Fuzzy Operator Design And Application Research For AUV Multi-objective Path Planning

Relying on large area trawler to catch aquatic organisms is easy to destroy the und erwater ecological environment,so using autonomous underwater vehicle（AUV）to c atch aquatic organisms has become a research hotspot.AUV fishing aims to identify t he target in the complex underwater environment with strong noise and weak light,an d plan the driving path in the environment with uncertain information so as to approac h the target and catch it accurately.At present,most fishing auvs face the following three main problems:（1）the reco gnition and perception ability of target recognition algorithm is poor in the underwater environment with weak light and strong noise.（2）The underwater environment is co mplex,not only the underwater information is uncertain,but also the data collected by the machine is not accurate.（3）As fishing robots mainly rely on manual control to ca rry out fishing operations,it is difficult to expand to autonomous fishing by machines.Aiming at these three problems,an underwater image enhancement method is propos ed to improve the recognition probability.Then fuzzy operators are used to better deal with the uncertain information in path planning.Finally,the application of fuzzy operator in multi-objective path planning of fishing AUV is further explored,and a complete fishing AUV system is proposed to meet the realistic demand.To sum up,the research work of this topic has the following three contributions:（1）This paper designs an underwater image enhancement algorithm based on adapt ive contrast stretching and PST.Firstly,adaptive contrast stretching was carried out fo r different color channels to alleviate the problem of image degradation caused by sele ctive absorption of light by seawater.Then,PST algorithm was used to obtain the edg e contour features of aquatic organisms,and convolution was used to remove the nois e points.Finally,the front and back scenes are separated based on contour features,an d the foreground objects are stretched nonlinear to highlight the objects.Experimental results show that the method used in this paper can better enhance underwater images with weak light and strong noise,and make underwater target recognition more accur ate.（2）This paper proposes an underwater path planning method based on IVIFWPMM_DST.Firstly,the IVIFWPMM_DSToperator is obtained by combining PA and MM operators,which is used to aggregate interval intuitionistic fuzzy numbers in THE DST framework.Then,according to the proposed IVIFWPMM_DSToperator,a fuzzy sorting method for multi-objective path planning is proposed.The optimal target direction can be obtained by using the proposed sorting method.Finally,according to the optimal target direction and underwater current information,the driving direction and speed are obtained.The experimental results show that this method not only overcomes the shortcomings of traditional methods in processing complex fuzzy information quickly,but also can deal with the shortcomings of experts’lack of experience,and the algorithm also meets the time and space requirements.（3）This paper designs a fishing AUV system based on IVIFWPMM_DST.Firstly,the SFR based algorithm is used to remove unclear video frames,then the underwater image enhancement method based on adaptive contrast stretching and PST is used to obtain the enhanced image,and then the target is identified,and then the local path planning method is used to plan the real-time path.Finally,closed-loop search and open-loop capture methods were designed to capture aquatic organisms.The simulation test shows that the machine can smoothly capture all water creatures in the operation area,and the capture speed is increased by about 9%.