CN110376574B

CN110376574B - Target tracking method based on multi-base passive sonar observation data

Info

Publication number: CN110376574B
Application number: CN201910599056.6A
Authority: CN
Inventors: 韩一娜; 杜力; 赵伟康; 杨益新
Original assignee: Northwestern Polytechnical University
Current assignee: Northwestern Polytechnical University
Priority date: 2019-07-04
Filing date: 2019-07-04
Publication date: 2023-01-17
Anticipated expiration: 2039-07-04
Also published as: CN110376574A

Abstract

The invention relates to a target tracking method based on multi-base passive sonar observation data, which is a tracker with good concealment, strong functions and stable performance, wherein the tracker can quickly converge the estimation of the current state of a target based on the measurement data of passive multi-sensors, and the convergence effect is very stable.

Description

Target tracking method based on multi-base passive sonar observation data

Technical Field

The invention belongs to the field of sensor information fusion, and particularly relates to a target tracking algorithm based on multi-base passive sonar observation data.

Background

Passive sonar detection locates a target by receiving noise radiated by the target in the water. Because the sonar does not emit signals, the sonar has the advantages of good concealment, difficulty in being attacked and the like, and is concerned. The single passive sensor can only observe the direction of a target, cannot obtain distance information, belongs to incomplete observation, information of a plurality of sensors needs to be fused to complete a target tracking task, and a target tracker based on passive measurement has no more complete flow form description, so that the research on the passive multi-sensor tracker is very necessary.

Disclosure of Invention

The technical problem solved by the invention is as follows: in order to make up for the defects of the prior art, the invention designs a target tracking method based on multi-base passive sonar observation data, re-describes a motion model and an observation model in a target tracker, and obtains the target tracking method which is suitable for receiving passive measurement and has reliable performance.

The technical scheme of the invention is as follows: a target tracking method based on multi-base passive sonar observation data comprises the following steps:

the method comprises the following steps: defining an application scene: a plurality of passive observation sonar base stations are distributed in a space in a distributed manner, nodes of each base station form a network in a self-organizing manner, a plurality of targets are arranged in an area among the plurality of sonar base stations, the targets radiate acoustic signals to the surroundings indiscriminately, the passive observation sonar base stations receive the acoustic signals, and the relative direction and the receiving frequency of the targets are obtained by using output results of signal processing (filtering, beam forming and matched filtering);

step two: according to the application scene in the first step, a target motion model and an observation model are created, wherein the target motion model is an approximate uniform velocity model, and the observation model is a nonlinear system;

step three: designing a nonlinear filter according to the motion model and the observation model in the step two, wherein the design is as follows: an extended Kalman filtering method is applied to convert the nonlinear system into an approximately linearized model around the filtered values. When the filter performs measurement data correlation, there are two timing structures: the measurement data of a plurality of directions returned at the same time are correlated, and the correlation of the measurement data continuously returned in the uniform advancing time sequence can be summarized into the same type of problems to be solved, and finally, a complete target tracking process is obtained, and the position coordinates of the target can be tracked.

The further technical scheme of the invention is as follows: the target motion model is an approximate uniform velocity model.

Effects of the invention

The invention has the technical effects that: based on the measured data of the passive multi-sensor, the tracker can quickly converge the estimation of the current state of the target, and the convergence effect is very stable, so that the tracker has the advantages of good concealment, strong function and stable performance.

Drawings

FIG. 1 is a schematic diagram of passive observation of a target by multiple sensors;

fig. 2 is an effect diagram of the present invention.

Detailed Description

The process of estimating the current state parameters of the tracked target and processing the measurement data received by the sensor is called a target tracking process.

Referring to fig. 1-2, a solution for applying passive sensor data to a target tracking algorithm framework includes creation of a motion model and an observation model, and processing of obtaining measurement data with respect to different time sequences under multi-sensor conditions since the observation model is a nonlinear system that needs to be converted into a linear system. The method is characterized in that: according to the specific application scene, establishing corresponding observation and motion models; redesigning the nonlinear filter based on the nonlinear observation model; the application scene matched with the multiple sensors adopts a time sequence structure different from a traditional filter.

The target state comprises the position, the speed and the radiation frequency of the target; the measurement of the target comprises the relative direction and the receiving frequency of the target; the doppler effect is taken into account in the reception frequency. Because the passive tracking measurement system is nonlinear, the linear Kalman filter used in the traditional target tracking process is not applicable, so the method is replaced by a classic nonlinear filter, namely an extended Kalman filter.

In the scenario observed by a passive multi-sensor tracker, measurement data received by a sensor is to be correlated, and two tasks of data correlation are subdivided, one is the correlation of multiple azimuth measurements returned at the same time, and the other is the correlation of measurement in time sequence, a time difference Δ t between frames in a filter is a variable controlled by a data source, and for data from different sensors at the same time, Δ t may be 0, so that the two tasks of correlation can be actually classified into one task.

The main contents of the invention are:

1. the application scenario of the invention is as follows:

the passive observation scene of multiple sonar (sensor) described in the present invention is shown in fig. 1, where multiple fixed sonar base stations for passive observation are distributed in space, and a target radiates acoustic signals to the surroundings without distinction, and the passive base stations receive the signals and use a signal processing algorithm to obtain some data (measurement) related to the state of the target, including the relative azimuth and receiving frequency of the target. These data can be used as the basis for the tracker to find and track the target.

2. The target motion state has the characteristics of unpredictable and random changes, and a target motion model of the current tracking system may not accord with the actual target motion situation, and these factors increase the difficulty of target state estimation and prediction, so that it is important to establish a suitable target motion model. A common mathematical model of the moving object is given here: the near uniform velocity (NCV) model.

Step 1: the motion model can be described as:

X _k+1 ＝Φ _k X _k +ω _k

wherein X is a system state vector including a position coordinate (X) _k ,y _k ) X-axis, y-axis velocity components

Frequency information

X _k Is the system state vector at time k, X _k+1 Is the system state vector at time (k + 1), Φ _k Is the state transition matrix at time k, ω _k Zero mean at time k, variance Q _k The Gaussian process noise is that the target motion speed is modeled into a wiener process, and then the uncertain speed parameters on the x axis and the y axis are q respectively _x ,q _y The frequency uncertainty parameter is q _f To obtain a covariance matrix Q _k The specific expression of (1).

Step 2: the observation equation:

passive multisensor returns the azimuth and frequency of the target, assuming Z _k Is a system observation vector at the time k, the azimuth of the target at the time kObservation of theta _k And frequency observation f _k Are independent of each other and have Gaussian errors of

v _k Representing the observation noise, c represents the speed of sound.

Indicating the coordinate position of the observation station, r _k Representing the distance of the observation station from the target, d _k /r _k Representing the radial velocity of the target with respect to the observation station.

In the formula, the atan2 function represents a two-parameter arc tangent function with a value range of (-pi, pi ].

And step 3: for the active tracker, the first measurement can be directly converted into position coordinates, and the covariance of the coordinates is used to construct the matrix P (1), for the passive measurement, the first measurement is the azimuth and frequency information, rather than the direct position coordinate information, and the estimation error is large because the position coordinate information is fuzzy, so that the initial value of P (1) and n 1 needs to be set large.

The initial state is described as follows:

wherein X (1Y 1) is the initial time system prediction state vector, and the position coordinate at this time is assumed to be (X) ₁ ,y ₁ ) Velocity of 0 and frequency of

P (1 < 1 >) represents the covariance matrix of the estimation error, where

Is a gaussian error of the position coordinates,

is a gaussian error in the velocity of the beam,

is the error in frequency.

And 4, step 4: non-linear filtering

Step 2, the direction and the frequency are obtained from the sensor, and the H matrix in the observation equation is nonlinear, so that a nonlinear filter, namely the most classical extended Kalman filtering, is used for approximating a nonlinear system to a linear system, the system prediction is described again, and the updating process is deduced as follows:

x (k + 1) is a system prediction state vector at the time k +1, and P (k + 1) is a covariance matrix representing a prediction error.

The linearization of the H matrix in the observation equation is as follows:

wherein

Wherein

J (k + 1) represents the equivalent observation matrix for the k +1 th time sequence, which is the partial derivative of the observation function H.

P(k+1|k+1)＝(I-L(k+1)J(k+1))P(k+1|k)

X (k + 1) is a state update vector at the moment of k +1, and P (k + 1) denotes that the state vector update step corresponds to the covariance matrix.

A problem with the tracker referred to in 2 is that it cannot converge if the measurements are always from the same fixed sonar. This problem can be avoided when there are multiple fixed sonar bss, and the method dynamically adjusts the timestamp of the filter, and the time delay Δ t varies only when the measured time variation is obtained, otherwise, the sensor data Δ t =0 at the same time obtained by different bss. (this paragraph is a limiting condition for Δ t in the filter)

Implementation example: this example demonstrates mainly the working effect of the filter under non-uniform timing, with the sensors distributed at (0, 0) m (0, 5000) m (5000, 0) m (5000 ) m, and the target moving from (1000 ) m to the upper right. The filter acquires the orientation of the target from the sensor. Sensor estimation targetError of orientation sigma _θ ＝0.1°,q _x ＝q _y =300. As can be seen from fig. 2, the solid line represents the real track of the target, the dotted line represents the filtered tracking result, and for the arbitrarily set initial filtering value (1500 ) m far from the target, the solid line and the dotted line are rapidly overlapped, so that the filter can be rapidly converged, and in the subsequent tracking process, the two lines can be always well overlapped, i.e., the error between the tracking result and the real track is small, and the tracker can well track the target.

Claims

1. A target tracking method based on multi-base passive sonar observation data is characterized by comprising the following steps:

the method comprises the following steps: defining an application scene: a plurality of passive observation sonar base stations are distributed in a distributed manner in space, nodes of each base station form a network in a self-organizing manner, a plurality of targets are arranged in an area among the plurality of sonar base stations, the targets radiate acoustic signals to the surroundings without distinction, the passive observation sonar base stations receive the acoustic signals and obtain the relative direction and the receiving frequency of the targets by using the output result of signal processing, and the method for signal processing comprises filtering, beam forming and matched filtering;

step two: according to the application scene in the first step, a target motion model and an observation model are established, wherein the target motion model is an approximate uniform velocity model, and the observation model is a nonlinear system;

step three: designing a nonlinear filter according to the motion model and the observation model in the step two, wherein the design is as follows: applying an extended Kalman filtering method to convert the nonlinear system into an approximate linearized model around the filtered value; when the filter performs measurement data correlation, there are two timing structures: the measurement data of a plurality of directions are returned at the same time for correlation, and the correlation of the measurement data continuously returned in uniform forward time sequence is solved by solving the same problem, finally a complete target tracking process is obtained, and the position coordinate of the target can be tracked;

step four: when the measured data correlation is performed in the filter in the third step, two timing structures exist: the measurement data of a plurality of directions are returned at the same time for correlation, the measurement data which are continuously returned in uniform forward time sequence are correlated, a correlation frame with universality is designed to complete two different correlation tasks by setting the time difference delta t between frames according to different time sequence structures, and finally a complete target tracking process is obtained and the position coordinates of a target can be tracked.