An Adaptive Interactive Multi-model Maneuvering Target Tracking Method

Abstract

The invention relates to the mobile tracking of vehicle objects in the field of transportation. Establish the mixed initialization input of each model, including the mixed initial conditions of each model and the covariance matrix of the mixed initial state; establish the uniform velocity (CV) and uniform acceleration (CA) motion model; the update refers to the covariance matrix of the calculation error according to the Kalman filter equation and Innovation; use the innovation in the Kalman filter results to construct the likelihood function of the target motion model, and calculate the Markov state transition probability matrix; use the Markov state transition probability matrix as the weight of switching between motion models for fusion estimated output. The invention solves the problem of error increase or loss of tracking caused by the mismatch between the filter model and the target motion model caused by the maneuvering of the target in the traditional interactive multi-model algorithm, and has the advantages of small computational complexity and good tracking effect. It can be used for target tracking of motor vehicles in the field of transportation.

Description

technical field [0001] The invention belongs to the field of maneuvering target tracking, and in particular relates to an adaptive interactive multi-model maneuvering target tracking method in the field of transportation when a moving car suddenly turns, accelerates, decelerates, and the like. Background technique [0002] Target tracking is to filter the target movement data received by the detection sensor combined with different observation sets generated by various uncertain information sources, and estimate the state parameters of the moving target, such as the distance, orientation, speed, acceleration, etc. of the target. [0003] The single-model maneuvering target tracking algorithm only performs well on non-maneuvering targets. When the target maneuvers, the tracking performance of the system decreases, and even causes the target to be lost. In order to track the maneuvering target, it is necessary to establish a reasonable motion model according to the motion cha...

AI Technical Summary

Problems solved by technology

In fact, when the motion model of the target has a certain tendency, the traditional IMM algorithm only realizes the "synthesis" between the motion models by mediating the weighting of the posterior probability of the motion model under different observation vector conditions, and does not take into account the Marl Irrationality of Cove's Transition Probability Matrix Design

[0004] In the field of transportation, when a moving car suddenly turns, accelerates, and decelerates, it also belongs to the maneuver of the target. Especially when the vehicle target suddenly enters a curve, the vehicle's motion state will be greater than that of the previous uniform speed or uniform acceleration linear motion. If it cannot effectively adapt to the maneuvering state of the target, it is easy to cause the loss of the target, and when the target is maneuvering, it is the moment when traffic accidents are most likely to occur

Images