结合IFF点迹与AIS的民船避警行为识别OA

To enhance the ability of the marine police command information system to detect abnormal vessel behavior,this study addresses the problem of detecting evasion behavior by civilian ships against police vessels.By integrating police vessel Identi-fication Friend or Foe(IFF)point traces and civilian ship AIS data,a detection method based on a PID-BP neural network predic-tion model and evasion coefficients is proposed.The method constructs a PID-BP neural network prediction model to predict the tra-jectory of civilian ships.It then calculates the evasion coefficients of civilian ships relative to police vessels based on their actual and predicted trajectories.By analyzing the differences in these evasion coefficients,the method computes the probability of evasion and compares it with a preset threshold to determine whether a civilian ship exhibits evasion behavior.Experimental results demonstrate that this method effectively identifies evasion behavior by civilian ships and has practical engineering value.