Precise tracking method based on nerve network for moving target

Abstract

The invented method for accurate tracking the maneuvering target based on neural net adopts double parallel structure being composed of two filters. The difference of states between the parallel double filters is picked-up as the character vector of the state of target to be estimate, which is input to the neutral net system. Based on the character vector input, the neural net system, which has been trained off-line before tracking estimation, outputs the predicted value of the acceleration variance of the target to be estimated at present time. Adaptive filtering the said predicted value based on the current statistical model obtains the fast and accurate tracking the target to be estimate. The invented method possesses favorable adaptive tracking ability.

Description

Technical field: [0001] The invention relates to a precise tracking method of maneuvering targets based on neural network (NN), which is used for high-precision positioning and prediction of maneuvering targets in systems such as intelligent transportation, robots, avionics, anti-ballistic missile defense and precision guidance, and belongs to intelligent information processing technology field. Background technique: [0002] For the precise tracking of maneuvering single targets, many maneuvering target models and adaptive tracking algorithms have been proposed at home and abroad since the 1980s, including differential polynomial models, uniform velocity and uniform acceleration models, time-dependent models, semi-Markov models, and maneuvering target models. "Current" statistical model (Zhou, H. and K.S.P. Kumar. A current statistical model and adaptive algorithm for estimating maneuvering targets. ALAA Journal of Guidance, Control, and Dynamics, 1984, 7(5), 596-602.). Th...

AI Technical Summary

Problems solved by technology

In order to overcome this difficulty, the interactive multi-model adaptive tracking method (Blom, H.A.P., and Bar-Shalom, Y., The interacting multiple model algorithms for systems with Markovian switchingcoefficients. IEEE Tran. Automatic Control, 1988, 33(8) , 780-783) was proposed. Although this method improves the tracking performance to a certain extent, the speed and accuracy of its tracking response are far from the requirements of the actual system, especially the precise tracking system.

Images