning and path planning.