CN117031473B - Underwater target collaborative track tracking method - Google Patents

Underwater target collaborative track tracking method Download PDF

Info

Publication number
CN117031473B
CN117031473B CN202311284679.7A CN202311284679A CN117031473B CN 117031473 B CN117031473 B CN 117031473B CN 202311284679 A CN202311284679 A CN 202311284679A CN 117031473 B CN117031473 B CN 117031473B
Authority
CN
China
Prior art keywords
uuv
target
underwater
time
state vector
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.)
Active
Application number
CN202311284679.7A
Other languages
Chinese (zh)
Other versions
CN117031473A (en
Inventor
邢文
刘建波
张勋
冯志光
姚思博
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.)
Qingdao Harbin Engineering University Innovation Development Center
Original Assignee
Qingdao Harbin Engineering University Innovation Development Center
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 Qingdao Harbin Engineering University Innovation Development Center filed Critical Qingdao Harbin Engineering University Innovation Development Center
Priority to CN202311284679.7A priority Critical patent/CN117031473B/en
Publication of CN117031473A publication Critical patent/CN117031473A/en
Application granted granted Critical
Publication of CN117031473B publication Critical patent/CN117031473B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
    • G01S15/66—Sonar tracking systems
    • 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
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/52—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00
    • G01S7/539—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S15/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • 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
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

An underwater target cooperative track tracking method belongs to the field of underwater target tracking. The invention aims to solve the problems that a Kalman filtering method used for tracking the track of the existing underwater target can generate a large amount of estimation errors and even lead to filtering divergence for complex and changeable underwater environments; and the target maneuver also can interfere with the filtering results. The process is as follows: A. in all UUV participating in target tracking, no. 1 UUV is used as a master UUV, other UVs are used as slave UVs, and only one slave UUV can send information to the master UUV at the same time; B. taking the self-measurement information of the main UUV and the self-measurement information of the slave UUV received by the main UUV under the same sampling time as a vector; the vector is an actual quantity measurement; and fusing the actual measurement of each sampling time by using a fusion algorithm until target tracking is completed. The invention is used for underwater target tracking.

Description

Underwater target collaborative track tracking method
Technical Field
The invention relates to the field of pure-azimuth target tracking, and belongs to a robust pure-azimuth underwater target collaborative track tracking method considering communication time lag.
Background
The pure-azimuth underwater target cooperative track tracking is a technology for precisely determining the position and the running track of a moving target by adopting a proper filtering estimation algorithm and combining a moving target kinematic model and passive sonar of an underwater vehicle to measure the relative azimuth information of a non-cooperative target. The method belongs to passive positioning, the tracking system does not radiate electromagnetic waves or acoustic waves, is not easy to detect by enemies, and has stronger battlefield viability.
Due to the limitations of the underwater acoustic channel, only one underwater acoustic communication machine can be in a transmitting state at the same time, so that the underwater acoustic communication must follow strict time sequence. This results in that the observed information of all nodes can reach the fusion center after a certain time lag, and the data fusion algorithm needs to be designed under the communication time lag.
The Kalman filtering method used for tracking the underwater target track has the highest state estimation precision under the condition of a linear state space model with Gaussian distribution noise. For complex and variable underwater environments, passive sonar may often be disturbed by outliers, resulting in non-gaussian heavy tail distribution of measurement noise. Meanwhile, the statistical characteristics of the observed noise of the target are often unknown or even time-varying, and the relative positions of the target and the sensor are related to the hydrologic environment. The performance of a kalman filter depends to a large extent on the a priori knowledge of the noise covariance matrix, and the use of a wrong a priori noise covariance matrix and measurement noise non-gaussian can produce a large amount of estimation errors and even result in filter divergence. In addition, the target maneuver may also interfere with the filtering result. It is desirable to design a robust method to address the problems of inaccuracy in system noise and measurement noise, and errors due to non-gaussian conditions.
Disclosure of Invention
The invention aims to solve the problems that a Kalman filtering method used for tracking the track of the existing underwater target can generate a large amount of estimation errors and even lead to filtering divergence for complex and changeable underwater environments; and the problem that the maneuvering of the target can also interfere the filtering result is solved, and an underwater target collaborative track tracking method is provided.
The underwater target collaborative track tracking method specifically comprises the following steps:
step A, observing underwater non-cooperative targets by all UUV participating in target tracking, wherein each UUV has different sampling frequencies;
the UUV 1 in all the UUV participating in target tracking is used as a master UUV, and other UVs are used as slave UVs;
the main UUV is used as a data fusion center;
each slave UUV transmits time slotsCrossing with each otherInformation is sent to the main UUV instead of the own time slot, and only one slave UUV can send information to the main UUV at the same time, so that mutual interference is avoided;
step B, slaveStarting at moment, measuring azimuth angle information of an underwater non-cooperative target relative to a main UUV by using a passive sonar on the main UUV as self-measurement information of the main UUV;
meanwhile, the azimuth information of the underwater non-cooperative target relative to the slave UUV and the corresponding sampling time are measured by the passive sonar on the slave UUV and used as the self-measurement information of the slave UUV, and the slave UUV sends the self-measurement information between the last communication end of the slave UUV and the current communication to the master UUV through underwater sound in own time slots;
taking the self-measurement information of the main UUV and the self-measurement information of the slave UUV received by the main UUV under the same sampling time as a vector;
the vector is the actual measurement；
And fusing the actual measurement of each sampling time by using a fusion algorithm until target tracking is completed.
The beneficial effects of the invention are as follows:
the invention aims to design a pure-azimuth underwater target co-location and tracking algorithm, which utilizes a plurality of Underwater Unmanned Vehicles (UUV) passive sonar to obtain azimuth information of targets, utilizes underwater acoustic communication machines to communicate with each other, and solves the problem of target tracking under the condition of unknown non-Gaussian measurement noise caused by complex underwater environment under the constraint condition of communication time lag.
The target track tracking scheme provided by the invention ensures that all measurement information can be effectively utilized under the condition of communication time lag, and ensures the real-time performance of state estimation. The data fusion algorithm can adapt to the statistical characteristics of process noise and observation noise, and has good robustness when the noise is unknown and is not Gaussian.
Drawings
Figure 1 is a flow chart of the data fusion algorithm of the present invention,is->Target posterior estimated state vector for each sample time,/->Is->A priori estimating a state vector of the object at the sampling time,/->For the target state vector correction value, +>For the actually measured azimuth information of the underwater non-cooperative target relative to the UUV,/the underwater non-cooperative target is a target of the UUV>For observing vector innovation, ->Is Kalman filtering gain, < >>Estimating a vector for the corrected state;
figure 2 is a communication and sampling timing diagram,is the firstn1 st sample data in self-measurement information from UUV last communication end to current communication in 1 st communication time slot,/>Is the firstnThe 1 st sampling data in the 1 st communication time slot from the end of UUV last communication to the 2 nd sampling data in the self-measurement information between the current communication>Is the firstnThe 1 st sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is the firstnThe 2 nd sampling data in the self-measuring information from the end of UUV last communication to the present communication in the communication time slot +.>Is the firstn2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the present communication in the 2 nd communication time slot,/>Is the firstnSample data of 3 rd in self-measurement information from UUV last communication end to current communication in 2 nd communication time slot,/>Is the firstnThe 2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is the firstnSample data 1 in the self-measurement information from the end of UUV last communication to the present communication in 3 rd communication time slot,/day>Is the firstnSample data of the 3 rd of the communication time slots from the last communication end of UUV to the 2 nd of the self-measurement information between the last communication end and the present communication, and +.>Is the firstnSelf-measurement from UUV last communication end to current communication in 3 rd communication time slotSample data 3 in the information;
is->1 st sample data in self-measurement information from UUV last communication end to current communication in 1 st communication time slot,/>Is->The 1 st sampling data in the 1 st communication time slot from the end of UUV last communication to the 2 nd sampling data in the self-measurement information between the current communication>Is->Sample data of 1 st of the communication time slots from the end of the last communication of UUV to the 3 rd of the self-measurement information of the present communication,/>Is->The 1 st sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is->The 2 nd sampling data in the self-measuring information from the end of UUV last communication to the present communication in the communication time slot +.>Is->2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the present communication in the 2 nd communication time slot,/>Is->The 2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is->Sample data 1 in the self-measurement information from the end of UUV last communication to the present communication in 3 rd communication time slot,/day>Is->Sample data of the 3 rd of the communication time slots from the last communication end of UUV to the 2 nd of the self-measurement information between the last communication end and the present communication, and +.>Is->3 rd sample data in the self-measurement information from the end of the last communication of UUV to the present communication in the communication time slot>Is->The 3 rd communication time slot from the last communication end of UUV to the current communicationThe 4 th sampling data in the self-measurement information;
is->1 st sample data in self-measurement information from UUV last communication end to current communication in 1 st communication time slot,/>Is->The 1 st sampling data in the 1 st communication time slot from the end of UUV last communication to the 2 nd sampling data in the self-measurement information between the current communication>Is->The 1 st sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is->The 2 nd sampling data in the self-measuring information from the end of UUV last communication to the present communication in the communication time slot +.>Is->2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the present communication in the 2 nd communication time slot,/>Is->The 2 nd sampling data in the self-measurement information from the end of the last communication of UUV to the communication of the present time in the communication time slot;
is->Sample data 1 in the self-measurement information from the end of UUV last communication to the present communication in 3 rd communication time slot,/day>Is->Sample data of the 3 rd of the communication time slots from the last communication end of UUV to the 2 nd of the self-measurement information between the last communication end and the present communication, and +.>Is->The 3 rd sampling data in the self-measurement information from the end of the last communication of UUV to the communication in the 3 rd communication time slot;
nis the firstnA number of communication time slots of the communication network,each transmission duration of each slave UUV is a time slot;
FIG. 3 is a diagram of simulation results of an embodiment, X/m is the east-west coordinate in terms of m in terms of Cartesian coordinate system, Y/m is the north-south coordinate in terms of m in terms of Cartesian coordinate system, and the modified CKF algorithm is the proposed algorithm of the present invention;
FIG. 4 is a simulated root mean square error plot of an embodiment, time/s is the simulation time in seconds, error/m is the root mean square error in meters, and the EKF algorithm is a standard extended Kalman filter algorithm.
Detailed Description
The first embodiment is as follows: the underwater target collaborative track tracking method in the embodiment comprises the following specific processes:
step A, observing underwater non-cooperative targets by all UUV participating in target tracking, wherein sampling is not necessarily synchronous, and each UUV has different sampling frequencies;
the UUV 1 in all the UUV participating in target tracking is used as a master UUV, and other UVs are used as slave UVs;
the main UUV is used as a data fusion center;
each slave UUV transmits time slotsInformation is alternately sent to the main UUV in own time slot, and only one slave UUV can send information to the main UUV at the same time, so that mutual interference is avoided;
step B, slaveStarting at moment, measuring azimuth angle information of an underwater non-cooperative target relative to a main UUV by using a passive sonar on the main UUV as self-measurement information of the main UUV;
meanwhile, the azimuth information and the corresponding sampling time of the underwater non-cooperative target relative to the slave UUV are measured by the passive sonar on the slave UUV, the slave UUV is used as self-measurement information of the slave UUV, and the self-measurement information (one slave UUV is communicated with a plurality of sampling data at one time) between the last communication of the slave UUV and the communication is sent to the master UUV through underwater sound in own time slots;
taking the self-measurement information of the main UUV and the self-measurement information of the slave UUV received by the main UUV under the same sampling time as a vector;
the vector is the actual measurement；
And fusing the actual measurement of each sampling time by using a fusion algorithm until target tracking is completed, so that the real-time performance of state estimation is ensured.
The method comprises the steps that a master UUV continuously obtains azimuth information of an actually measured underwater non-cooperative target relative to the master UUV, each slave UUV continuously obtains azimuth information of the actually measured underwater non-cooperative target relative to the slave UUV and corresponding sampling time, and the slave UUV time-sharing sends self measurement information between the last communication end and the current communication end to the master UUV;
one slave UUV communicates multiple samples at a time.
The second embodiment is as follows: the present embodiment differs from the specific embodiment in that the fusion algorithm is as follows:
step one, establishing an operation model of an underwater non-cooperative target;
underwater non-cooperative targets such as enemy submarines;
step two, obtaining a nonlinear underwater non-cooperative target observation model and actual quantity measurement;
the actual quantity measurement is azimuth angle information of the underwater non-cooperative target which is actually measured relative to the UUV;
step three, obtaining a nonlinear underwater non-cooperative target observation model based on the step twoTarget posterior estimated state vector for each sample time +.>And->Posterior error covariance matrix for each sample time +.>；
Step four, obtaining a target state vector correction value by using the cerebellar neural network CMAC based on the step three；
Step five, based on step threeTarget posterior estimated state vector for each sample time +.>And step four target state vector correction value +.>Obtaining corrected state estimation vector to complete the +.>Target tracking at each sampling time;
step six, repeatedly executing the steps three to five (using the actual measurement of the next moment)) Until target tracking is completed.
Other steps and parameters are the same as in the first embodiment.
And a third specific embodiment: the difference between the embodiment and the first or second embodiments is that in the first step, an operation model of an underwater non-cooperative target (such as an enemy submarine) is built; the specific process is as follows:
the discrete time state equation of the running model of the underwater non-cooperative target is as follows:
（1）
wherein,is a non-cooperative target under water>The state vector of the moment of time,is a non-cooperative target under water>State vector of time of day->For inputting vectors, ++>0 mean Gaussian noise sequence,>for state transition matrix>Acceleration is the underwater non-cooperative target;
representing the north-east direction (north is the direction of north) in the Cartesian coordinate systemyThe east is the axisxAxis) underwater non-cooperative target position, +.>Is underwater non-cooperative target speed,/-> Acceleration is the underwater non-cooperative target;
the state transition matrixThe expression is
（2）
Wherein,respectively, the maneuvering frequencies of the underwater non-cooperative targets in the east and north directions are +.>Is the sampling period;
the input vectorThe expression is
（3）
Wherein,indicating transpose,/->、/>、/>、/>、/>、/>Representing intermediate variables.
Other steps and parameters are the same as in the first or second embodiment.
The specific embodiment IV is as follows: the difference between the present embodiment and one to three embodiments is that, in the second step, a nonlinear underwater non-cooperative target (such as an enemy submarine) observation model and an actual measurement are obtained:
（4）
wherein,for UUV position, ++>Numbering UUV, ++>Representing an underwater non-cooperative target location;
the actual amount is measured，/>，/>Is thatNumber of time measurement, +.>Representation->Measuring all quantities at the moment; unknown measurement noise->Mean 0, variance +.>Variance->Determining by UUV sensor precision; />Representing the relative azimuth angle of the non-cooperative target as an observation equation;
the actual quantity measurement is the azimuth angle information of the underwater non-cooperative target which is actually measured relative to the UUV.
Other steps and parameters are the same as in one to three embodiments.
Fifth embodiment: the difference between the present embodiment and the first to fourth embodiments is that the adaptive outlier robust volume kalman filter is applied to target tracking in the third step;
obtaining a second nonlinear underwater non-cooperative target observation model and actual quantity measurement based on the second stepTarget posterior estimated state vector for each sample time +.>And->Posterior error covariance matrix for each sample time +.>The method comprises the steps of carrying out a first treatment on the surface of the The specific process is as follows:
step III, byEstimated state vector of time object +.>First->Error covariance matrix of individual sample times +.>Calculate->A priori estimated state vector of the object at the sampling time +.>First->A priori error covariance matrix of the individual sample times +.>；
（5）
（6）
Wherein,for transposition->A process noise covariance matrix;
step III, step II, step IIIA priori error covariance matrix of the individual sample times +.>Performing cholesky decomposition to obtain triangular matrix +.>；
Based on triangular matrixAnd->A priori estimated state vector of the object at the sampling time +.>Obtaining a volume point of the state vector;
obtaining a predicted observed volume point based on a state vector volume point；
Obtaining expectations of predictive observations based on volumetric points of the predicted observations；
Expectations based on predictive observationsFirst->A priori estimated state vector of the object at the sampling time +.>Volume point of predicted observance ∈ ->Obtaining a cross covariance matrix of the state vector and the prediction observance quantity by the volume point of the state vector;
said cholesky is a square root method;
the specific process is as follows:
（7）
（8）
（9）
（10）
（11）
wherein,for->Performing cholesky decomposition to obtain a triangular matrix;
is a state vector +>Is>A plurality of volume points;
is->A plurality of volume points;
hope +.>Is>A plurality of volume points;
is a serial number;
is a state vector +>Is a dimension of (2);
is an observation equation;
for predicting observance->Is not limited to the desired one;
a cross covariance matrix for the state vector and the predicted observed quantity;
the [1]]Is all 1Vector;
step III, a cross-cooperation party based on state vectors and prediction observablesDifference matrixAnd->A priori error covariance matrix of the individual sample times +.>Constructing a pseudo-linear measurement matrix>；
Due to the observation equationNonlinear, non-direct derivation of the measurement matrix, pseudo-linear measurement matrix +.>Can equivalently use the error covariance matrix +.>And cross covariance matrix->A representation;
（12）
step three, four, in practical application, accurately measuring the noise covariance matrixIt is difficult to obtain, and the on-line real-time estimation is carried out on the measurement output;
a measurement noise covariance matrix in the form of an exponentially weighted average is calculated,
（13）
（14）
wherein,is->Measuring a noise covariance matrix at a moment;
is->Measuring a noise covariance matrix at a moment;
for measuring the prediction error vector;
the azimuth information of the underwater non-cooperative target relative to the UUV is actually measured;
is->A time-of-day cycle parameter;
is->Time cycle parameter, taking initial value ∈ ->；
Is a progressive factor, usually +.>=0.9~0.999；
Step III, calculating a nominal observation vector based on step III, step III and step IIIAnd error covariance matrix->；
Based on the third step, the second step, the third step, the fourth step and the unknown parametersCalculating a correction error covariance matrix->；
Iterative determination of unknown parametersUntil the condition is satisfied, obtaining the minimum fixed point iteration times of ensuring convergence>A value;
the specific process is as follows:
calculating a nominal observation vector based on the third step, the first step, the second step and the third stepSum-of-error covariance matrix；
（15）
（16）
Based on the third step, the second step, the third step, the fourth step and the unknown parametersCalculating a correction error covariance matrix->；
（17）
Is->Parameters of the secondary iteration;
motionless point iteration parameter determinationInitializing->Cycling:
（18）
（19）
up toThen->；
Wherein,correcting an error covariance matrix;
respectively +.>Parameter values for the second iteration;
for iteration->Correcting the error covariance matrix after the secondary;
as an auxiliary function, the parameters +.>In->Increment in the multiple iterations;
is a trace of the matrix;
is a proportional matrix;
is->Number of time measurement;
is unknown parameter->Is a good approximation of the estimate of (a);
for ensuring convergence of the minimum number of iterations +.>A value;
is the minimum iteration number;
is an infinitely small real number;
step III, six, updating Kalman filtering gainUpdate->Target posterior estimated state vector for each sample time +.>Update->Posterior error covariance matrix for each sample time +.>The method comprises the steps of carrying out a first treatment on the surface of the The specific process is as follows:
wherein,for iteration->The error covariance matrix after the secondary;
is Kalman filtering gain;
is the nominal observation vector;
is a unit matrix;
and->Namely +.>Target posterior estimated state vector of sampling time +.>A posterior error covariance matrix of the individual sample times.
Other steps and parameters are the same as in one to four embodiments.
Specific embodiment six: the difference between the present embodiment and one to fifth embodiments is that, in the fourth step, the target state vector correction value is obtained based on the third step by using a cerebellar neural network CMAC, where the CMAC network is used to reduce the estimation error caused by model uncertainty and nonlinearity in the maneuvering process; the specific process is as follows:
cerebellar neural network CMAC input: kalman filter gainObservation vector innovationDifference between predicted value and filtered estimated value +.>；
Neural network output: target state vector correction value；
Wherein,a correction value for the target state vector;
the azimuth information of the underwater non-cooperative target relative to the UUV is actually measured;
is an observation equation;
is->Estimating a state vector by a target posterior of each sampling time;
is->The state vector is estimated a priori for the target at each sample time.
Other steps and parameters are the same as in one of the first to fifth embodiments.
Seventh embodiment: this embodiment differs from the first to sixth embodiments in that the fifth step is based on the third stepTarget posterior estimated state vector for each sample time +.>And step four target state vector correction value +.>Obtaining corrected state estimation vector, complete->Tracking a target at moment;
the corrected state estimation vector is:
（23）。
other steps and parameters are the same as in one of the first to sixth embodiments.
One of the slave UUV is inThe information is sent to the main UUV at the sampling moment, and the information content is thatAll observations in between; the main UUV is +.>Finishing receiving at the moment, and starting a communication time slot of the next slave UUV;
each transmission time for each slave UUV is time slot, < > for each slave UUV>The total number of all slave UUVs.
The communication protocol uses Time Division Multiple Access (TDMA), each slave UUV allocates corresponding time slots, data are alternately transmitted in own time slots, and only one slave UUV can transmit information at the same time, so that mutual interference is avoided;
the communication and sampling timing of the 3 slave UUVs is shown in fig. 2.
The 1 st sampling data in the self-measurement information from the end of the last communication of UUV to the present communication is thatSample 2 is +.>The method comprises the steps of carrying out a first treatment on the surface of the A slave UUV having a plurality of sampled data communicated at a time;
nis the firstnIn a number of communication cycles of the communication,is->And slave UUVs.
The invention is described in further detail below with reference to the drawings and the detailed description.
The invention will be further explained by taking a specific multi-UUV cooperative target trajectory tracking as an example. The number of UUVs involved in target track following is 3. UUV No. 1 is used as a data fusion center, information is sent in turn, all measurement information in a communication period is sent to the fusion center, and each transmission time length is. The passive sonar sampling period is equal to +.>Is not fixed and sampling is not synchronous, filtering initial value error +.>. The measured noise is affected by the outlier and is in heavy tail step by step, which satisfies,
wherein the method comprises the steps ofExpressed as probability 0.90->Mean value 0, variance ++>Normal distribution of->Taking 3 °
Fig. 3 shows simulation results lasting 1000s, and fig. 4 shows error comparison of the improved volume kalman filter algorithm and the extended kalman filter algorithm, wherein the algorithm has higher convergence speed and estimation accuracy.
The present invention is capable of other and further embodiments and its several details are capable of modification and variation in light of the present invention, as will be apparent to those skilled in the art, without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (6)

1. A collaborative trajectory tracking method for an underwater target is characterized by comprising the following steps: the method comprises the following specific processes:
step A, observing underwater non-cooperative targets by all UUV participating in target tracking, wherein each UUV has different sampling frequencies;
the UUV 1 in all the UUV participating in target tracking is used as a master UUV, and other UVs are used as slave UVs;
the main UUV is used as a data fusion center;
each slave UUV transmits time slot T for each time 0 Information is alternately sent to the main UUV in own time slot, and only one slave UUV can send information to the main UUV at the same time, so that mutual interference is avoided;
step B, starting from the time t', measuring azimuth angle information of the underwater non-cooperative target relative to the main UUV by using a passive sonar on the main UUV as self-measurement information of the main UUV;
meanwhile, the azimuth information of the underwater non-cooperative target relative to the slave UUV and the corresponding sampling time are measured by the passive sonar on the slave UUV and used as the self-measurement information of the slave UUV, and the slave UUV sends the self-measurement information between the last communication end of the slave UUV and the current communication to the master UUV through underwater sound in own time slots;
taking the self-measurement information of the main UUV and the self-measurement information of the slave UUV received by the main UUV under the same sampling time as a vector;
the vector is the actual quantity measurement Z k ；
Fusing the actual measurement of each sampling time by using a fusion algorithm until target tracking is completed;
the fusion algorithm is as follows:
step one, establishing an operation model of an underwater non-cooperative target;
step two, obtaining a nonlinear underwater non-cooperative target observation model and actual quantity measurement;
the actual quantity measurement is azimuth angle information of the underwater non-cooperative target which is actually measured relative to the UUV;
step three, obtaining a target posterior estimation state vector of the kth sampling time based on the nonlinear underwater non-cooperative target observation model in the step twoAnd a posterior error covariance matrix P for the kth sample time k∣k ；
Step four, obtaining a target state vector correction value by using the cerebellar neural network CMAC based on the step three
Step five, estimating a state vector based on the target posterior of the kth sampling time in the step threeAnd step four target state vector correction value +.>Obtaining a corrected state estimation vector, and completing target tracking of the kth sampling time;
and step six, repeatedly executing the steps three to five until target tracking is completed.
2. The underwater target collaborative trajectory tracking method according to claim 1, wherein: establishing an operation model of the underwater non-cooperative target in the first step; the specific process is as follows:
the discrete time state equation of the running model of the underwater non-cooperative target is as follows:
wherein,for the state vector of the underwater non-cooperative target at the moment k, X k-1 U is the state vector of the underwater non-cooperative target at the time k-1 k For inputting vectors, W k Is 0 mean Gaussian noise sequence, F k|k-1 For state transition matrix>Acceleration is the underwater non-cooperative target;
(x k ,y k ) Representing the position of the non-cooperative target under water,is underwater non-cooperative target speed,/->Acceleration is the underwater non-cooperative target;
the state transition matrix F k|k-1 The expression is
Wherein alpha is 1 ,α 2 Respectively the maneuvering frequencies of the underwater non-cooperative targets, and t is a sampling period;
the input vector U k The expression is
Wherein T represents the transpose,representing intermediate variables.
3. The underwater target collaborative trajectory tracking method according to claim 2, characterized in that: in the second step, a nonlinear underwater non-cooperative target observation model and actual quantity measurement are obtained:
wherein, (x) l ,y l ) For UUV position, l is UUV number, (x) k ,y k ) Representing an underwater non-cooperative target location;
the actual amount is measuredZ k ＝[z 1 ,…,z l ,…,z N′ ] T N' is the number of the measurement of the k moment and { z 1 ,…,z l ,…,z N′ -all quantity measurements at time k; unknown measurement noise v k The mean value is 0, and the variance is R; h (X) k ) Representing the relative azimuth angle of the non-cooperative target as an observation equation;
the actual quantity measurement is the azimuth angle information of the underwater non-cooperative target which is actually measured relative to the UUV.
4. A method for collaborative trajectory tracking of an underwater target according to claim 3 wherein: in the third step, a target posterior estimation state vector of the kth sampling time is obtained based on the nonlinear underwater non-cooperative target observation model and the actual quantity measurement in the second stepAnd a posterior error covariance matrix P for the kth sample time k∣k The method comprises the steps of carrying out a first treatment on the surface of the The specific process is as follows:
step three, estimating state vector of k-1 moment targetAnd the k-1 th sampling time error covariance matrix P k-1|k-1 Calculating a priori estimated state vector of the object at the kth sampling time>Priori error covariance matrix P of kth sampling time k|k-1 ；
Wherein T is transposed, Q k A process noise covariance matrix;
step three, the prior error covariance matrix P of the kth sampling time k|k-1 Performing cholesky decomposition to obtain triangular matrix S k|k-1 ；
Based on triangular matrix S k|k-1 And a priori estimated state vector of the object at the kth sampling timeObtaining a volume point of the state vector;
obtaining a predicted observed volume point based on a state vector volume point
Obtaining expectations of predictive observations based on volumetric points of the predicted observations
Expectations based on predictive observationsA priori estimated state vector of the object at the kth sampling time +.>Volume point of predictive observance>The method comprises the steps of obtaining a cross covariance matrix of a state vector and a prediction observance quantity by a volume point of the state vector;
said cholesky is a square root method;
the specific process is as follows:
wherein S is k|k-1 For P k|k-1 A triangular matrix obtained after cholesky decomposition,for state vector X k Zeta of the ith volume point of (c) i For the i-th volume point->Hope +.>I is the sequence number and m is the state vector X k Dimension of->Is an observation equation; />For predicting observance->P is as follows xz,k|k-1 A cross covariance matrix for the state vector and the predicted observed quantity;
the [1] is an m multiplied by 1 vector of which the elements are all 1;
step III, a cross covariance matrix P based on state vectors and prediction observables xz , k|k-1 And the prior error covariance matrix P of the kth sampling time k|k-1 Constructing a pseudo-linear measurement matrixThe expression is:
step three, calculating a measurement noise covariance matrix in an exponential weighted average form,
wherein R is k Measuring a noise covariance matrix for k time, R k-1 The noise covariance matrix is measured for time k-1,to measure the prediction error vector, Z k The azimuth information of the underwater non-cooperative target relative to the UUV is actually measured; beta k For the cycle parameter at time k, beta k-1 Taking the initial value beta as the cyclic parameter at the moment k-1 0 =1, b is a progressive factor;
step III, calculating a nominal observation vector based on step III, step III and step IIISum-of-error covariance matrix
Based on the third step, the second step, the third step, the fourth step and the unknown parameter theta k Calculating a correction error covariance matrix P zz,k∣k-1 ；
Iterative determination of unknown parameter θ k Until the condition is satisfied, theta when the minimum fixed point iteration number is obtained k A value;
the specific process is as follows:
calculating a nominal observation vector based on the third step, the first step, the second step and the third stepAnd error covariance matrix->
Based on the third step, the second step, the third step, the fourth step and the unknown parameter theta k Calculating a correction error covariance matrix P zz,k∣k-1 ；
Parameters for the j-th iteration;
motionless point iteration parameter determinationInitialization->And (3) circulation:
up toThen->
Wherein P is zz,k∣k-1 Correcting an error covariance matrix;the parameter values of the j and j+1 iterations are respectively;the corrected error covariance matrix after iteration j times; />As an auxiliary function, the parameters +.>Increment in the j+1th iteration; tr { } is the trace of the matrix, u k Is a proportional matrix, N' is the number of the measurement of the k moment, and +.>Is an unknown parameter theta k Estimated approximation of>θ to ensure a converged minimum number of iterations k Value of N min Epsilon is an infinitely small real number for the minimum iteration number;
step III, six, updating Kalman filtering gainUpdating the target posterior estimated state vector for the kth sample time +.>Updating a posterior error covariance matrix P for a kth sample time k∣k The method comprises the steps of carrying out a first treatment on the surface of the The specific process is as follows:
wherein,for iteration N min Error covariance matrix after the second time, +.>For the kalman filter gain,as a nominal observation vector, I n Is a unit matrix; />And P k∣k The target posterior estimated state vector of the kth sampling time and the posterior error covariance matrix of the kth sampling time are obtained.
5. The underwater target cooperative track following method according to claim 4, wherein: the target state vector correction value is obtained by using a cerebellar neural network CMAC based on the third step; the specific process is as follows:
cerebellar neural network CMAC input: kalman filter gainObservation vector innovation->Difference between predicted value and filtered estimated value +.>
Neural network output: target state vector correction value
Wherein,for the target state vector correction value, Z k For the actually measured azimuth information of the underwater non-cooperative target relative to the UUV,/the underwater non-cooperative target is a target of the UUV>For the observation equation +.>Estimating a state vector for the target posterior at the kth sampling time,>the state vector is estimated a priori for the target at the kth sample time.
6. The underwater target cooperative track following method according to claim 5, wherein: in the fifth step, the state vector is estimated based on the target posterior of the kth sampling time in the third stepAnd step four, target state vector correctionValue->Obtaining a corrected state estimation vector, and completing target tracking at the moment k;
the corrected state estimation vector is:
CN202311284679.7A 2023-10-07 2023-10-07 Underwater target collaborative track tracking method Active CN117031473B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311284679.7A CN117031473B (en) 2023-10-07 2023-10-07 Underwater target collaborative track tracking method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311284679.7A CN117031473B (en) 2023-10-07 2023-10-07 Underwater target collaborative track tracking method

Publications (2)

Publication Number Publication Date
CN117031473A CN117031473A (en) 2023-11-10
CN117031473B true CN117031473B (en) 2024-01-12

Family

ID=88628510

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311284679.7A Active CN117031473B (en) 2023-10-07 2023-10-07 Underwater target collaborative track tracking method

Country Status (1)

Country Link
CN (1) CN117031473B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2169422A1 (en) * 2008-09-24 2010-03-31 Whitehead Alenia Sistemi Subacquei S.p.A. System and method for acoustic tracking an underwater vehicle trajectory
JP2011247624A (en) * 2010-05-24 2011-12-08 Furuno Electric Co Ltd Underwater detection device and underwater detection method
CN102323586A (en) * 2011-07-14 2012-01-18 哈尔滨工程大学 UUV (unmanned underwater vehicle) aided navigation method based on current profile
CN110044356A (en) * 2019-04-22 2019-07-23 北京壹氢科技有限公司 A kind of lower distributed collaboration method for tracking target of communication topology switching
CN112711025A (en) * 2019-10-24 2021-04-27 中国科学院声学研究所 Underwater multi-station combined multi-target tracking method and system
CN114415157A (en) * 2021-12-30 2022-04-29 西北工业大学 Underwater target multi-model tracking method based on underwater acoustic sensor network
CN116125375A (en) * 2022-12-07 2023-05-16 云南民族大学 CKF-SLAM (continuous wave inertial navigation system-SLAM) -based improved unmanned underwater vehicle dynamic target tracking method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9223021B2 (en) * 2012-10-22 2015-12-29 The United States Of America As Represented By The Secretary Of The Army Method and system for motion compensated target detection using acoustical focusing

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2169422A1 (en) * 2008-09-24 2010-03-31 Whitehead Alenia Sistemi Subacquei S.p.A. System and method for acoustic tracking an underwater vehicle trajectory
JP2011247624A (en) * 2010-05-24 2011-12-08 Furuno Electric Co Ltd Underwater detection device and underwater detection method
CN102323586A (en) * 2011-07-14 2012-01-18 哈尔滨工程大学 UUV (unmanned underwater vehicle) aided navigation method based on current profile
CN110044356A (en) * 2019-04-22 2019-07-23 北京壹氢科技有限公司 A kind of lower distributed collaboration method for tracking target of communication topology switching
CN112711025A (en) * 2019-10-24 2021-04-27 中国科学院声学研究所 Underwater multi-station combined multi-target tracking method and system
CN114415157A (en) * 2021-12-30 2022-04-29 西北工业大学 Underwater target multi-model tracking method based on underwater acoustic sensor network
CN116125375A (en) * 2022-12-07 2023-05-16 云南民族大学 CKF-SLAM (continuous wave inertial navigation system-SLAM) -based improved unmanned underwater vehicle dynamic target tracking method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
基于UUV的目标非声探测技术发展及趋势分析;李阁阁等;水下无人系统学报;第31卷(第4期);第510-520页 *
基于变维卡尔曼滤波的UUV目标运动优化估计方法;严浙平;郝悦;王千一;边信黔;;传感技术学报(第12期);第1637-1642页 *

Also Published As

Publication number Publication date
CN117031473A (en) 2023-11-10

Similar Documents

Publication Publication Date Title
CN109459040B (en) Multi-AUV (autonomous Underwater vehicle) cooperative positioning method based on RBF (radial basis function) neural network assisted volume Kalman filtering
CN108896047B (en) Distributed sensor network collaborative fusion and sensor position correction method
CN107390199B (en) A kind of radar maneuvering target tracking waveform design method
CN109581281B (en) Moving target positioning method based on arrival time difference and arrival frequency difference
CN109671100B (en) Distributed variable diffusion combined coefficient particle filter direct tracking method
CN115307643A (en) Double-responder assisted SINS/USBL combined navigation method
Wanasinghe et al. Decentralized cooperative localization approach for autonomous multirobot systems
CN117390498B (en) Flight capability assessment method of fixed wing cluster unmanned aerial vehicle based on Transformer model
CN117031473B (en) Underwater target collaborative track tracking method
CN109341690A (en) A kind of efficient combined navigation self-adaptive data fusion method of robust
Witzgall et al. Single platform passive Doppler geolocation with unknown emitter frequency
CN114445459B (en) Continuous-discrete maximum correlation entropy target tracking method based on variable decibel leaf theory
CN112835020B (en) Rigid body positioning method for non-line-of-sight parameter estimation
CN114111796B (en) Parallel fusion positioning method and system of underwater unmanned robot based on information gain
CN112285697B (en) Multi-sensor multi-target space-time deviation calibration and fusion method
CN115728710A (en) Robust TDOA (time difference of arrival) positioning method based on variable center maximum entropy criterion
CN113503891B (en) SINSDVL alignment correction method, system, medium and equipment
CN109282820B (en) Indoor positioning method based on distributed hybrid filtering
CN109856624B (en) Target state estimation method for single-radar linear flight path line
CN107590509B (en) Cherenov fusion method based on maximum expectation approximation
CN112946568A (en) Radiation source track vector direct estimation method
CN109684771A (en) Maneuvering target state prediction optimization method based on interactive multi-model
CN111796271B (en) Target tracking method and device under constraint of proportional guidance destination
Oliveira et al. GNSS-Denied Joint Cooperative Terrain Navigation and Target Tracking Using Factor Graph Geometric Average Fusion
RU2263927C2 (en) Method of evaluating parameters of trajectory of radio-frequency radiation sources in two-positioned passive goniometrical radar station

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
GR01 Patent grant
GR01 Patent grant