Turning rate acquisition method for target tracking

Abstract

The invention provides a turning rate acquisition method for target tracking, which comprises the following steps of: acquiring a state vector of a target at a moment k, and establishing a state equation comprising a turning rate; estimating the turning rate at the moment k to obtain an estimated value of the turning rate; and in combination with the estimated value of the turning rate sum at the k-1 moment, performing weighted summation to obtain the turning rate at the current moment.

Description

Technical Field

[0001] The present invention relates to a radar target tracking technology, in particular to a method for obtaining the turning rate of target tracking. Background Art

[0002] The problem of target tracking widely exists in civil and military fields such as national defense and military, deep space exploration, port collision avoidance, and vehicle navigation. In recent years, with the popularization of "low, slow, and small" aircraft, the increasingly developed ground traffic, and the more complex environment, the maneuverability of the targets detected by radar has become stronger and stronger. When using the conventional single-model filtering tracking algorithm to track maneuvering targets, there is often a problem of mismatch between the tracking model and the target model, which will lead to filter divergence and it is difficult to achieve high-precision tracking of maneuvering targets.

[0003] The interactive multiple model algorithm can better solve the defects of the single-model algorithm because its model set can contain multiple motion models and the models can be switched with each other. It is a relatively effective method and is widely used in the field of maneuvering target tracking. In the maneuvering target scenario, the target will inevitably make a turning motion. Therefore, the model set adopted by the interactive multiple model usually includes a coordinated turn model. However, the traditional coordinated turn model uses a fixed turning rate when calculating the state transition matrix. When the actual turning rate of the target is different from the set turning rate, the state transition matrix of the model cannot be adjusted in time, and the filtering accuracy of the algorithm will be greatly affected, and even the situation of filter divergence will occur, resulting in the algorithm being difficult to adapt to various current high-maneuverability target tracking scenarios. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for obtaining the turning rate of target tracking, including:

[0005] Step S100, obtaining the state vector X(k) of the target at time k, and establishing a state equation including the turning rate;

[0006] Step S200, estimating the turning rate at time k to obtain an estimated value of the turning rate ;

[0007] Step S300, combining the turning rate at time k - 1 and the estimated value of , and performing weighted summation to obtain the turning rate at the current time .

[0008] Further, in step S100, for the state vector at time k, establish the corresponding state equation

[0009]

[0010] where, and respectively represent the distances of the target on the x-axis and y-axis. and respectively represent the speeds of the target on the x-axis and y-axis. represents the turning rate of the target. and represent the acceleration disturbance terms on the x-axis and y-axis, and T represents the sampling interval.

[0011] Furthermore, step S200 includes:

[0012] Step S201, obtaining the filtered estimated values X(k - 1) and X(k) at times k - 1 and k;

[0013] Step S202, obtaining the measurement value z(k + 1) at time k + 1, where the measurement value also includes the positions and speeds on the x-axis and y-axis;

[0014] Step S203, obtaining the distances a(k), a(k + 1), and c(k) of the target points at times k - 1 and k, k and k + 1, and k - 1 and k + 1;

[0015] Step S204, obtaining the estimation of the turning rate

[0016]

[0017] where is the angle between the line connecting the target point at time k and the target point at time k + 1 and the extension line of the line connecting the punctuation point at time k - 1 and the punctuation point at time k, is the angle between the line connecting the target point at time k and the target point at time k + 1 and the line connecting the punctuation point at time k - 1 and the punctuation point at time k.

[0018] Furthermore, in step S300, , where α is the weight coefficient.

[0019] Compared with the prior art, the present invention has the following advantages: In the case where the turning rate of the target is unknown, the present invention estimates the turning rate in real time by using the track and measurement information, adjusts the state transition matrix of the turning model, and establishes a model that matches the actual motion state of the target. The present invention can improve the track filtering accuracy of the interactive multi-model algorithm, is applied to complex and changeable maneuvering target scenarios, expands the application scenarios of the algorithm, and enhances the maneuvering target tracking ability of the algorithm.

[0020] The present invention will be further described below with reference to the accompanying drawings of the specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a schematic flow chart of the method of the present invention.

[0022] Figure 2 Schematic diagram of the model for obtaining the turning rate estimation value.

[0023] Figure 3 Comparison chart of the simulation results between the algorithm of the present invention and the AT-IMM algorithm.

[0024] Figure 4 Comparison chart of the RMSE errors between the algorithm of the present invention and the AT-IMM algorithm in each direction. Detailed implementation manners

[0025] Combined with Figure 1 , a method for obtaining the turning rate of target tracking includes the following steps:

[0026] Step S100: Obtain the state vector X(k) of the target at time k, and establish a state equation including the turning rate;

[0027] Step S200: Estimate the turning rate at time k to obtain the estimated value of the turning rate ;

[0028] Step S300: Combine the turning rate at time k-1 and the estimated value , and perform weighted summation to obtain the turning rate at the current time.

[0029] In step S100, for the state vector at time k, establish the corresponding state equation

[0030]

[0031] wherein, and respectively represent the distances of the target on the x-axis and y-axis, and respectively represent the speeds of the target on the x-axis and y-axis, represents the turning rate of the target, and represent the acceleration disturbance terms on the x-axis and y-axis, and T represents the sampling interval.

[0032] In step S200, as Figure 2 shown, establish a turning rate calculation model, and the establishment method is as follows:

[0033] Step S201: Obtain the filtered estimated values X(k-1) and X(k) at times k-1 and k;

[0034] Step S202: Obtain the measurement value z(k + 1) at the (k + 1)-th moment, where the measurement value also includes the positions and velocities on the x-axis and y-axis.

[0035] Step S203: Obtain the distances a(k), a(k + 1), and c(k) of the target point at the (k - 1)-th and k-th moments, the k-th and (k + 1)-th moments, and the (k - 1)-th and (k + 1)-th moments.

[0036] Step S204: Obtain the estimation of the turning rate

[0037]

[0038] where is the angle between the line connecting the target points at the k-th and (k + 1)-th moments and the extension line of the line connecting the punctuation points at the (k - 1)-th and k-th moments, is the angle between the line connecting the target points at the k-th and (k + 1)-th moments and the line connecting the punctuation points at the (k - 1)-th and k-th moments.

[0039] In Step S300, , where α is the weight coefficient, and the turning rate estimation value can be adjusted by modifying α and the turning rate at the previous moment to change their proportions.

[0040] Comparative Example

[0041] Through the above method for obtaining the turning rate of target tracking, in this embodiment, an Adaptive Turn Model Interactive Multiple Model (AT-IMM) is constructed. This embodiment mainly compares the traditional Interactive Multiple Model (IMM) algorithm and the Adaptive Turn Model-based Interactive Multiple Model (AT-IMM) algorithm. In the designed simulation scenario, the total movement time of the target is 100 s, and the initial state of the target is The target moves at a constant speed of 5 m / s in the Z direction throughout the process. In the X and Y directions, it moves in a straight line at a constant speed of 5 m / s from 0 to 10 s (both in the X and Y directions are 5 m / s). From 10 to 30 s, it switches to a uniform right-turning motion with a turning rate of -0.26 rad / s. From 30 to 50 s, it makes a uniform right-turning motion with a turning rate of -0.52 rad / s. From 50 to 60 s, it switches to a straight-line motion. From 60 to 80 s, it switches to a uniform left-turning motion with a turning rate of 0.26 rad / s. From 80 to 100 s, it makes a uniform left-turning motion with a turning rate of 0.52 rad / s. The initial turning rate of the model used is 0.05 rad / s. The simulation results are as Figure 3 shown.

[0042] From Figure 3It can be seen that the filtered track of the AT-IMM algorithm (track with pentagrams) is more in line with the true motion track of the target (solid track) as a whole compared to the filtered track of the CIMM algorithm (dashed track). From 10 to 30 s, both the AT-IMM algorithm and the IMM algorithm initially deviated from the true track of the target, but the AT-AIMM algorithm was able to quickly adjust and get closer to the target track. From 30 to 50 s, the turning rate of the target increased, and the turning rate estimated by the IMM algorithm was significantly too large. This is because the IMM algorithm could not adjust the turning rate in time, resulting in a large deviation between the measurement value and the estimated track. At this time, the estimated track was incorrect and could not reflect the motion state of the target. The AT-IMM algorithm can estimate and adjust the turning rate in the turning model in real time, making the measurement value and the estimated track basically keep in step, so as to more accurately represent the maneuvering situation of the target. Similarly, within 60 to 80 s, when the target makes a left-turn maneuver, compared with the IMM algorithm, the AT-IMM algorithm can quickly adjust to the convergence state. After the turning rate of the target changes from 80 to 100 s, it can still stably track the moving target.

[0043] Define the root mean square error obtained after N Monte Carlo simulation experiments as:

[0044]

[0045] Where, represents the true value of the i-th state component of the j-th Monte Carlo simulation at time k, represents the filtered estimated value. Take the target state vector to contain the position and velocity information of the target, which is

[0046]

[0047] Define the position errors of the target in the x, y, and z-axis directions as RMSE x 、RMSE y and RMSE z . The root mean square error of the target's spatial position at time k is

[0048]

[0049] The RMSE errors of the two algorithms in each direction obtained by simulation are as shown in the following table and Figure 4 shown, Figure 4 The dashed line in represents the IMM algorithm, and the dotted line represents the AT-IMM algorithm.

[0050]

[0051] From Figure 4It can be seen from the above table that, compared with the IMM algorithm, the AT-IMM algorithm has a significant improvement in the filtering accuracy in the X direction and the Y direction. Among them, the RMSE error in the X direction is reduced by 65.85%, the RMSE error in the Y direction is reduced by 63.87%, and the RMSE error of the spatial position is reduced by 65.36%. This experimental scenario verifies the tracking performance of the AT-IMM algorithm in the variable turning rate maneuvering scenario.

Claims

1. A method for obtaining a turning rate for target tracking, characterized in that: include: Step S100, obtaining the state vector X(k) of the target at time k, and establishing a state equation including the turning rate; Step S200, estimating the turning rate at time k to obtain an estimated value of the turning rate ; Step S300, combining the turning rate at time k-1 The estimated value of , weighted sum to get the turning rate at the current moment .

2. The method according to claim 1, characterized in that In step S100, for the state vector at time k , establish the corresponding state equation ， in, and Represents the distance of the target on the x-axis and y-axis respectively, and Represent the speed of the target on the x-axis and y-axis respectively, represents the turning rate of the target, and represents the acceleration disturbance term on the x-axis and y-axis, and T represents the sampling interval.

3. The method according to claim 2, characterized in that Step S200 includes: Step S201, obtaining the filter estimation values ​​X(k-1) and X(k) at time k-1 and time k; Step S202, obtaining the measurement value z(k+1) at time k+1, wherein the measurement value also includes the position and speed of the x-axis and y-axis; Step S203, obtaining the distances a(k), a(k+1) and c(k) of the target point at time k-1 and k, time k and k+1, and time k-1 and k+1; Step S204, obtaining an estimate of the turning rate , ， in, is the angle between the line connecting the target point at time k and the target point at time k+1 and the extended line connecting the mark point at time k-1 and the mark point at time k, It is the angle between the line connecting the target point at time k and the target point at time k+1 and the line connecting the mark point at time k-1 and the mark point at time k.

4. The method according to claim 3, characterized in that In step S300, , where α is the weight coefficient.

Images