依托船舶自动识别系统(Automatic Identification System,AIS)数据,利用云计算并结合聚类算法,对船舶历史数据进行轨迹聚类分析,构建船舶航行正常轨迹模型,为实时检测船舶异常轨迹奠定基础,进而为提高水上交通监管智能化水平提供新方法...依托船舶自动识别系统(Automatic Identification System,AIS)数据,利用云计算并结合聚类算法,对船舶历史数据进行轨迹聚类分析,构建船舶航行正常轨迹模型,为实时检测船舶异常轨迹奠定基础,进而为提高水上交通监管智能化水平提供新方法。针对目前轨迹聚类算法效率低等问题,基于Spark内存计算技术及数据分区思想,提出一种改进的并行子轨迹聚类算法SPDBSCANST(Parallel DBSCAN of Sub Trajectory Based on Spark)。以长江航道武汉段船舶航行数据为例进行试验验证,并通过可视化方式呈现。结果表明,改进后的算法的聚类效率和效果都有明显提升。展开更多
The study on artificial intelligence(AI) methods for tuning of particle accelerators has been reported in many literatures.This paper presents tuning method for agent-based control systems of transport lines in the ca...The study on artificial intelligence(AI) methods for tuning of particle accelerators has been reported in many literatures.This paper presents tuning method for agent-based control systems of transport lines in the case of sensor/actuator failures.The method uses model-based tracking concept to relax the demand on sensor data.The condition for successful operation of the stated scheme is derived,and the concept is demonstrated through simulation by applying it to the model of microtron,transport line-1 and booster of indus accelerator.The results show that this approach is very effective in transport line control during sensor/actuator failures.展开更多
文摘依托船舶自动识别系统(Automatic Identification System,AIS)数据,利用云计算并结合聚类算法,对船舶历史数据进行轨迹聚类分析,构建船舶航行正常轨迹模型,为实时检测船舶异常轨迹奠定基础,进而为提高水上交通监管智能化水平提供新方法。针对目前轨迹聚类算法效率低等问题,基于Spark内存计算技术及数据分区思想,提出一种改进的并行子轨迹聚类算法SPDBSCANST(Parallel DBSCAN of Sub Trajectory Based on Spark)。以长江航道武汉段船舶航行数据为例进行试验验证,并通过可视化方式呈现。结果表明,改进后的算法的聚类效率和效果都有明显提升。
文摘The study on artificial intelligence(AI) methods for tuning of particle accelerators has been reported in many literatures.This paper presents tuning method for agent-based control systems of transport lines in the case of sensor/actuator failures.The method uses model-based tracking concept to relax the demand on sensor data.The condition for successful operation of the stated scheme is derived,and the concept is demonstrated through simulation by applying it to the model of microtron,transport line-1 and booster of indus accelerator.The results show that this approach is very effective in transport line control during sensor/actuator failures.