Multi-dimensional high-precision track intelligent prediction method based on line segment clustering

A technology of intelligent prediction and track prediction, applied in the field of target tracking

CN111858565AActive Publication Date: 2020-10-30BEIHANG UNIV
6 Cites 4 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Publication Date
2020-10-30

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a multi-dimensional high-precision track intelligent prediction method based on line segment clustering. The method includes discretizing the continuous track data; carrying out abrupt change longitude data processing, data cleaning and normalization processing; compressing track data by using a Douglas-Peucker algorithm; using a DBSCAN clustering algorithm for clustering flight paths, selecting flight path clusters corresponding to emergency situations according to different emergency situations under multi-dimensional factors, performing flight path prediction througha flight path prediction neural network model, and finishing a multi-dimensional high-precision flight path prediction task. According to the invention, the original flight path data is compressed, so that the calculation pressure is greatly reduced under the condition of reserving flight path characteristics, the operation time is shortened, and the operation efficiency is improved; a convolution and LSTM neural network model is adopted, convolution is used for feature extraction, and the track prediction precision of the LSTM neural network model can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical field

[0001] The invention relates to the technical field of target tracking, in particular to a multi-dimensional high-precision track intelligent prediction method based on line segment clustering. Background technique

[0002] For the specific task of mastering the trajectory of a suspicious moving target, on the one hand, it is necessary to learn the law of the suspicious moving target’s trajectory and predict the future trajectory of the moving target with high accuracy. On the other hand, it is necessary to consider multi-dimensional factors under realistic conditions. Such as emergencies, international hot spots, weather conditions, etc., the track information of moving targets can be predicted with high accuracy.

[0003] The method of collecting track data of moving targets (such as ships) in the ocean is usually to obtain the data information of the current position of the ship intermittently. According to the different state of the ship, the position is broadca...

Examples

Embodiment 1

[0104] The first step is to train a neural network model for track prediction.

[0105] (1) According to the track characteristics of the moving target on the sea, choose the appropriate time interval to discretize the original continuous historical track data. After abrupt longitude data processing, data cleaning and normalization, the original training set is formed.

[0106] Since the input and output of the neural network algorithm used in model training of historical track data are usually discrete data, it is necessary to use an equidistant method to discretize the original continuous historical track data to facilitate subsequent Input the track prediction neural network model for training, the time interval is Δt. If the value of Δt is too large, the track data will lose more track information; if the value of Δt is too small, it will make the track data too dense and increase the calculation pressure. Therefore, it needs to be based on the characteristics of the actual tra...