Font Size: a A A

Research On Ship Collision Avoidance Dangerous Degree Based On Neural Network In Complicated Environment

Posted on:2023-05-08Degree:MasterType:Thesis
Country:ChinaCandidate:X F ChenFull Text:PDF
GTID:2531307040482564Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Collisional incidents account for more than half of all marine traffic accidents,according to the Ministry of Transportation’s Maritime Traffic Accident Report.The value of ship collision hazard degree is one of the criteria to judge whether avoidance action should be taken next.Furthermore,the ship collision danger degree can provide a reference for the choice of ship avoidance time.Therefore,the ship collision hazard is one of the key areas of research in the field of maritime traffic.The traditional collision hazard mathematical model only involves the distance between ships,speed and other factors to calculate the ship collision hazard,which cannot ful y reflect the influence of other factors on the ship collision hazard such as the navigation water environment.Especially when the ship navigates in narrow waters with the complex environment such as a harbor,the ship’s maneuvering space is small and the ship’s navigation is influenced by wind,waves,visibility and other environmental factors.The environmental factors of sailing waters also affect the degree of collision risk of ships.With the development of electronic technology,the maritime observation equipment increases,and the navigation data information that can be collected considerably increases as wel.Therefore,it is feasible to consider the influence of environmental factors on ship collisions.The main research work and innovative results are as follows.(1)The existing algorithm for calculating the collision hazard of a ship only considers the influence of the ship’s own factors such as ship speed and distance on the collision hazard.Therefore the influence of speed,distance,ship’s meeting angle,ship’s tonnage,visibility,wind,waves,density of ships in navigable water and other factors on collision are taken into account in the study.The degree of association between eight influencing factors and collision hazard is analyzed by utilizing gray correlation analysis.(2)The ship collision hazard is divided into macro collision hazard and micro collision hazard.The influence of visibility,wind,waves,ship density and other environmental factors on ship collision is cal ed macro collision hazard.The influence of ship speed,distance,ship encounter angle,ship tonnage and other factors on ship collision is cal ed micro collision hazard.The data of the above factors are fuzzified and used as the input of the macro collision hazard neural network and micro collision hazard neural network.And the macro collision hazard and micro collision hazard of the ship are calculated by the neural network model respectively.For the disadvantage of the slow convergence speed of the BP neural network,the whale optimization BP neural network algorithm is designed.The whale algorithm is used to provide the optimal initial weights and thresholds to the BP neural network for macroscopic collision hazard and microscopic collision hazard calculation.(3)Hierarchical analysis and entropy power method are employed to find the weight ratio of macroscopic collision hazard degree and microscopic collision hazard degree.The collision hazard degree and micro-collision hazard degree are fused,and final y,the comprehensive collision hazard degree value of the ship is obtained.(4)In order to reflect the collision risk of the ship more intuitively,the ship collision risk simulation system is designed using MATLAB software.The simulation system can display the ship’s motion,environmental information,macroscopic collision hazard,microscopic collision hazard and comprehensive collision hazard,and the ship pilot can intuitively understand the ship’s collision risk degree.The simulation experiment verifies that the new algorithm for calculating the ship collision hazard can reflect the influence of environmental factors on the ship collision hazard very wel.The algorithm can objectively and realistical y reflect the influence of the navigational water environment on the ship collision hazard,guide the ship pilot to accurately evaluate the ship motion collision risk and reasonably choose the ship collision avoidance action time,which has certain practicality.Especially,it is more effective to use in the case of dense ships in navigable waters and narrow waterways.
Keywords/Search Tags:Fuzzy neural network, Macro collision risk, Micro collision risk, Hierarchical analysis, Entropy
PDF Full Text Request
Related items