CN-121114962-B - Target tracking algorithm based on multi-station foundation radar relay

CN121114962BCN 121114962 BCN121114962 BCN 121114962BCN-121114962-B

Abstract

The invention discloses a target tracking algorithm based on multi-station foundation radar relay, which comprises the following steps of firstly, constructing a target tracking model; the method comprises the steps of preprocessing multi-radar measurement data, constructing a weighted observation fusion observation equation, independently operating a local filter, estimating the global optimal state of a main filter, optimizing the result feedback, processing the measurement data of the multi-station foundation radar through the weighted observation fusion technology, effectively expanding the radar detection range, solving the problem of insufficient coverage of a single sensor in the traditional relay tracking method, reducing the influence of single sensor measurement noise on an estimation result, improving the estimation precision, converting the prior knowledge of target motion into a probabilistic form through a constructed target tracking model, and improving the tracking precision through utilizing the prior knowledge expressed in the probability form by using a detection tracking joint processing method based on the Bayesian theory.

Inventors

SONG SHENMIN
YANG YUHANG
LIU JINGANG

Assignees

哈尔滨工业大学

Dates

Publication Date: 20260508
Application Date: 20250929

Claims (6)

1. The target tracking algorithm based on multi-station foundation radar relay comprises the following steps of constructing a target tracking model, preprocessing multi-radar measurement data, constructing a weighted observation fusion observation equation, independently operating a local filter, estimating a global optimal state of a main filter, and optimizing result feedback, and is characterized in that: In the first step, the position, the speed and the acceleration of a target under a geocentric fixedly connected coordinate system are selected as state quantities, and a target motion equation and a measurement equation are established as a target tracking model; In the second step, the original measurement data acquired by a plurality of foundation radars are subjected to space alignment processing, and radar measurement data under different coordinate systems are converted into the same geocentric fixation coordinate system; In the third step, a weighting method is adopted to combine the measurement equations of a plurality of radars, and a weighted observation fusion observation equation is constructed; In the fourth step, the measurement data preprocessed in the second step is input into each local filter, and each local filter is independently operated by adopting a Kalman filtering algorithm based on a target motion equation to obtain a local target state estimation value and an estimation error covariance matrix and transmits the local target state estimation value and the estimation error covariance matrix to the main filter; in the fifth step, the main filter fuses the preprocessed measurement data by using the weighted observation fusion observation equation constructed in the third step, and calculates and obtains a global target state estimation value and a corresponding global estimation error covariance matrix based on an optimal distribution fusion algorithm by combining the outputs of all local filters; in the sixth step, the global estimation result output by the main filter is fed back to each local filter, the local filter adjusts the self filtering parameters by using the feedback information, and the local target state estimation precision at the next sampling moment is optimized; In the third step, the process of constructing the weighted observation fusion observation equation comprises the steps of merging radar measurement equations by using an augmented observation vector method to obtain a concentrated fusion observation equation: , , , , , Wherein the method comprises the steps of In order to centralize the fusion of the observation vectors, In order to fuse the array of observations, The Kalman filtering algorithm is applied to the state equation and the fusion observation equation to obtain the centralized observation fusion global optimal Kalman filter ; From the concentrated fusion of observation equations, a state vector is obtained The weighted least squares estimate of (2) is: , Then there is a weighted observation fusion observation equation: , wherein the weighted fused observation vector And observing white noise The method comprises the following steps of: , , the variance matrix of (2) is: , The Kalman filter for the weighted observation fusion can be obtained by applying Kalman filtering algorithm to the state equation and the weighted observation fusion observation equation ; Wherein the method comprises the steps of Is the first The measurement vector of the portion radar, Is the first An observation matrix of the portion radar, Is the first Observing white noise of the partial radar; in the fifth step, the optimal distribution fusion algorithm specifically comprises: , , Wherein the method comprises the steps of And To merge the center to the target state estimation and the error covariance matrix, And Is the first The radar pair target state estimates and their estimated error covariance matrix, And For the fusion center to forecast the target state and forecast error covariance matrix, And Is the first The radar forecast the target state and forecast the error covariance matrix, Is the radar quantity.
2. The target tracking algorithm based on multi-station ground-based radar relay according to claim 1, wherein in the first step, the target motion equation is specifically: Location based Speed and velocity of And acceleration Constructed target state vector The method comprises the following steps: , As a result of: , , Then there are: , in the formula, As a result of the location of the object, Is the average value of the acceleration, is assumed to be constant in the sampling period, Relative to the target acceleration Is used to determine the amount of deviation of (c), Is the reciprocal of the maneuver time constant, Is system noise; The continuous form of the target motion equation is: , The discretized form of the target motion equation is: , the state transition matrix is: , The array is as follows: , the system noise driving matrix is: , in the formula, For the system sampling step size, Is the reciprocal of the maneuver time constant, Smaller values indicate greater maneuver intensity and for cornering maneuvers, And, for evasion maneuvers, And, in the case of an atmospheric disturbance, 。
3. The target tracking algorithm based on multi-station ground-based radar relay according to claim 1, wherein in the first step, the measurement equation is specifically: , Wherein the method comprises the steps of As the true state vector of the object, Is the total number of radars.
4. The target tracking algorithm based on multi-station foundation radar relay according to claim 1, wherein in the second step, the spatial alignment processing specifically comprises the steps of taking a geocentric fixation coordinate system as a reference coordinate system, and uniformly converting target position and speed measurement data acquired by each radar under a local coordinate system thereof into the geocentric fixation coordinate system through a coordinate conversion formula.
5. The method of claim 1, wherein in the fourth step, the independent operation of Kalman filtering algorithm includes two stages of prediction and updating, the state prediction value at the current time is calculated by the local state estimation value and the state transition matrix at the sampling time above the prediction stage, and the prediction value is corrected by the current measurement data in the updating stage to obtain the local target state estimation value.
6. The target tracking algorithm based on multi-station ground-based radar relay according to claim 1, wherein in the sixth step, the global estimation result includes a global target state estimation value, a global estimation error covariance matrix and a process noise matrix.

Description

Target tracking algorithm based on multi-station foundation radar relay Technical Field The invention relates to the technical field of computers, in particular to a target tracking algorithm based on relay of a multi-station foundation radar. Background In network center operations, the sensors are required to cooperate to better detect the target. Relay tracking is an important component of collaborative combat, and has wide application prospect in practice. When each sensor has a different observation area and all the target track data is detectable but the resulting track has no intersection, this situation similar to a relay game is called relay tracking. In the prior art, the relay tracking method specifically comprises the steps that after a certain sensor platform receives target observation data and completes primary tracking processing, the data is transmitted to a sensor corresponding to an observation area into which a target possibly enters according to the movement trend of the target, the sensor waits for the target to appear in the detection range and continues tracking, and the target processing is carried out by adopting the traditional detection and tracking method. The method has the following defects that 1, each sensor only works by depending on a self-fixed observation area, the observation range of a plurality of sensors is not integrated and expanded, the whole detection range is limited by the physical detection boundary of a single sensor, when the motion track of a target exceeds the detection range of the single sensor and an adjacent sensor transmitted by relay, the problem of tracking interruption is easy to occur, and 2, in the existing relay tracking method, a subsequent sensor only passively waits for the target in the detection range of the sensor, the prior knowledge of the motion information of the target provided by the previous sensor is not utilized, and the tracking precision is low. Disclosure of Invention The invention aims to provide a target tracking algorithm based on multi-station foundation radar relay so as to solve the problems in the background technology. The target tracking algorithm based on the multi-station foundation radar relay comprises the following steps of constructing a target tracking model, preprocessing multi-radar measurement data, constructing a weighted observation fusion observation equation, independently operating a local filter, estimating a global optimal state of a main filter, and carrying out feedback optimization on a result; In the first step, the position, the speed and the acceleration of a target under a geocentric fixedly connected coordinate system are selected as state quantities, and a target motion equation and a measurement equation are established as a target tracking model; In the second step, the original measurement data acquired by a plurality of foundation radars are subjected to space alignment processing, and radar measurement data under different coordinate systems are converted into the same geocentric fixation coordinate system; In the third step, a weighting method is adopted to combine the measurement equations of a plurality of radars, and a weighted observation fusion observation equation is constructed; In the fourth step, the measurement data preprocessed in the second step is input into each local filter, and each local filter is independently operated by adopting a Kalman filtering algorithm based on a target motion equation to obtain a local target state estimation value and an estimation error covariance matrix and transmits the local target state estimation value and the estimation error covariance matrix to the main filter; in the fifth step, the main filter fuses the preprocessed measurement data by using the weighted observation fusion observation equation constructed in the third step, and calculates and obtains a global target state estimation value and a corresponding global estimation error covariance matrix based on an optimal distribution fusion algorithm by combining the outputs of all local filters; in the sixth step, the global estimation result output by the main filter is fed back to each local filter, and the local filter adjusts its own filtering parameters by using the feedback information, so as to optimize the estimation accuracy of the local target state at the next sampling time. Preferably, in the first step, the target motion equation is specifically: Location based Speed and velocity ofAnd accelerationConstructed target state vectorThe method comprises the following steps: As a result of: Then there are: in the formula, As a result of the location of the object,Is the average value of the acceleration, is assumed to be constant in the sampling period,Relative to the target accelerationIs used to determine the amount of deviation of (c),Is the reciprocal of the maneuver time constant,Is system noise; The continuous form of the target motion equation is: The discretized form of