Heterogeneous radar data fusion method and system for maritime radar system

Abstract

The invention belongs to the technical field of maritime radars and navigation and positioning, and particularly discloses a heterogeneous radar data fusion method and system for a maritime radar system. The method comprises the following steps: firstly, by taking a target ship trajectory obtained by a GPS / Beidou as a space-time reference basis, carrying out processing such as acquisition and extraction, data cleaning and preprocessing, joint matching based on time and trajectory, coordinate calibration and conversion and the like on original data of two heterogeneous radars, and converting the original data of the two heterogeneous radars into a unified GPS / Beidou coordinate system; then, a self-adaptive weighted trajectory fusion method based on improved local information entropy and a multi-scale sliding window is provided, probability modeling and local information entropy calculation are carried out on observation residual errors in the sliding windows of different spatial scales, and in combination with the prior weight of radar performance, the fusion weight of trajectory points is subjected to self-adaptive adjustment; therefore, high-precision track fusion of heterogeneous radar data is realized.

Description

Technical Field

[0001] This invention belongs to the field of marine radar and navigation and positioning technology, and specifically relates to a heterogeneous radar data fusion method and system for marine radar systems, which is particularly suitable for high-precision trajectory fusion between two heterogeneous radar data. Background Technology

[0002] With the intelligent development of global maritime transportation, the demand for dynamic perception of complex sea conditions (such as fog and strong clutter) has significantly increased. Radar systems have become core sensors for ship navigation, maritime situation monitoring, and collision avoidance decision-making. Among the current mainstream maritime radar equipment, solid-state pulse compression radar (SPL) is widely used for shore-based monitoring and long-range target early warning due to its long detection range and strong electromagnetic penetration capability. Electronically Scanned Array Radar (ESA), with its higher spatial resolution and faster target refresh rate, has become the preferred equipment for high-precision near-field tracking of intelligent ships. These two types of radar have fundamental differences in their working mechanisms. Specifically, SPL uses pulse compression technology to achieve long-range detection, and the data format is mainly polar coordinates (range, azimuth). ESA radar, on the other hand, uses phased array technology to achieve electronic scanning and outputs point cloud data in Cartesian coordinates (X, Y). The heterogeneity of the two presents a natural challenge for data fusion.

[0003] In the dynamic maritime environment, data from a single radar often fails to comprehensively and accurately reflect the true state of a target vessel. To improve the accuracy of target detection and trajectory tracking, and enhance the robustness of the system under complex conditions, multi-radar information fusion technology has become a key research focus in recent years. Fusion technology not only requires spatial coordinate transformation and temporal alignment of data from different radars, but also deep data fusion at the target association and trajectory reconstruction levels. This fully leverages the complementary advantages of multi-source perception to achieve collaborative perception and joint estimation of targets. Therefore, there is an urgent need for a multi-radar fusion method adapted to the dynamic characteristics of ship targets in the maritime environment. This method should be able to fuse target observation data from different types of radars under a unified reference coordinate system, construct a unified multi-source target trajectory representation, improve the positioning accuracy and tracking continuity of ship targets in complex environments, and thus provide stable and reliable perception support for intelligent ship navigation systems and maritime monitoring platforms.

[0004] Existing multi-radar fusion methods largely draw upon multi-sensor fusion technologies from the land transportation sector. These methods are typically designed for sensors such as lidar and cameras, relying on relatively stable environmental characteristics and fixed motion patterns, and are usually based on high-precision GPS / IMU-assisted positioning. Directly transferring these fusion strategies to the maritime domain presents significant challenges, such as high uncertainty in target motion states, lack of environmental structure, and strong radar measurement noise. Furthermore, maritime radar data often contains strong dynamic interference and background clutter, further increasing the difficulty of multi-source data correlation and fusion. In summary, directly transferring existing multi-sensor fusion strategies from the land transportation sector to the maritime environment presents the following technical problems: 1. Lack of multi-radar data fusion methods suitable for the maritime environment: Existing multi-sensor fusion methods mostly serve land transportation systems and are difficult to directly apply to the data fusion of pulse compression radar and electronically scanned radar in the maritime environment. Maritime scenarios involve highly dynamic targets and complex background interference, and there is a lack of dedicated fusion technologies that can efficiently adapt to the characteristics of marine sensors. 2. Difficulty in spatiotemporal alignment between multiple radar data, resulting in weak fusion foundation: Differences in sampling frequency, resolution, and coordinate systems among different types of radar make it difficult to achieve accurate time synchronization and spatial mapping before target registration and data fusion, severely affecting fusion accuracy and stability. 3. Large target information matching errors, leading to insufficient fusion accuracy and reliability: Different radar observation targets in the maritime environment have different feature representation methods, making target matching prone to mismatches or omissions. This results in discontinuous and abrupt changes in the fused trajectory, impacting the overall performance and safety assurance capabilities of the ship navigation system. Summary of the Invention

[0005] The purpose of this invention is to propose a heterogeneous radar data fusion method for maritime radar systems. This method converts two different types of radar data into a unified GPS / BeiDou coordinate system through data calibration. At the same time, it proposes an adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding window to achieve adaptive weighted trajectory fusion between two heterogeneous radar data, thereby improving the trajectory fusion accuracy between heterogeneous radars in dynamic maritime environments.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A heterogeneous radar data fusion method for marine radar systems includes the following steps: Step 1. Using the target ship trajectory obtained by GPS / BeiDou as the spatiotemporal reference benchmark, the raw data of the two heterogeneous radars are converted to a unified GPS / BeiDou coordinate system through the steps of data collection and extraction, data cleaning and preprocessing, joint matching based on time and trajectory, coordinate calibration and transformation, thus completing the data calibration before fusion. Step 2. An adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding window is proposed. The observation residuals are probabilistically modeled and local information entropy is calculated within sliding windows of different spatial scales. Combined with the prior weights of heterogeneous radar performance, the fusion weights of trajectory points are adaptively adjusted to achieve the fusion of heterogeneous radar trajectories.

[0007] Furthermore, based on the aforementioned heterogeneous radar data fusion method for marine radar systems, this invention further proposes a corresponding heterogeneous radar data fusion system for marine radar systems, which adopts the following technical solution: A heterogeneous radar data fusion system for marine radar systems includes the following modules: The data calibration module is used to collect and extract raw data from two heterogeneous radars, clean and preprocess the data, perform joint matching based on time and trajectory, and perform coordinate calibration and transformation, using the target ship trajectory obtained by GPS / BeiDou as a spatiotemporal reference benchmark. It converts the raw data of the two heterogeneous radars into a unified GPS / BeiDou coordinate system, thus completing the data calibration before fusion. The data fusion module proposes an adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding windows. It performs probabilistic modeling and local information entropy calculation on the observation residuals within sliding windows of different spatial scales, and adaptively adjusts the fusion weights of trajectory points by combining the prior weights of heterogeneous radar performance, so as to achieve the fusion of heterogeneous radar trajectories.

[0008] The present invention has the following advantages: As described above, this invention discloses a heterogeneous radar data fusion method for maritime radar systems. This method comprises two parts: data calibration before fusion and data fusion itself. In the data calibration stage, this invention modularizes steps such as data preprocessing, trajectory matching, and coordinate transformation, unifying data formats, reducing intermediate conversion steps, lowering computational complexity, and improving overall processing efficiency. Furthermore, by introducing a unified GPS / BeiDou reference trajectory as the fusion benchmark, combined with spatial rigidity transformation and time synchronization mechanisms, it ensures a high degree of spatial and temporal consistency between different types of radar data (e.g., pulse compression radar and electronically scanned radar). This lays a unified data foundation for subsequent fusion weight calculation and residual modeling, effectively reducing data drift and error accumulation, and significantly improving the target positioning and tracking accuracy after fusion. Through the tight coupling of the data calibration and data fusion stages, this invention ensures the comparability of residuals and the accuracy of weight estimation, resulting in higher consistency and reliability in the fusion process. In the data fusion section, this invention innovatively proposes an adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding windows. Within sliding windows of different spatial scales, probabilistic modeling and local information entropy calculation of observation residuals are performed. Combined with prior weights of heterogeneous radar performance, the fusion weights of trajectory points are adaptively adjusted to achieve the fusion of heterogeneous radar trajectories. This method, through the combined effect of multi-scale residual information entropy and prior performance weights, can significantly improve the contribution of low-noise observation data while dynamically suppressing the influence of high-noise or uncertain observation points, thereby effectively improving the continuity, smoothness, and accuracy of the fused trajectory. This invention is particularly suitable for rapidly changing maritime environments. Employing an efficient data association and matching algorithm combining timestamp nearest neighbor matching and spatial nearest neighbor matching, it can quickly establish correspondences between different radar observation data in large-scale data streams, ensuring the real-time and correctness of trajectory point pairs, meeting the timeliness requirements of online fusion, and possessing excellent real-time response capabilities. It can accurately fuse multi-source radar information during rapid ship movement, improving the system's dynamic perception capability and navigation robustness in complex scenarios. This invention is applicable to various types of maritime radar systems, has good versatility, realizes data fusion of heterogeneous multi-source sensing systems, and provides basic support for building a high-precision maritime intelligent sensing platform. Attached Figure Description

[0009] Figure 1 This is a flowchart of the heterogeneous radar data fusion method for a maritime radar system in Embodiment 1 of the present invention; Figure 2 The images shown are the pulse compression radar, electronic scanning radar, and GPS trajectory maps after data preprocessing in Embodiment 1 of the present invention. Figure 3 This is a comparison diagram of the converted electronically scanned radar and GPS trajectories in Embodiment 1 of the present invention; Figure 4 This is a comparison diagram of the converted solid-state pulse compression radar and GPS trajectory in Embodiment 1 of the present invention; Figure 5 This is a diagram showing the trajectory alignment effect after conversion in Embodiment 1 of the present invention; Figure 6 This is a flowchart of the weighted trajectory fusion algorithm based on local information entropy in Embodiment 1 of the present invention; Figure 7 This is a comparison diagram of the fused trajectories in Embodiment 1 of the present invention; Figure 8 This is a comparison diagram of the fused trajectory and the true AIS value in the longitude direction in Embodiment 1 of the present invention; Figure 9 This is a comparison diagram of the fused trajectory and the true AIS value in the latitudinal direction in Embodiment 1 of the present invention; Figure 10 This is a comparison chart of the total error between the fused trajectory and the true AIS value (latitude and longitude) in Embodiment 1 of the present invention; Figure 11 This is an error distribution map of latitude and longitude distance after being processed by the local information entropy weighted fusion algorithm in Embodiment 1 of the present invention; Figure 12 This is the error distribution diagram of the flight speed after the local information entropy weighted fusion algorithm in Embodiment 1 of the present invention; Figure 13 This is the error distribution diagram of the heading after the local information entropy weighted fusion algorithm in Embodiment 1 of the present invention; Figure 14 This is the error distribution diagram of the trajectory after the local information entropy weighted fusion algorithm in Embodiment 1 of the present invention. Detailed Implementation

[0010] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: Example 1 This embodiment describes a heterogeneous radar data fusion method for maritime radar systems. This method uses the high-precision trajectory of a target vessel acquired through GPS / BeiDou as a spatiotemporal reference benchmark, fusing heterogeneous observation data from two different types of radars, A and B. Through coordinate system transformation, time synchronization calibration, and trajectory feature matching, the spatiotemporal differences between the multi-source radars are eliminated. Under a unified geographic coordinate system, an optimized fusion algorithm is used to jointly process the multi-source target information, achieving high-precision joint estimation and dynamic reconstruction of the vessel trajectory. This method can significantly improve the spatiotemporal consistency and robustness of multi-radar sensing data in complex maritime environments, providing highly reliable fusion sensing technology support for real-time target detection, continuous tracking, and intelligent navigation.

[0011] In this embodiment 1, radar A and radar B are, for example, a solid-state pulse compression radar and an electronically scanned array radar, respectively. This embodiment 1 achieves data fusion of solid-state pulse compression radar and electronically scanned array radar through the proposed steps, aiming to significantly improve the spatial consistency, temporal synchronization and anti-interference robustness of trajectory data.

[0012] The method of this invention is generally divided into two main steps: a data calibration step before fusion and a data fusion step. The data calibration step before fusion includes raw data acquisition and extraction, data cleaning and preprocessing, time- and trajectory-based joint matching, and coordinate calibration and transformation. The data fusion step uses adaptive weighted trajectory fusion based on improved Local Information Entropy (LIE) and a multi-scale sliding window. These two steps complement each other, ultimately forming a unified, high-precision multi-source fused ship trajectory.

[0013] like Figure 1 As shown, the heterogeneous radar data fusion method for a maritime radar system in this embodiment includes the following steps: Step 1. Using the target ship trajectory obtained by GPS / BeiDou as the spatiotemporal reference benchmark, the raw data of the two heterogeneous radars are converted to a unified GPS / BeiDou coordinate system through the steps of data collection and extraction, data cleaning and preprocessing, joint matching based on time and trajectory, and coordinate calibration and transformation, thus completing the data calibration before fusion.

[0014] Step 1.1. Raw data collection and extraction.

[0015] First, trajectory data of the target vessel is collected from various heterogeneous sensing devices, mainly including solid-state pulse compression radar, electronically scanned radar, and the BeiDou / GPS system. Among them, solid-state pulse compression radar and electronically scanned radar provide observation data containing fields such as longitude, latitude, track number (hjid), timestamp, speed over ground (SOG), and course over ground (COG).

[0016] The BeiDou / GPS system provides high-precision latitude and longitude location and timestamp information for unified reference. Step 1.1 establishes the basic data interface between the multi-source sensing systems, providing basic data support for subsequent registration and fusion.

[0017] The following example illustrates how to convert the raw data from two heterogeneous radars to a unified GPS coordinate system.

[0018] Step 1.2. Data cleaning and preprocessing.

[0019] Due to the inconsistent data structures, observation accuracy, and sampling frequencies of different types of radar equipment, it is necessary to perform unified cleaning and standardization of the raw data to ensure data comparability and consistency. First, key fields (such as track ID, longitude, latitude, timestamp, speed, and heading) are extracted from each source data; missing values, duplicate records, and stationary records with a speed of 0 are deleted; timestamps are converted to UTC format and normalized in seconds; all latitude and longitude data are uniformly represented using angle units (°); tracks are arranged in ascending order of timestamp, and linear interpolation is used to fill in track gaps to enhance time alignment accuracy; radar data is coarsely aligned with GPS tracks to lay the foundation for subsequent fine matching.

[0020] Step 1.3. Joint matching based on time and trajectory.

[0021] To ensure the spatiotemporal consistency of the same target across different radar data sources, a matching method combining time windows and spatial constraints is employed. First, data from pulse compression radar and electronically scanned radar are grouped separately by track number (hjid) and treated as the same target. Within a set time window (±2s), data is filtered based on conditions including a spatial distance of less than 50 meters, a speed difference of less than 1 knot, and a heading difference of less than 10 degrees. A nearest neighbor matching strategy is then used to select track point pairs that meet these constraints. Finally, a high-quality set of matched point pairs is output for calibration and fusion.

[0022] Step 1.4. Perform coordinate calibration and transformation based on the improved ICP algorithm.

[0023] To accurately map all radar observation data to a unified geographic coordinate system (GPS), an improved Iterative Closest Point (WICP) coordinate calibration algorithm is proposed.

[0024] Compared to the traditional ICP algorithm, the improved ICP calibration algorithm proposed in this embodiment introduces a weighted centroid mechanism to enhance robustness and matching accuracy in the presence of observation noise and incomplete point cloud overlap.

[0025] The detailed process for improving the ICP calibration algorithm is as follows: Step 1.4.1. Initial Setup and Point Cloud Definition. Assume the trajectory points in the radar data constitute the source point cloud. GPS trajectory data constitutes the target point cloud. .

[0026] in , ∈R 2Let represent two-dimensional coordinate points, where m and n are the number of source and target points, respectively. Initialize the rigid transformation matrix T = [R|t]. Where R ∈ R 2×2 Let be a rotation matrix, t∈R 2 Let R be the translation vector. Initially set R=I, t=0.

[0027] Step 1.4.2. Nearest point matching. For each source point... ∈P, find the nearest Euclidean distance matching point in the target point cloud Q. ∈Q, construct a set of matching point pairs .

[0028] Step 1.4.3. Weighted centroid calculation. To enhance the influence of key points in the calibration, confidence weights are introduced. (Usually related to factors such as point-to-point spacing and velocity consistency), calculate the weighted centroid for matched point pairs: , .

[0029] in , These are the weighted centroids of the source point cloud and the target point cloud, respectively.

[0030] Step 1.4.4. Decentralization of point clouds.

[0031] Centering all matching points yields a set of points with zero mean: , .

[0032] Step 1.4.5. Solve for the covariance matrix and rotation matrix.

[0033] Construct the weighted covariance matrix: .

[0034] Perform singular value decomposition (SVD) on H: .

[0035] Obtain the rotation matrix: .

[0036] To ensure that R is a positive definite rotation matrix (i.e., det(R) = 1), V needs to be corrected if det(R) = -1.

[0037] Step 1.4.6. Solve for the translation vector.

[0038] Solve for the translation vector based on the weighted centroid difference: .

[0039] Step 1.4.7. Rigid transformation update and iteration.

[0040] Combine the current rotation and translation into a new rigid transformation matrix T=[R|t], and update the coordinates of the source point cloud: .

[0041] Repeat steps 1.4.2 to 1.4.7 until the convergence condition is met (the convergence condition here refers to, for example, the change in the objective function error being lower than the threshold ε, or reaching the maximum number of iterations N).

[0042] Step 1.4.8. Solve the objective function using the optimal transformation.

[0043] Finally, the optimal rigid transformation is obtained by minimizing the following objective function:

[0044] .

[0045] in, Let w represent the Euclidean norm. i It is the confidence weight of the i-th pair of matching points.

[0046] Through the above calibration steps, the optimal rigid transformation matrix of the pulse compression radar and electronically scanned radar relative to the GPS coordinate system is obtained, and then all radar data are uniformly mapped to a high-precision geographic coordinate system to achieve consistent alignment of the coordinate systems.

[0047] This invention constructs a unified fusion framework based on reference trajectories, fusing observation data from pulse compression radar and electronically scanned array radar to establish target-level correspondences and improve the accuracy of target matching. By leveraging a spatial transformation model and a time synchronization mechanism, it achieves consistent representation of data from different radar systems in both spatial and temporal dimensions, providing a unified spatial framework for the weighted fusion of heterogeneous radars in subsequent steps, and significantly improving the performance of the fused trajectory in terms of position accuracy and time synchronization.

[0048] Step 2. An adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding window is proposed. The observation residuals are probabilistically modeled and local information entropy is calculated within sliding windows of different spatial scales. Combined with the prior weights of heterogeneous radar performance, the fusion weights of trajectory points are adaptively adjusted to achieve the fusion of heterogeneous radar trajectories.

[0049] After completing the coordinate calibration of multi-source radar data and unifying it to the GPS geographic coordinate system, this invention proposes an adaptive weighted trajectory fusion method based on improved Local Information Entropy (LIE) and multi-scale sliding windows to further improve the spatial accuracy and continuity of the fused trajectory and address the residual offset problem caused by different radar observation errors. This method dynamically assigns fusion weights based on differences in radar performance, achieving high-precision fusion of heterogeneous radar trajectories. Its core idea is to perform probabilistic modeling and local information entropy calculation on the observation residuals within sliding windows of different spatial scales, and adaptively adjust the fusion weights of trajectory points by combining prior weights based on radar performance, thus achieving robust estimation and weighted fusion to address data uncertainties.

[0050] like Figure 6 As shown, step 2 specifically involves: Step 2.1. Construct a set of matching trajectory point pairs for solid-state pulse compression radar and electronic scanning radar, and calculate the Euclidean distance in space for each pair of matching trajectory points in the set, i.e., the observation residual.

[0051] Each pair of matching trajectory points consists of the observation coordinates of the solid-state pulse compression radar and the electronically scanned radar at the same time.

[0052] Defined in step 1, during the coordinate calibration stage, the set of matching trajectory point pairs with high matching accuracy is as follows: .

[0053] in For radar A at time Observation coordinates , For radar B at the same time Corresponding matching observation coordinates The trajectory data of radar A (such as pulse compression radar) and radar B (such as electronically scanned radar) have been unified to the same GPS coordinate system. The Euclidean distance in space for each pair of matching trajectory points, i.e., the observation residual, is calculated using the following formula: ; All Construct distance vector Where N is the number of matching point pairs, .

[0054] Step 2.2. For each pair of matched trajectory points, the distribution probability of the observation residuals in the neighborhood is statistically analyzed within a sliding window at different spatial scales, and the local information entropy at the corresponding spatial scale is calculated.

[0055] For each pair of matched trajectory points Pi, within sliding windows of different spatial scales, the probability distribution of observed residuals in the neighborhood is statistically analyzed, and the corresponding local information entropy is calculated. Here, different spatial scales refer to sliding windows of different radii. For example, three sliding windows with different radii, r1=10m, r2=20m, and r3=50m, can be used to statistically analyze the residual distribution at different scales.

[0056] For each spatial scale r m The sliding window is defined with the following neighborhood set: .

[0057] in For trajectory points The coordinates; For the trajectory point Coordinates of trajectory points at the same or adjacent moments; The Euclidean distance between the two points; For scale Below are the trajectory points It is the set of trajectory points in the neighborhood of the center.

[0058] The observation residuals of each pair of matching trajectory points in the neighborhood, i.e., the Euclidean distance between the observation coordinates of the two radars for the same target, are denoted as... Divide the data into B intervals and count the frequency in each interval. Calculate the residual probability distribution: .

[0059] in This indicates that the observation residual within the neighborhood falls into the first... The probability of each interval; Indicates the first The frequency of each interval, i.e. the number of trajectory points; This represents the total number of trajectory points within the neighborhood.

[0060] Calculate the current spatial scale r m Local information entropy The formula is as follows: .

[0061] Finally, the local information entropy at different spatial scales is weighted and fused to obtain the trajectory points. Multi-scale local information entropy: .

[0062] in, To prevent tiny positive numbers from reaching the zero of the logarithm, Weights for different scales.

[0063] This invention introduces a multi-scale sliding window mechanism into the residual analysis of multi-radar observations. By probabilistically modeling the distribution of observation residuals across multiple spatial scales (i.e., windows with different radii, such as 10 meters, 20 meters, and 50 meters), local information entropy indices are calculated for each scale. Multi-scale analysis can comprehensively characterize the local consistency and uncertainty features of observation residuals, avoiding noise sensitivity due to excessively small neighborhoods or smoothing distortion caused by excessively large neighborhoods at a single scale. This method enables the fusion algorithm to simultaneously perceive fine-grained local anomalies and large-scale trends, significantly improving the spatial accuracy, anomaly detection capability, and robustness of trajectory fusion.

[0064] Step 2.3. Design prior weights based on the performance of the two heterogeneous radars.

[0065] To reasonably reflect the observation reliability of different radars under different ranging ranges, target characteristics, and observation environments, prior weights are used. , The values ​​are assigned based on the historical measurement accuracy, noise level, stability index, and ranging capability of each radar.

[0066] Define X∈{A,B} to represent radar A or radar B, then the prior weights , Calculate using the following formula: .

[0067] in denoted as the root mean square error of radar X's historical observations, corresponding to the historical measurement accuracy. The smaller the value, the lower the long-term observation error of the radar, and the greater the prior weight. It represents the moving average of the observation residuals calculated within a fixed-length sliding time window (e.g., ±5 seconds) centered on the current trajectory point timestamp. It corresponds to the short-term noise level and stability index. The larger the value, the stronger the recent observation volatility, and the lower the prior weight. These are constant coefficients used to adjust the relative initial weights of the radar; To prevent division by zero of constants; Adjust parameters to observe residual sensitivity; This is the distance correction factor, corresponding to the ranging capability. It is used to dynamically adjust the weight based on the current distance R of the target being measured. It usually decreases monotonically as the distance increases to reflect the impact of decreased accuracy in long-distance observation.

[0068] The current target distance refers to the distance between the target ship observed by the radar and the radar itself.

[0069] Step 2.4. For each pair of matched trajectory points, design a corresponding adaptive fusion weight function based on the observation residual, multi-scale local information entropy, and prior weight parameters, and obtain its adaptive fusion weight under radar A and radar B respectively.

[0070] To fully integrate the spatial residuals and uncertainties of trajectory observations, as well as the performance differences of the radar systems themselves, an adaptive fusion weighting function was designed for each pair of matched trajectory points to dynamically determine the contribution of each observation data in the weighted fusion process. By measuring the confidence and reliability of an observation value for a particular trajectory point, a larger weight value indicates a higher proportion of that observation in the fused trajectory. Each pair of matched trajectory points is located on radar X. The specific formula for calculating the fusion weight is as follows: ; in, Let i be the fusion weight of trajectory point i under radar X; Let i be the observation residual of trajectory point i under radar X; Let i be the multi-scale local information entropy of trajectory point i under radar X; Let X be the prior performance weight of radar X; The Sigmoid function maps the result of a linear combination to the interval (0,1), preventing the weights from being infinitely amplified or approaching zero. , The parameters are adjusted to control the sensitivity of the influence of observation residuals and multi-scale local information entropy on the weights.

[0071] When the residual distance of the trajectory point Smaller, information entropy A lower value indicates that the observation is more stable and reliable. When the value tends to be negative, the Sigmoid output is close to 1, and the weights... Larger values ​​indicate a decrease in observation quality or an increase in uncertainty. The Sigmoid output tends towards 0, and the weights automatically decrease, thus weakening their impact on the fusion result. Prior weights... This is used to combine the historical performance differences of different radars (such as the ranging capability's ability to distinguish small targets, and the accuracy performance of electronic scanning / pulse compression in different ranges) for global correction, so that different data sources can be reasonably divided in the fusion process.

[0072] To address the significant performance differences between electronically scanned radar (ESR) and pulse compression radar (PCR) in terms of detection range, resolution, response time, and noise level, this invention dynamically designs a priori weighting function based on historical measurement mean square error (RMSE), residual moving average, and ranging capability. This priori weighting not only reflects the average accuracy of each type of radar under typical conditions but also incorporates performance degradation trends within the target ranging interval (e.g., decreased accuracy of ESR at long ranges) for correction. This allows the weight allocation to be adjusted in real-time according to observation conditions, thereby rationally determining the contribution ratio of different radar data in the fusion process. This mechanism fully leverages the advantages of high-performance radars within suitable ranges, effectively suppressing the negative impact of low-confidence observations on the fusion results and improving the overall trajectory reliability. This fusion weighting can dynamically identify and enhance the contribution of reliable observation points while suppressing noisy points and high-uncertainty observations, effectively improving the overall accuracy and robustness of the fused trajectory and meeting the accuracy requirements of multi-source radar fusion in complex maritime environments.

[0073] Step 2.5. After obtaining the adaptive fusion weights of each pair of matched trajectory points under the two heterogeneous radars, Radar A and Radar B, the final fused trajectory coordinates are calculated by combining the observation coordinates of the two heterogeneous radars and using a weighted average strategy.

[0074] After step 2.4, the adaptive fusion weights of each pair of matched trajectory points i under radar A and radar B are obtained. , Then, a weighted average strategy is used to calculate the final fused trajectory coordinates. The specific process is as follows: Let radar A and radar B be at time... The observed coordinates are respectively , .

[0075] The coordinates of the merged trajectory points The calculation formula is as follows: ; in, , Let be the fusion weights of the i-th trajectory point in radar A and radar B respectively.

[0076] This weighted average formula is essentially a confidence-weighted interpolation, which makes data sources with higher weights contribute more to the fusion result, ensuring that the fused trajectory has higher stability and continuity while preserving observation accuracy. Using this method, all trajectory point pairs can be fused point-by-point to generate a complete, high-precision multi-source radar trajectory sequence.

[0077] The proposed entropy-driven weighting strategy quantitatively reflects the uncertainty of observation data for each pair of matched trajectory points based on local information entropy: the more concentrated the residual distribution and the lower the entropy value, the stronger the observation consistency and the higher the reliability, thus assigning higher weights during fusion; when the entropy value is high, it indicates that the residual distribution of observations around the trajectory point is scattered and fluctuates greatly, with significant uncertainty, thereby automatically reducing its impact on the fusion result. This mechanism originates from the meaning of entropy in information theory (the lower the entropy, the more ordered and reliable the information), and does not rely on fixed thresholds or empirical parameters. By mapping the residuals and entropy to the weight range (0~1) through the Sigmoid function, it can both improve the contribution of reliable observations and dynamically suppress outliers, ensuring the continuity and accuracy of the fused trajectory. This method has low computational complexity, strong robustness, and is easy to embed into real-time multi-radar sensing systems, significantly improving fusion quality and system stability.

[0078] Step 2.6. Repeat steps 2.2 to 2.5 above until the weighted fusion calculation of trajectory points at each time point is completed, and connect all fused trajectory coordinates according to the time series to construct the final fused trajectory sequence.

[0079] After completing the weighted fusion calculation of trajectory points at each time step, all fused points By connecting the time series data, the final fused trajectory sequence is constructed. N represents the number of matching point pairs, where i ∈ [1, N].

[0080] This fused trajectory retains the spatial resolution of both types of radar observation data, and significantly reduces the influence of single radar noise and residuals through multi-scale entropy weighting and dynamic weighting mechanisms, thus possessing higher smoothness and higher accuracy.

[0081] The multi-radar data fusion method based on local information entropy weighting in this embodiment can efficiently convert data from solid-state pulse compression radar and electronically scanned radar, thereby achieving the fusion processing of different types of radar data.

[0082] Of course, in this embodiment, radar A and radar B are not limited to data fusion between solid-state pulse compression radar and electronically scanned array radar. For example, it can also be data fusion between other types of radar, such as data fusion between X-band and S-band radars with different operating frequencies, or data fusion between pulse Doppler radar and continuous wave radar.

[0083] Furthermore, in order to comprehensively evaluate the fusion result quality of the fusion method proposed in this invention, AIS trajectory was selected as the true value reference, and the following multiple indicators were used for error analysis. The calculation formulas for each indicator are as follows: RMSE (Root Mean Square Error): The overall deviation between the quantized fusion trajectory and the true value. The calculation formula is: .

[0084] in, To merge the coordinates of the trajectory points, Here are the AIS ground truth coordinates at the corresponding time, and N is the total number of samples.

[0085] Maximum error: .

[0086] Minimum error: .

[0087] Mean error: .

[0088] Standard deviation: .

[0089] Figure 1 This is a flowchart illustrating the fusion method of solid-state pulse compression radar, electronically scanned radar, and GPS. The diagram shows the overall process of data fusion between various radar systems (solid-state pulse compression radar and electronically scanned radar) and GPS.

[0090] The process is divided into several stages, including raw data acquisition, data preprocessing, trajectory matching, coordinate system transformation, weighted trajectory fusion based on local information entropy, and error assessment and truth value comparison, gradually realizing the calibration and fusion of radar data. The inputs, outputs, and operational steps of each stage are clearly presented, demonstrating the entire process from raw data to the final fusion result.

[0091] Figure 2 This is a preprocessed trajectory map of pulse-pressure radar, electronically scanned radar, and GPS. The image shows the raw trajectory data of the pulse-pressure radar, electronically scanned radar, and GPS before coordinate transformation.

[0092] The blue trajectory line represents the target vessel's trajectory collected by the shipborne GPS system, the yellow trajectory represents the target vessel's trajectory scanned by the pulse compression radar, and the red trajectory line represents the target vessel's trajectory acquired by the electronically scanned radar.

[0093] Figure 3 This is a comparison chart of the converted electronically scanned radar (ESDR) and GPS tracks. The blue track line represents the trajectory of the shipborne GPS system, serving as a reference. The red track line is the ESDR data before conversion; the yellow track line is the ESDR track obtained after coordinate transformation.

[0094] Figure 4This is a comparison chart of the converted solid-state pulse compression radar (SPL) data and the GPS trajectory. The chart shows a comparison between the SPL data before and after conversion and the GPS trajectory. The blue trajectory line represents the trajectory of the shipborne GPS system, used as a reference. The yellow line represents the SPL trajectory before conversion, and the dark blue line represents the SPL trajectory obtained after coordinate transformation.

[0095] Figure 5 This demonstrates the alignment effect between all radar data and GPS tracks after applying the ICP algorithm. Pulse compression radar, electronically scanned radar, and GPS exhibit consistent tracks in a unified coordinate system. The blue track line represents the GPS track as a reference. The red line represents the converted track from electronically scanned radar, and the dark blue line represents the converted track from solid-state pulse compression radar.

[0096] Figure 7 This is a trajectory comparison diagram after fusion using the method of this invention. This diagram shows the comparison effect of all radar data and GPS trajectories after the local information entropy weighted fusion algorithm. The vertical axis represents latitude, and the horizontal axis represents longitude (for ease of display, each scale value on the horizontal axis needs to be superimposed with 120.197 to obtain the actual longitude; the actual longitude on the leftmost side is 120.197). The black trajectory line is the GPS trajectory, serving as a reference after fusion; the red trajectory line is the trajectory after fusion of solid-state pulse compression radar and electronically scanned radar. Figure 7 It can be seen that the trajectory fused by the method of the present invention has a high degree of consistency with the true value of AIS, which indicates that the trajectory fused by the present invention has high accuracy.

[0097] Figure 8 , Figure 9 and Figure 10 The comparison of errors between the fused trajectory and the true AIS value in the longitude and latitude directions, as well as the temporal variation of the total longitude and latitude error, are presented. Specifically, Figure 8 and Figure 9 It intuitively reflects the positioning accuracy of the fused trajectory in the east-west and north-south directions. By comparing it with the AIS ground truth, the performance of the fusion algorithm in different spatial dimensions can be evaluated. Figure 10 This comprehensively demonstrates the dynamic trend of the overall spatial error of the fused trajectory relative to the true value over time, reflecting the temporal stability and accuracy of the fusion process. Statistical indicators of the error show that the root mean square error (RMSE) of the fused trajectory is 27.485 meters, the maximum error is 81.725 meters, the minimum error is 0.188 meters, and the mean error is 24.515 meters. This indicates that the proposed fusion method can effectively reduce target positioning errors and enhance the spatial accuracy and consistency of the trajectory.

[0098] Overall, Figures 8 to 10Error analysis verified that the proposed method has high accuracy and reliability in multi-source radar data fusion, which can meet the requirements of maritime intelligent perception systems for high-precision target tracking. In the figure, the vertical axis represents the error, and the horizontal axis represents the timestamp (a timestamp is a digital sequence that identifies the moment an event occurs. For ease of display, each scale value on the horizontal axis needs to be superimposed with 1721800000 to obtain the actual timestamp, and the leftmost actual timestamp is 1721800000).

[0099] like Figures 11 to 14 The error distribution maps for latitude and longitude distance, speed, heading, and trajectory are shown respectively after applying the local information entropy weighted fusion algorithm. Figure 12 , Figure 13 The displayed speed and heading error diagrams also reflect the error control effect of this invention on different motion parameters, which can improve positioning accuracy and effectively suppress measurement deviations in speed and heading.

[0100] The fusion method of this invention not only enhances the ability of multi-source radar systems to identify, locate and track targets, but also significantly improves perception robustness and navigation safety in complex maritime scenarios, and has good engineering practical value and promotion prospects.

[0101] Example 2 This embodiment 2 describes a heterogeneous radar data fusion system for a maritime radar system. This system is based on the same inventive concept as the heterogeneous radar data fusion method for a maritime radar system in embodiment 1 above.

[0102] The heterogeneous radar data fusion system for marine radar systems in this embodiment 2 includes the following modules: The data calibration module is used to collect and extract raw data from two heterogeneous radars, clean and preprocess the data, perform joint matching based on time and trajectory, and perform coordinate calibration and transformation, using the target ship trajectory obtained by GPS / BeiDou as a spatiotemporal reference benchmark. It converts the raw data of the two heterogeneous radars into a unified GPS / BeiDou coordinate system, thus completing the data calibration before fusion. The data fusion module proposes an adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding windows. It performs probabilistic modeling and local information entropy calculation on the observation residuals within sliding windows of different spatial scales, and adaptively adjusts the fusion weights of trajectory points by combining the prior weights of heterogeneous radar performance, so as to achieve the fusion of heterogeneous radar trajectories.

[0103] It should be noted that in the heterogeneous radar data fusion system for maritime radar systems in this embodiment 2, the implementation process of the functions and roles of each functional module is detailed in the implementation process of the corresponding steps of the method in the above embodiment 1, and will not be repeated here.

[0104] Example 3 This embodiment 3 describes a computer device. The computer device includes a memory and one or more processors. Executable code is stored in the memory. When the processor executes the executable code, it implements the steps of the heterogeneous radar data fusion method for a maritime radar system described in embodiment 1 above.

[0105] In this embodiment, the computer device can be any device or apparatus with data processing capabilities, and will not be described in detail here.

[0106] Example 4 This embodiment 4 describes a computer-readable storage medium storing a program that, when executed by a processor, is used to implement the steps of the heterogeneous radar data fusion method for a maritime radar system described in embodiment 1 above.

[0107] The computer-readable storage medium can be an internal storage unit of any device or apparatus with data processing capabilities, such as a hard disk or memory, or an external storage device of any device with data processing capabilities, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc.

[0108] Of course, the above description is only a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. It should be noted that any equivalent substitutions or obvious modifications made by those skilled in the art under the guidance of this specification fall within the scope of this specification and should be protected by the present invention.

Claims

1. A heterogeneous radar data fusion method for a marine radar system, characterized in that, Comprising the following steps: Step 1. Take the target ship trajectory obtained by GPS / Beidou as the space-time reference datum, collect and extract the original data of two heterogeneous radars, clean and preprocess the data, jointly match based on time and trajectory, coordinate calibration and conversion, and convert the original data of two heterogeneous radars to the unified GPS / Beidou coordinate system to complete the data calibration before fusion; Step 2. An adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding window is proposed, the observation residual is probabilistically modeled and the local information entropy is calculated in the sliding window of different spatial scales, and the fusion weight of the trajectory point is adaptively adjusted by combining the prior weight of the performance of the heterogeneous radar, so as to realize the fusion of the heterogeneous radar trajectory.

2. The heterogeneous radar data fusion method for maritime radar system according to claim 1, wherein, Step 2 is specifically: Step 2.

1. Construct a set of matched trajectory points of radar A and radar B, and calculate the Euclidean distance of each pair of matched trajectory points in the set, i.e. the observation residual; Each pair of matched trajectory points is composed of observation coordinates of radar A and radar B at the same time; Step 2.

2. For each pair of matched trajectory points, the distribution probability of observation residuals in the neighborhood is respectively calculated in the sliding window of different spatial scales, and the local information entropy under the corresponding spatial scale is calculated; Then the local information entropy under each spatial scale is integrated to obtain the multi-scale local information entropy of the trajectory point; Step 2.

3. Prior weight design based on the performance of the two heterogeneous radars; Step 2.

4. For each pair of matched trajectory points, a corresponding adaptive fusion weight function is designed based on the observation residual, the multi-scale local information entropy and the prior weight parameter, and the adaptive fusion weight under radar A and radar B is obtained respectively; Step 2.

5. After obtaining the adaptive fusion weight of each pair of matched trajectory points under radar A and radar B, the final fusion trajectory coordinates are calculated by using the weighted average strategy combined with the observation coordinates of the two heterogeneous radars; Step 2.

6. Repeat steps 2.2 to 2.5 above until the weighted fusion calculation of all matched trajectory points at each time is completed, and connect all the fusion trajectory coordinates in time sequence to construct the final fusion trajectory sequence.

3. The heterogeneous radar data fusion method for maritime radar system according to claim 1, wherein, In step 1, the heterogeneous radars are solid-state pulse compression radars and electronic scanning array radars, or X-band and S-band radars with different operating frequencies, or pulse Doppler radars and continuous wave radars.

4. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, Step 2.1 is specifically: The obtained matching track point pair set in the step 1 coordinate calibration stage is ; wherein is the observation coordinate of radar A at time , is the corresponding matched observation coordinate of radar B at the same time , the trajectory data of radar A and radar B are both unified to the same GPS coordinate system;​​ The Euclidean distance of each pair of matched trajectory points in space is calculated, and the formula is as follows: ； All constituting distance vectors ; where N is the number of matching point pairs, .

5. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, Step 2.2 is specifically: For each spatial scale r m a neighborhood set is defined as follows: ； wherein is the coordinate of a trajectory point ; is the coordinate of a trajectory point at the same time or adjacent time instant as the trajectory point ; is the Euclidean distance between two points is a neighborhood trajectory point set centered at the trajectory point ; The observation residual of each pair of matching track points in the neighborhood, i.e. the Euclidean distance of two radar pairs observing the same target coordinates, is denoted as , is divided into B intervals, and the frequency of each interval is counted , and the residual distribution probability is calculated: ； wherein denotes the probability that the observed residual in the neighborhood falls into the th interval; denotes the frequency of the th interval, i.e. the number of trajectory points; denotes the total number of trajectory points in this neighborhood; Compute the local information entropy under the current spatial scale r m , as follows: ​ ； Finally, the local information entropy of different spatial scales is fused to obtain the multi-scale local information entropy of the trajectory point : ； wherein, to prevent a small positive number from being zero, are weights of different scales.

6. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, the step 2.3 is specifically: Define X∈{A,B} to represent radar A or radar B, then the prior weights , Calculate using the following formula: ； wherein is the radar X historical observation root mean square error; represents a moving average of the fixed length observation residuals centered on the current trajectory point timestamp; is a constant coefficient used to adjust the relative initial weight of the radar; for preventing division by zero constant; for observing residual sensitivity adjustment parameter; for distance correction factor, used to dynamically adjust radar weight according to ranging distance, to correct the observation accuracy decrease caused by distance increase.

7. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, the step 2.4 is specifically: a fusion weight of each pair of matched track points under radar X, The specific calculation formula is as follows: ； where, is the fusion weight of the trajectory point i under radar X; is the observation residual of the trajectory point i under radar X; is the multi-scale local information entropy of the trajectory point i under radar X; is the prior performance weight of radar X; is the Sigmoid function; , are the adjustment parameters, respectively used to control the influence sensitivity of the observation residual and the multi-scale local information entropy on the weight.

8. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, the step 2.5 is specifically: The adaptive fusion weight of each pair of matched track points i at radar A and radar B is obtained after step 2.4 、 After that, the final fusion track coordinates are calculated by using the weighted average strategy, and the specific process is as follows: Let radar A and radar B be at observation coordinates , , at time t = t0 The coordinates of the fused trajectory points The calculation formula is as follows: ； wherein, , are the fusion weights of the i-th trajectory point in radar A and radar B, respectively.

9. The heterogeneous radar data fusion method for maritime radar system according to claim 2, wherein, After the weighted fusion of the trajectory points at each time is completed, all the fusion points are connected in time sequence to construct the final fusion trajectory sequence After the weighted fusion of the trajectory points at each time is completed, all the fusion points are connected in time sequence to construct the final fusion trajectory sequence ; N is the number of matched point pairs, i∈[1,N].

10. A heterogeneous radar data fusion system for a marine radar system, characterized by comprises the following modules: a data calibration module, which is used to take the target ship trajectory obtained by GPS / Beidou as the space-time reference datum, to collect and extract the original data of two heterogeneous radars, to clean and pretreat the data, to jointly match based on time and trajectory, to calibrate and convert coordinates, to convert the original data of two heterogeneous radars to a unified GPS / Beidou coordinate system, and to complete the data calibration before fusion; a data fusion module, which is used to propose an adaptive weighted trajectory fusion method based on improved local information entropy and multi-scale sliding window, to perform probability modeling and local information entropy calculation on observation residuals in different spatial scale sliding windows, and to combine the prior weight of the performance of the heterogeneous radars to adaptively adjust the fusion weight of the trajectory points, so as to realize the fusion of the heterogeneous radar trajectories.

Images