Target track prediction method based on cycle neural network

Abstract

The invention discloses a target track prediction method based on the cycle neural network, belongs to the radar target tracking field and mainly solves problems of simple model, low complexity, pooruniversality and learning incapability existing in a track prediction method in the prior art. The method comprises steps that firstly, the same type of radar measurement traces and tracking track data in various scenarios are collected, cooperative target track is collected through employing cooperative target information receiving equipment, and the data is eliminated and corrected to form an original track data set; secondly, the target track prediction cycle neural network is constructed, a training sample characteristic vector is set, and a track training set is generated; and lastly, based on a cooperative track training set and a radar track training set, the target track prediction cycle neural network is trained and optimized, and a target track prediction method matched with a radar is generated. The method is advantaged in that the prediction algorithm is generated through automatic training, the application scope is wide, adaptation scenes are a lot, and the practical effect is good.

Description

technical field [0001] The invention belongs to the field of radar target tracking, and provides a target track prediction method, relates to the construction, training and generation of a cyclic neural network, and is suitable for radar data processing systems. Background technique [0002] Target tracking is the core and key technology of radar data processing. By establishing the correspondence between each frame of radar measurement data and different real targets, and filtering and estimating, the target's motion trajectory and motion parameters are obtained, so as to achieve real-time continuous control of the target individual. . Target tracking includes multiple technical links such as track initiation, track filtering, track prediction, and point-to-point correlation, among which track prediction is a key bottleneck link, which plays a connecting role and directly determines the target tracking effect. If the track prediction is inaccurate, it will affect the resul...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to propose a target track prediction method, aiming to solve the problems of the existing target track prediction method model simple, low complexity, poor adaptability and unable to learn

Images