The Strategy Research Of Ship Intelligent Automatic Collision Avoidance

Ship collision accident occupies a large proportion in the maritime traffic accidents.With the development of maritime trade and people's growing demand for marine resources,a large number of ship collision accidents occur every year,which brought the world a huge loss of life and property,and severely damaged the original marine ecological environment.How to avoid collision among ships is an important part of the ship to ensure safe navigation and has been the international navigation society attaches great importance to the problem.Thereinto,reducing errors manipulation is pivotal to improve the safety of navigation and cut down ship collision accidents.Currently,countries all over the word are seeking effective auxiliary decision system of ship intelligent collision avoidance in order to reduce collision accidents at sea.That is why,intelligent collision avoidance assist research methods of ship are carried out based on detailed analysis of existing research status at home and abroad collision avoidance systems and collision avoidance methods.The main works are completed as following:Firstly,the judgment of the risk of collision is the precondition of ship collision avoidance.Therefore,the model of ship collision risk index(CRI)is established by fuzzy membership function,which lays the foundation to the determination of single ship collision avoidance opportunity.Aiming at the problem that collision risk is only for single-ship collision avoidance dangerous judgment,dangerous target ship and ship dangerous sort under the case of multi-ships can rarely achieved accurately and fleetly.Based on grey correlation degree to determine key ship and the analysis of relationship between factors affecting the risk of collision and risk of collision,and safe distance approach among own ship and targets ships is used in grey correlation decision,optimization of grey relational analysis model is used to select dangerous target ship and determine the risk ranking in the case of multi-ships encounter,the accuracy and superiority of optimized multi-ships encounter risk model are verified by comparison with another methods.Meanwhile,the extensible interactive interface of risk of collision sorting for multi-ships encounter is established to embody the practicability of model,which lays a foundation for the realization of ship collision avoidance.Secondly,aiming at the problem that weight coefficient selection uncertainty in ship collision avoidance scheme evaluation leads to decreased accuracy of the assessment program and applicability of evaluation scheme can not be used widely,a deterministic collision avoidance scheme evaluation mathematical model for ships encounter is proposed in thispaper,the collision avoidance process of obliged ship is divided into five stages,and every stages is analyzed based on the certain collision impact parameters,the feasibility of collision avoidance scheme can be determined by assessment parametric curves which are changed as the different collision avoidance stages,and that the validity of the evaluation mathematical model of collision avoidance is confirmed by comparing to actual collision avoidance movement modeling,the deterministic collision avoidance scheme evaluation mode provides a reliable guarantee for the assessment of implementation of ship automatic collision avoidance scheme.Thirdly,aim at the issue that ship automatic collision avoidance schemes based on existing heuristic algorithms exhibit the characteristics of slowing convergence speed and low accuracy of uncertain solving,which lead to generate collision avoidance scheme slower and influences the crew analyzing and judgment.Bacterial foraging algorithm with the advantage of fast parallel search speed is improved by particle swarm optimization and adaptive replication bacterial to enhance global searching ability and quick local convergence ability in this paper,the strategy of ship intelligent automation collision avoidance based on optimized bacterial foraging algorithm is realized,collision avoidance scheme including the parameters of collision avoidance angle,reorientation angle,collision avoidance time and reorientation time can be gained under the condition of bacterial foraging coding and objective function of collision avoidance.To the end,the ship collision avoidance simulation platform under different encounter situations is designed in a simulation software development environment.Optimization collision avoidance parameters are chosen and displayed,and a dynamic route of ship collision avoidance is generated.Based on the deterministic collision avoidance scheme evaluation,the effectiveness and feasibility of ship collision avoidance scheme are verified.Fourthly,aim at the problem that collision avoidance guidance is not provided in COLREG collision avoidance decision under multi-ships encounter,as a result,some difficulties are caused to decision-making and evaluation of multi-ships collision avoidance problems,and navigation and control field of multi-ships encounter is still a scabrous problem.Based on optimization bacterial foraging algorithm and realization of two-ship collision avoidance,the multi-ships collision avoidance strategy of “circulatory avoidance for dangerous ships” is obtained in this paper,on the premise of the grey correlation analysis,the strategy is carried out by determining key collision avoidance ships and re-anglicizing with target ship encounter situation among the own ship and target ships,based on collisionavoidance objective function with the characteristic of “economic” and “safe”.Finally,a series of optimal steering angles of our ship collision avoidance are gained by dynamic ship collision avoidance simulation test,and effectiveness and feasibility of the strategy are verified by deterministic collision avoidance scheme evaluation.This research proposes a new idea for intelligent avoidance collision under multi-ships encounter conditions.Finally,In view of shortcoming that due to the intelligent algorithm of random variables,the strategy of ship intelligent collision avoidance which employs existing intelligent algorithm can not show the consistent collision parameter after collision avoidance action,the strategy of ship collision avoidance path planning based on adaptive wolf colony search algorithm is raised in the paper?Compared to other intelligent algorithm,the wolf colony search algorithm(WCA)has characteristics of good global convergence and excellent calculation robustness.Besides,to accelerate the convergence of the quality and speed of search collision avoidance scheme,wolf basic step of each direction dimension in the search space is calculated,and search step length and moving step length of every artificial wolf are adjusted in order to avoid cross-border;meanwhile adaptive surrounded step length is also adopted to avoid falling into local optimum.at last,the feasibility of the scheme is verified by simulation collision avoidance movement of own ship and target ship;effectiveness is verified by the deterministic collision avoidance scheme evaluation;and superiority is verified by comparison with other algorithms from time of collision avoidance generation and collision avoidance stability,in addition,the interfaces of dynamics simulation between own ship and target ship are achieved,the research of this paper open up the new ideas for ship automatic collision avoidance system development.