船舶模拟驾驶系统障碍物自动识别方法  

Automatic obstacle identification method for ship simulation system

作　　者：刘晓峰[1] LIU Xiao-feng(Institute of Navigation Technology,Jiangsu Maritime Institute,Nanjing 211170,China)

机构地区：[1]江苏海事职业技术学院航海技术学院,江苏南京211170

出　　处：《舰船科学技术》2022年第22期144-147,共4页Ship Science and Technology

基　　金：江苏省现代教育技术研究课题(2021-R-91015)。

摘　　要：本文研究船舶模拟驾驶系统障碍物自动识别方法,满足船舶模拟驾驶系统在未知环境下的避障需求。船舶模拟驾驶系统的电子导航雷达单元,利用激光雷达传感器采集船舶环境信息的激光点云数据;控制中心依据所采集激光点云数据,通过自适应距离阈值聚类法聚类激光点云数据,提取障碍物特征向量。将提取障碍物特征向量作为支持向量机的输入,利用粒子群优化算法确定支持向量机的最优核参数,利用设置最优核参数的支持向量机,输出船舶模拟驾驶系统障碍物自动识别结果。实验结果表明,船舶模拟驾驶系统采用该方法,自动识别模拟驾驶时的静态障碍物以及动态障碍物,满足船舶安全航行需求。This paper studies the automatic obstacle identification method of the ship simulation driving system to meet the obstacle avoidance requirements of the ship simulation driving system in unknown environment.The electronic navigation radar unit of the ship navigation simulation system collects the laser point cloud data of ship environmental information by using lidar sensors.According to the collected laser-point cloud data,the control center uses adaptive distance threshold clustering method to cluster the laser-point cloud data and extract the obstacle feature vectors.The extracted obstacle feature vectors were set as the input of the support vector machine,and the optimal kernel parameters of the support vector machine were determined by the particle swarm optimization algorithm.The automatic obstacle recognition results of the ship simulation system were output by the support vector machine with the optimal kernel parameters.The experimental results show that this method can automatically identify static and dynamic obstacles during simulated driving,and meet the requirements of safe navigation of ships.

关 键 词：船舶 模拟驾驶系统 障碍物 自动识别方法 支持向量机 粒子群 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]