CN116520310A - Maneuvering multi-target tracking method under Doppler blind area - Google Patents

Maneuvering multi-target tracking method under Doppler blind area Download PDF

Info

Publication number
CN116520310A
CN116520310A CN202310493468.8A CN202310493468A CN116520310A CN 116520310 A CN116520310 A CN 116520310A CN 202310493468 A CN202310493468 A CN 202310493468A CN 116520310 A CN116520310 A CN 116520310A
Authority
CN
China
Prior art keywords
target
doppler
measurement
model
representing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310493468.8A
Other languages
Chinese (zh)
Inventor
国强
卢宇翀
王亚妮
戚连刚
卢芳葳
黄帅
卡柳日内.尼古拉
任海宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Engineering University
Original Assignee
Harbin Engineering University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Engineering University filed Critical Harbin Engineering University
Priority to CN202310493468.8A priority Critical patent/CN116520310A/en
Publication of CN116520310A publication Critical patent/CN116520310A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/66—Radar-tracking systems; Analogous systems
    • G01S13/72—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar
    • G01S13/723—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar by using numerical data
    • G01S13/726—Multiple target tracking
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
    • G01S13/50—Systems of measurement based on relative movement of target
    • G01S13/58—Velocity or trajectory determination systems; Sense-of-movement determination systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Software Systems (AREA)
  • Algebra (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Probability & Statistics with Applications (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention belongs to the field of maneuvering multi-target tracking, and particularly relates to a multi-target tracking method under a Doppler blind zone, which comprises the following steps: constructing a Doppler radar detection probability model based on the target motion state, the clutter notch function and the sensor state; based on the Doppler radar detection probability model, a measurement model for estimating and tracking the state of the target is constructed, and a prediction process is carried out to obtain the prior density of the multiple targets; in the algorithm updating process, an additional pseudo measurement component and an enhanced measurement component are added, so that under the condition that the target loses measurement in the Doppler blind area, a higher weight is still kept to prevent the track from being cut, and the track is redistributed after the target leaves the blind area, so that the normal operation of the tracker is ensured; based on the Doppler radar measurement set, carrying out an updating process, and calculating the multi-target posterior density of the IMM-GLMB algorithm; and extracting a target motion state according to the IMM-GLMB model to obtain a current target motion state estimation result.

Description

Maneuvering multi-target tracking method under Doppler blind area
Technical Field
The invention belongs to the field of maneuvering multi-target tracking, and particularly relates to a maneuvering multi-target tracking method under a Doppler blind zone based on an interactive multimode generalized tag Bernoulli (Interacting multiple model-Generalized labeled multi-Bernoulli, IMM-GLMB) filter.
Background
The biggest characteristic of the Doppler radar is that Doppler measurement can be additionally obtained while position measurement is obtained, and target tracking performance can be effectively improved by introducing Doppler measurement into a tracking algorithm, however, due to the fact that Doppler dead zones (Doppler Blind Zone, DBZ) inevitably exist in physical limitation of a Doppler Lei Fa sensor, when the Doppler measurement of a target falls within a minimum detectable speed interval of the Doppler Lei Fa sensor, the Doppler radar cannot detect a moving target, and problems such as track interruption, temporary extinction, restarting, batch interruption and the like are caused. The presence of doppler dead zones can greatly increase the complexity of the doppler radar multi-target tracking problem. In the radar tracking process, the motion state of a target and the matching degree of a motion model in a tracker can directly influence the performance of the tracker, when the motion state and the matching degree of the motion model in the tracker are matched, the tracking performance can be improved, otherwise, the tracking performance can be deteriorated, even filtering divergence is caused, the target tracked in an actual tracking scene usually cannot keep a single operation state, such as the motion changes of an aircraft such as turning, rolling and diving in the flight process, and the single motion model cannot guarantee the tracking precision.
Therefore, how to reduce the influence of the Doppler blind area on the tracker and track the maneuvering target under the blind area coverage is the research focus of the students in the related field.
Disclosure of Invention
In order to overcome the problems in the prior art, the invention provides a multi-target tracking method under a Doppler blind zone.
The technical scheme for solving the technical problems is as follows:
the invention provides a multi-target tracking method under a Doppler blind zone, which comprises the following steps:
constructing a Doppler radar detection probability model based on the target motion state, the clutter notch function and the sensor state;
based on the Doppler radar detection probability model, a measurement model for estimating and tracking the state of the target is constructed, and a prediction process is carried out to obtain the prior density of the multiple targets;
adding additional pseudo measurement components and enhanced measurement components in the measurement model updating process; calculating the multi-target posterior density;
and extracting the target motion state according to the measurement model to obtain a current target motion state estimation result.
Further, the constructing a radar detection probability model based on the minimum detectable speed, the clutter notch function and the detection probability of the Doppler Lei Fa sensor specifically includes:
wherein p is D,k (x) Probability of detecting the target for the doppler Lei Fa sensor; p is p D Representing the detection probability when the Doppler velocity of the target is far from the Doppler blind zone;representing the normalization factor; r is R f ＝MDV 2 (2 ln 2) represents the variance of the pseudo-metric in the Doppler domain; />Representing a pseudo-metrology function; />Is a pseudo-measurement matrix; x represents the motion state of the target;
the clutter notch n c The function is expressed as the difference in target doppler velocity relative to nearby clutter doppler velocity;
when the target enters the Doppler blind zone, i.e. n c <MDV, wherein MDV is the minimum detectable speed, and the Doppler Lei Fa sensor has difficulty in obtaining the measured value of the target, and p is the time D,k (x) Approaching 0; conversely, when the target is far from the Doppler blind zone, n c > MDV, p at this time D,k (x) Will tend to be p D 。
Further, the multi-target prior density is:
wherein X represents a set of target states, and delta (X) represents a tag-different indicator, defined asI.e. when the labels of the elements in set X differ, Δ (X) =1; conversely, Δ (X) =0; delta (·) represents the dirac delta function, defined as: /> Label for showing survivalWeight of->Representing the new born tag->Weights of (2); />And p +，γ( L) represents the viable target density and the nascent target density, respectively, for a given set of tags L; />For normalizing parameters->τ (μ) represents the model probability when the model is μ; c represents a component->Representing a set of components; the prediction set weight is:
the single target mixing density is:
in the method, in the process of the invention,a survival tag->Weight of->Representing the new born tag->Weights of (2);and p +，γ (. L) represents viable target density and neo-target density, respectively.
Further, the multi-target posterior density is:
where τ (μ) represents the model probability when the model is μ;
wherein, in model mu r The following single-target posterior probability densities are:
the average value, covariance and weight are measured for the conventional loss; />The mean value, covariance and weight of the pseudo measurement are obtained; />The mean value, covariance and weight of the position measurement are measured; />To enhance the measurement mean, covariance, and weight;
the update weight is:
the single target normalization constant is:
in the above-mentioned method, the step of,conventional loss measurement weights; />Representing a pseudo-measurement weight; />Representing position measurement weights; />Representing an enhanced measurement weight; θ (l) represents the track association map with label l, and when θ (l) =0, it indicates that the measurement is not associated with the track, and the track is missed, θ (l)>0, the track and the measurement information are updated; delta 0 (X) represents a generalized Croner-delta function.
Compared with the prior art, the invention has the following technical effects:
according to the invention, the Doppler Lei Fa sensor is modeled as a function related to the state of a moving target and the azimuth of the current Doppler Lei Fa sensor according to the change rule of the detection probability of the Doppler Lei Fa sensor, the modeled detection probability model is substituted into the MM-GLMB model, the pseudo measurement component and the enhanced measurement component are added on the basis of the original conventional loss measurement component and the position measurement component, and when the observed target enters a Doppler blind area, the corresponding weights of the pseudo measurement component and the enhanced measurement component are increased despite the loss of measurement information, so that the track can be prevented from being cut off, track information can be immediately allocated after the target exits the blind area, and the influence of the blind area on a tracker is reduced. Meanwhile, the motion state of the target is estimated by adopting an interactive multi-model algorithm, so that the method can adapt to maneuvering change of the target and improve tracking precision. The method has strong robustness and adaptability, and can realize effective tracking of maneuvering multiple targets in a complex radar environment.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a diagram of IMM-GLMB filtering according to the present invention;
FIG. 2 is a diagram of IMM-GLMB-MDV filtering according to the present invention;
FIG. 3 is a graph of OSPA (Optimal sub-pattem assignment) distance comparison for the present invention;
FIG. 4 is a flow chart of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. The particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
Aiming at the problem that the existing GLMB (Generalized Labeled Multi-Bernoulli, generalized label multi-Bernoulli) algorithm is hidden in a Doppler blind area facing in the process of tracking a maneuvering target, the tracking performance suddenly drops and even tracking cannot be continued, the invention provides a maneuvering multi-target tracking method suitable for the GLMB under the Doppler blind area.
The invention uses MDV (Minimum Detectable Velocity, minimum detectable speed) information to model the detection probability, so that the target can keep the corresponding track to have higher weight even if no measurement information exists after entering the blind area, thereby avoiding cutting, and when the target exits the blind area, the track information can be immediately redistributed, thereby reducing the influence of Doppler blind area on the tracker.
Referring to fig. 4, a multi-target tracking method under a doppler blind area includes the steps of:
and step 1, constructing a Doppler radar detection probability model based on the target motion state, the clutter notch function and the Doppler Lei Fa sensor state.
Set the motion state of the target at the moment k asWherein (x) k ,y k ) The target location component is represented as such,representing a target velocity component; correspondingly, the Doppler Lei Fa sensor state is +.>Wherein, the liquid crystal display device comprises a liquid crystal display device,representing the Doppler Lei Fa sensor position component, < ->Representing the doppler Lei Fa sensor velocity component.
Probability p of target detection by Doppler Lei Fa sensor D,k (x) Clutter notch function n c Direct influence, n c Represented as the difference in the target doppler velocity relative to the nearby clutter doppler velocity of the sensor.
Wherein, the liquid crystal display device comprises a liquid crystal display device,doppler measurements representing measurement targets:
representing background clutter Doppler measurements:
in the above, x k The matrix of states of the object is represented,respectively x-axis coordinates, x-direction speed, y-axis coordinates, y-direction speed, x s Is Doppler radar state matrix, < >>
The primary function of the clutter notch is to suppress clutter, but at the same time, it also affects the probability of detection of the target. The probability of detection is modeled here as a function p of the state of the moving object and the orientation of the current Doppler Lei Fa sensor D ＝p D,k (x)。
When the target enters the Doppler blind zone, i.e. n c <MDV (minimum detectable speed), doppler Lei Fa sensor has difficulty obtaining a measurement of the target, at which point p D,k (x) Approaching 0; conversely, when the target is far from the Doppler blind zone, n c > MDV, at this time, p D,k (x) Will tend to be p D ，p D Indicating the probability of detection when the doppler velocity of the target is far from the doppler dead zone, the doppler dead zone can be described as DBZ = { x|n c (x k )<MDV, probability p of target detection by Doppler Lei Fa sensor D,k (x) The method comprises the following steps:
p D,k (x)≈p D [1-exp(-(n c (x k )/MDV) 2 log2)]
because the detection probability in the above method is in an exponential form, the detection probability needs to be converted into a Gaussian formSubsequent calls are made for the nonlinear clutter notch function n c At the predicted valueThe first-order taylor expansion is performed nearby, and the following steps are obtained:
wherein the pseudo-metrology functionThe method comprises the following steps:
the pseudo-measurement matrix is:
in the above formula:
x, y axis coordinates of the predicted values, respectively, ">The x, y-axis direction speeds of the predicted values, respectively.
To approximate the above n c (x) Carrying in to obtain:
in the above-mentioned method, the step of,p D representing the probability of detection when the doppler velocity of the target is far away from DBZ; r is R f ＝MDV 2 (2 ln 2) represents the variance of the pseudo-metric in the Doppler domain;representing the normalization factor. />Gaussian density, H, representing mean m, variance P k For the observation matrix, R is the position measurement noise covariance.
And 2, constructing a measurement model for estimating and tracking the target state based on the Doppler radar detection probability model, and performing a prediction process to obtain the multi-target prior density.
Let the multi-target state be x= { X at time k-1 1 ,x 2 ,…x |X| X= [ x ] k-1 ,μ k-1 ,l k-1 ]E X; x, mu and l respectively represent the motion state, the motion model and the label state corresponding to the target.
Defining a Markov model jump transition probability matrix as follows:
wherein χ (μ) i |μ j ) Representing the motion model as mu j Jump to mu i Is a probability of (2).
The object moves to the moment k, and the state jumps from the x state to the x + The markov state transition density of states is expressed as:
φ(x + ,μ + ,l + |x,μ,l)＝φ(x + |x,μ + ,l)χ(μ + |μ)
the IMM algorithm (interactive multi-model algorithm) needs to use model probability to mix the target motion state before prediction, and the multi-target prior density of the IMM-GLMB after model state mixing:
wherein, the prediction set weight is:
the single target mixing density is:
in the above-mentioned method, the step of,a survival tag->Weight of->Representing the new born tag->Is a weight of (2). />And p +，γ (. L) represents viable target density and neo-target density, respectively, for a given tag set L. X is x + Represents the target prediction state, mu + And l represent the corresponding model and label, respectively.
The single target densities of surviving and nascent targets were:
wherein, the liquid crystal display device comprises a liquid crystal display device,for the surviving component weight, +.>And->Mean and covariance of survival components, respectively, +.>For new target weight, ++>And->The mean and covariance of the nascent object, respectively.
For the normalization parameters:
in the above equation, τ (μ) represents the model probability when the markov model is μ.
The parameters required for the density transfer process are given by:
the weights of the survival target predicted densities are:
in the method, in the process of the invention,represents the scaling factor, p S Representing the probability of survival of the target at the next time.
The survival target needs to interactively input the input state before the prediction step:
mean value after use state mixingSum covariance P i (c,o) (mu, l) temporal prediction was performed.
Wherein, the liquid crystal display device comprises a liquid crystal display device,and->Mean and covariance of survival components, F (μ) + ) Represents a state transition matrix, Q (μ) + ) Representing the noise covariance.
The weight of the new target prediction density is as follows:
in the method, in the process of the invention,indicating the probability of existence of the new target.
Since the new target has no prior model probability, a model mixing step is not needed, and the mean value and covariance of the new target can be directly calculated.
Step 3, adding additional pseudo measurement components and enhanced measurement components in the updating process of the measurement model, keeping high weight to prevent the track from being cut under the condition that the target loses measurement in the Doppler blind zone, and reallocating the track after the target leaves the blind zone to ensure the normal operation of the tracker; and calculating the multi-objective posterior density.
Let the measurement set of the Doppler Lei Fa sensor at the k moment be Z= { Z 1 ,z 2 ,…,z Z }，z＝[y c ,y d ]，y c Representing position measurement information, y d Indicating doppler measurement information.
The multi-target posterior density at this time is:
wherein, in model mu r The following single-target posterior probability densities are:
the update weight is:
the single target normalization constant is:
in the above description, θ (l) represents a track association map with label l, and when θ (l) =0, it indicates that measurement is not associated with track, and at this time, track is missed, θ (l)>And 0, updating the track and the measurement information. Delta 0 (X) represents a generalized Croner-delta function. The definition is as follows:
conventional loss measurementThe mean, covariance, and weights in (a) are given by:
wherein the method comprises the steps ofP i (c) And->Mean, covariance, and weights obtained for the prediction process.
Pseudo-measurementThe mean, covariance, and weights in (1) are obtained from the following equation:
wherein the pseudo-measurement gain K i,f Given by the formula:
the pseudo-measurement innovation covariance is:
the weights of the pseudo measurement components are:
in the above, c f Representing the normalization factor, given in modeling the detection probability in step one
Likelihood probability of a componentThe method comprises the following steps:
position measurementThe mean, covariance and weights of the data are obtained by processing position measurements and Doppler measurements, and the position measurement information y is firstly utilized c Updating parameters:
wherein H is c The target observation matrix is represented, and the gain and the innovation covariance of the position measurement are respectively:
then use Doppler information y d And (5) sequentially updating:
the weights of the location components are:
wherein, the liquid crystal display device comprises a liquid crystal display device,the clutter intensities, denoted as position component and doppler component, respectively.
The Doppler measurement gain and covariance are:
wherein R is c Sum sigma d The noise standard deviation of the position measurement and the Doppler measurement are respectively determined.
Enhanced metrologyThe mean, covariance, and weights in (1) are obtained from the following equation:
the gain and covariance of the enhancement measurement are:
enhanced metrology matrixBy putting->Is obtained by the following formula.
Wherein:
weights of the componentsGiven by the formula:
and 4, extracting the target motion state according to the measurement model to obtain a current target motion state estimation result.
Unlike multi-model, IMM algorithms require model state blending after all component calculations are completed to perform target state extraction. The following demolding mixing process is as follows:
the model probability updating process comprises the following steps:
wherein c is a normalization parameter
It should be noted that, in this embodiment, the symbols of the components are expressed in a consistent manner, and the subscript is only used to distinguish the different components.
The experimental conditions in the invention are as follows: the observation scene is a [ -1000 1000](m)×[-1000 1000](m) two-dimensional plane, the scanning period of the Doppler Lei Fa sensor is 1s, the observation time is 100s, and the motion model set of the target consists of a uniform linear motion, a right turning model (a cooperative turning model with a turning rate of 2 DEG/s) and a left turning (-a cooperative turning model with a turning rate of 2 DEG/s). In this experimental scenario, there are3 observations appear at different positions at different times, the new positions are (-400, 380), (-400, 410), (-700, 400), the motion state of the target is shown in table 1, and the state transition equation is: x is x + =f (μ) x+v, where the state transition matrices corresponding to the three motion models are obtained by bringing ω=0 (when μ=1), ω=2pi/180 (when μ=2), ω= -2pi/1 (when μ=3) into F (μ), respectively 1 )，F(μ 2 ) And F (mu) 3 )。
TABLE 1 motion state of objects
Where T represents the scan period of the doppler Lei Fa sensor, t=1s, and v represents process noise with covariance Q mean of 0. Sigma (sigma) v The survival probability per target is p =10m/s s =0.99, the saturation value of the doppler Lei Fa sensor detection probability is p D =0.98, a priori velocity standard deviation σ of the target s 17m/s, standard deviation sigma of position measurement noise c Standard deviation sigma of Doppler observed noise =10m/s d The clutter points per cycle obey a poisson distribution with an average value of 25, and the positions of each clutter point are uniformly distributed in the measurement range. Model probability of nascent target is τ 0 ＝[1/3 1/3 1/3]For comparison convenience, pruning parameters t=10 for all filters -5 Combining threshold u=4 and maximum number of gaussian components J max =100, the multi-objective extraction threshold is 0.5.
The initial motion state of the target is:
x 1 ＝[-400 7 380 -7]
x 2 ＝[-400 -6 410 -10]
x 3 ＝[-700 3 400 -14]
the Markov matrix for switching between the target motion models is set as follows:
fig. 1 and fig. 2 are respectively an IMM-GLMB filtering result and an IMM-GLMB-MDV filtering result, it can be seen that measurement is lost immediately after a target enters a doppler blind zone, and normal tracking is not possible even after the target exits the blind zone, and a filter incorporating MDV information loses a part of measurement in the blind zone but can be recovered immediately after the target exits the blind zone, so that the influence of the doppler blind zone on the performance of the tracker is effectively reduced.
In summary, the direct influence of the Doppler blind area on the radar can cause continuous loss of measurement on the moving target, when the target enters the Doppler blind area, once the continuously lost measurement point meets the track withdrawal threshold, the target is missed, and after the target leaves the Doppler blind area, the new measurement information has a larger error with the distance of the inherent threshold, so that the target is misdetected. According to the invention, the detection probability of the Doppler Lei Fa sensor is modeled aiming at the formation cause of the Doppler blind area, and the IMM algorithm is applied to the GLMB tracking algorithm according to the characteristic of high mobility of the observation target, so that the maneuvering multi-target tracking under the coverage of the Doppler blind area is realized.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.

Claims (4)

1. The multi-target tracking method under the Doppler blind zone is characterized by comprising the following steps of:
constructing a Doppler radar detection probability model based on the target motion state, the clutter notch function and the sensor state;
based on the Doppler radar detection probability model, a measurement model for estimating and tracking the state of the target is constructed, and a prediction process is carried out to obtain the prior density of the multiple targets;
adding additional pseudo measurement components and enhanced measurement components in the measurement model updating process; calculating the multi-target posterior density;
and extracting the target motion state according to the measurement model to obtain a current target motion state estimation result.
2. The method for multi-target tracking under a doppler blind zone according to claim 1, wherein the constructing a radar detection probability model based on a minimum detectable speed, a clutter notch function and a detection probability of a doppler Lei Fa sensor specifically comprises:
wherein p is D,k (x) Probability of detecting the target for the doppler Lei Fa sensor; p is p D Representing the detection probability when the Doppler velocity of the target is far from the Doppler blind zone;representing the normalization factor; r is R f ＝MDV 2 (2 ln 2) represents the variance of the measurement in the Doppler domain; />Representing a pseudo-metrology function; />Is a pseudo-measurement matrix; x represents the motion state of the target;
clutter notch function n c Expressed as the difference in target doppler velocity relative to nearby clutter doppler velocity;
when the target enters the Doppler blind zone, i.e. n c <MDV, wherein MDV is the minimum detectable speed, and the Doppler Lei Fa sensor has difficulty in obtaining the measured value of the target, and p is the time D,k (x) Approaching 0; conversely, when the target is far from the Doppler blind zone, n c > MDV, p at this time D,k (x) Will tend to be p D 。
3. The method for multi-target tracking under doppler blind areas according to claim 2, wherein the multi-target prior density is:
wherein Δ (X) represents a label-dissimilarity indicator defined asI.e. when the labels of the elements in set X differ, Δ (X) =1; conversely, Δ (X) =0; delta (·) represents the dirac delta function, p (c) Representing the state probability density, ω, of a single target (c) Weights representing the corresponding components ∈ ->Is a normalization parameter;
the label is L (X + ) Weight of +.>Representing single target density for a given tag setClosing L; />For normalizing parameters->τ (μ) represents the model probability when the model is μ; c represents a component->Representing a set of components; the prediction set weight is:
the single target mixing density is:
in the method, in the process of the invention,a survival tag->Weight of->Representing the new born tag->Weights of (2);and p +，γ (. L) represents viable target density and neo-target density, respectively.
4. A method for multi-target tracking under doppler blind zone according to claim 3, wherein the multi-target posterior density is:
where τ (μ) represents the model probability when the model is μ;
wherein, in model mu r The following single-target posterior probability densities are:
the average value, covariance and weight are measured for the conventional loss; />The mean value, covariance and weight of the pseudo measurement are obtained; />The mean value, covariance and weight of the position measurement are measured; />To enhance the measurement mean, covariance, and weight;
the update weight is:
the single target normalization constant is:
in the above-mentioned method, the step of,conventional loss measurement weights; />Representing a pseudo-measurement weight; />Representing position measurement weights; />Representing an enhanced measurement weight; θ (l) represents the track association map with label l, and when θ (l) =0, it indicates that the measurement is not associated with the track, and the track is missed, θ (l)>0, the track and the measurement information are updated; delta 0 (X) represents a generalized Croner-delta function.
CN202310493468.8A 2023-05-04 2023-05-04 Maneuvering multi-target tracking method under Doppler blind area Pending CN116520310A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310493468.8A CN116520310A (en) 2023-05-04 2023-05-04 Maneuvering multi-target tracking method under Doppler blind area

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310493468.8A CN116520310A (en) 2023-05-04 2023-05-04 Maneuvering multi-target tracking method under Doppler blind area

Publications (1)

Publication Number Publication Date
CN116520310A true CN116520310A (en) 2023-08-01

Family

ID=87391754

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310493468.8A Pending CN116520310A (en) 2023-05-04 2023-05-04 Maneuvering multi-target tracking method under Doppler blind area

Country Status (1)

Country Link
CN (1) CN116520310A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117784115A (en) * 2023-12-26 2024-03-29 兰州理工大学 Gaussian process regression model multi-expansion target PMBM tracking method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117784115A (en) * 2023-12-26 2024-03-29 兰州理工大学 Gaussian process regression model multi-expansion target PMBM tracking method

Similar Documents

Publication Publication Date Title
CN114859339A (en) Multi-target tracking method based on millimeter wave radar
Attari et al. An SVSF-based generalized robust strategy for target tracking in clutter
Punchihewa et al. A generalized labeled multi-Bernoulli filter for maneuvering targets
Smith et al. Systematic analysis of the PMBM, PHD, JPDA and GNN multi-target tracking filters
CN116520310A (en) Maneuvering multi-target tracking method under Doppler blind area
Zhang et al. Robust multi-target tracking under mismatches in both dynamic and measurement models
Feigl et al. Recurrent neural networks on drifting time-of-flight measurements
CN111257865A (en) Maneuvering target multi-frame detection tracking method based on linear pseudo-measurement model
Wu et al. MM-GLMB filter-based sensor control for tracking multiple maneuvering targets hidden in the Doppler blind zone
Scheel et al. Using separable likelihoods for laser-based vehicle tracking with a labeled multi-Bernoulli filter
CN111679251A (en) Radar-type interference resisting method based on radar infrared dual-mode fusion
CN106291530B (en) A kind of probabilistic data association optimization method based on nearest neighbor method
Fang et al. A tracking approach for low observable target using plot-sequences of multi-frame detection
CN111274529B (en) Robust Gao Sini Weisal PHD multi-expansion target tracking algorithm
Jing et al. Process noise identification based particle filter: an efficient method to track highly manoeuvring targets
Liu et al. Multiple hypothesis method for tracking move‐stop‐move target
CN109188422A (en) A kind of Kalman filtering method for tracking target decomposed based on LU
CN113280821B (en) Underwater multi-target tracking method based on slope constraint and backtracking search
Qiu et al. Multiple targets tracking by using probability data association and cubature Kalman filter
CN111679270B (en) Multipath fusion target detection algorithm under scene with uncertain reflection points
Jonsson et al. Multi-target tracking with background discrimination using PHD filters
Zhang et al. Model‐switched Gaussian sum cubature Kalman filter for attitude angle‐aided three‐dimensional target tracking
Hongxin et al. Modified joint probabilistic data association with classification-aided for multitarget tracking
Ikram et al. A new data association method for 3-D object tracking in automotive applications
Hajiramezanali et al. Stochastic differential equations for modeling of high maneuvering target tracking

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination