Research On Risk Assessment And Control Strategies For Intelligent Ships Navigation

In recent years,with the development of artificial intelligence,big data,and new generation information technology,Maritime technology has also undergone significant changes.In terms of the development of intelligent ships,more and more countries are investing large amounts of manpower and material resources in research,testing and application,with a view to occupying the commanding heights of intelligent ship technology.With the development of intelligent ships in the future,the navigation environment at sea will also undergo significant changes,from traditional manned driving at sea to mixed scenarios of manned and unmanned driving at sea,and ultimately to completely unmanned driving at sea.Although it needs to go through a long period of time,it will be the inevitable result of economic and technological development.It is of great theoretical significance and practical value to study the identification,evaluation,coupling and evolution mechanism of the risks of maritime navigation of intelligent ships,and then study their control strategies.This thesis takes intelligent ships as the research object,and combines system theory,coupling theory,and catastrophe theory to explore and study the risk identification,evaluation,coupling,evolution,and control strategies of intelligent ship navigation at sea.This thesis mainly conducts research work from the following five aspects.(1)Research on the construction method of intelligent ship navigation risk identification and evaluation index system.Based on theories and methods such as Delphi,entropy,and correlation coefficients,and based on the principles of risk factor identification,a risk factor identification process for intelligent ship navigation is established from the four aspects of“ ship remote control center environment emergency management”.The risk factors are identified according to their importance and global relevance,and key risk factors that may exist during the intelligent ship navigation process are identified.Based on this,a method for constructing an intelligent ship navigation risk evaluation index system based on the Delphi entropy improved correlation coefficient model is proposed.The intelligent ship navigation risk evaluation index system including ships,remote control centers,environment,and emergency management is constructed,and the concept of overall identification of the evaluation index system is proposed based on entropy theory to test the rationality of the constructed evaluation index system.(2)Research on risk assessment methods for intelligent ship navigation.Based on the constructed evaluation index system,a method for intelligent ship navigation risk assessment based on DS fuzzy weighted distance Bayesian network is proposed,which fully considers the similarities and differences between intelligent ships’ and traditional ships’ navigation risks.In response to the problems of lack of historical case data,high uncertainty,numerous influencing factors,and difficulty in determining the logical relationship between factors in intelligent ships navigation risk assessment,this thesis proposes a method for intelligent ships navigation risk assessment based on DS fuzzy weighted distance Bayesian network.This method combines the extracted factors that affect the navigation risk of intelligent ships to determine the hierarchical input nodes of the Bayesian network,quantify the network input parameters,and construct a partial network topology structure;Different from traditional risk assessment methods that rely on historical case data,an expert knowledge fusion method based on DS evidence theory is proposed to construct the remaining Bayesian network topology structure;By introducing fuzzy set valued statistical theory to calculate the influence weight of parent nodes,a conditional probability automatic allocation algorithm based on fuzzy weighted distance is proposed to effectively reduce the complexity of parameter calculation for intermediate nodes;Based on the constructed evaluation model,a cascaded risk quantification analysis mechanism is further designed using a fuzzy comprehensive evaluation strategy to ensure the quantification and visualization of risk assessment results.Finally,combined with practical intelligent ship navigation cases,a risk assessment simulation experiment based on DS fuzzy weighted distance Bayesian network was conducted to verify the effectiveness of the proposed method.(3)Study the coupling mechanism of risk factors in intelligent ship navigation.Based on historical maritime accident data and expert questionnaire surveys,as well as system theory and coupling theory,the N-K model and entropy TOPSIS coupling coordination model are used to measure the coupling effect of intelligent ship navigation risk factors,study its coupling mechanism,and compare the advantages and disadvantages and applicability conditions of the two models mentioned above.The results indicate that both methods can be used for coupling analysis of intelligent ship navigation risks,but there are differences in time considerations,coupling coordination development,and coupling state regulations.The former focuses more on the coupling and coordinated development of navigation,while the latter focuses more on the interrelationships of risk factors throughout the entire navigation.Using these two methods can better analyze the coupling mechanism of intelligent ship navigation risk factors.(4)Study the risk evolution mechanism of intelligent ship navigation.By integrating the risk sources,risk receptors,risk factors generated by risk sources,and catastrophe types in risk evolution in a certain accident into a tree network structure,this structure is defined as an intelligent ship navigation risk accident catastrophe tree.On this basis,folding,cusp,swallowtail,and butterfly catastrophe models for the risk evolution of intelligent ship navigation were derived and constructed,and the risk potentials,equilibrium surfaces of risk evolution,and bifurcation sets composed of control variables of these four models were analyzed.We studied the characteristics of each catastrophe model,including multimodality,sudden jumps,unreachability,divergence,and hysteresis.Finally,taking an intelligent ship collision accident as the background,a case simulation study of folding catastrophe and cusp catastrophe was conducted using a navigation simulator.(5)Research on risk management strategies for intelligent ship navigation.Based on the above research results,the connotation and principles of intelligent ship navigation risk management and control are elaborated,and the goals and processes of risk management and control are provided.Specific control strategies were proposed from three aspects:decision-making,management,and operation.Specific risk control suggestions were provided from three aspects: risk evaluation,risk coupling,and risk evolution.The research conclusion of this thesis lays a solid theoretical foundation for identifying,evaluating,and controlling the risks of intelligent ship navigation,and to some extent meets the urgent needs of the shipping industry to clarify and improve the safety of intelligent ships.At the same time,it can provide reference for the safety design of intelligent ships,effectively ensuring the safe operation of intelligent ships.