Data processing method, device, equipment, medium and vehicle

Abstract

The present invention provides a data processing method, device, equipment, medium, and vehicle, relating to the field of automotive environmental perception technology. The method includes: obtaining the current vehicle network data and multiple vehicle-mounted sensor data of the vehicle; performing a credibility assessment on the current vehicle network data to obtain a credibility score for the current vehicle network data, and performing a credibility assessment on the multiple vehicle-mounted sensor data to obtain a credibility score for each of the multiple vehicle-mounted sensor data; and fusing the current vehicle network data and the multiple vehicle-mounted sensor data using the respective credibility scores of the current vehicle network data and the multiple vehicle-mounted sensor data as fusion weights to obtain fused data. The present invention can achieve collaborative perception fusion of V2X and vehicle-mounted sensors based on dynamically adjusted fusion weights, thereby enhancing the environmental adaptability of V2X collaborative perception and improving the detection accuracy of the driving environment around the vehicle.

Description

Technical Field

[0001] The present invention relates to the field of automobile environment perception technology, and in particular to a data processing method, device, equipment, medium and vehicle. Background Art

[0002] With the accelerating trend toward intelligent and connected vehicles in the automotive industry, the integration of V2X (Vehicle to Everything) technology and single-vehicle intelligence has become an irreversible trend. V2X-based collaborative perception can overcome the limitations of single-vehicle perception, enhance safety redundancy, and optimize group decision-making, making it a key technology for achieving high-level autonomous driving. However, current collaborative perception technologies based on V2X and on-board sensors rely on preset fixed weights for collaborative perception fusion. This results in technical flaws such as poor environmental adaptability and low detection accuracy of the vehicle's surrounding driving environment. Summary of the Invention

[0003] The present invention provides a data processing method, apparatus, device, medium and vehicle, aiming to overcome the above-mentioned problems or at least partially solve the above-mentioned problems.

[0004] A first aspect of the present invention provides a data processing method, comprising:

[0005] Obtaining current vehicle networking data and multiple vehicle-mounted sensor data of the vehicle, wherein the current vehicle networking data represents detection results of other vehicles around the vehicle on the driving environment around the vehicle, and the multiple vehicle-mounted sensor data represents detection results of multiple vehicle-mounted sensors of the vehicle on the driving environment around the vehicle;

[0006] Performing a credibility evaluation on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data, and performing a credibility evaluation on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data;

[0007] The current Internet of Vehicles data and the multiple vehicle-mounted sensor data are fused using their respective credibility scores as fusion weights to obtain fused data, which represents the final detection result of the driving environment around the vehicle.

[0008] A second aspect of the present invention provides a data processing device, comprising:

[0009] A data acquisition module is used to obtain the current vehicle network data and various vehicle sensor data of the vehicle, wherein the current vehicle network data represents the detection results of other vehicles around the vehicle on the driving environment around the vehicle, and the various vehicle sensor data represents the detection results of various vehicle sensors of the vehicle on the driving environment around the vehicle;

[0010] a scoring determination module, configured to perform a credibility assessment on the current IoV data to obtain a credibility score for the current IoV data, and to perform a credibility assessment on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data;

[0011] The data fusion module is used to fuse the current Internet of Vehicles data and the multiple vehicle-mounted sensor data using their respective credibility scores as fusion weights to obtain fused data, where the fused data represents the final detection result of the driving environment around the vehicle.

[0012] The third aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the computer program is executed by the processor, it implements the data processing method of the first aspect of the present invention.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the data processing method of the first aspect of the present invention is implemented.

[0014] A fifth aspect of the present invention provides a vehicle comprising at least a controller, which, when executed, implements the data processing method of the first aspect of the present invention.

[0015] In the data processing method provided by the present invention, the credibility of the vehicle's current IoV data and various onboard sensor data are separately assessed, resulting in a credibility score for each. These credibility scores are then used as fusion weights to fuse the current IoV data and the various onboard sensor data, yielding fused data that represents the final detection results of the vehicle's surrounding driving environment. In this way, the present invention performs a real-time quantitative assessment of the credibility scores of the vehicle's multimodal perception data and, based on dynamically adjusted fusion weights, achieves collaborative perception fusion between V2X and onboard sensors, thereby enhancing the environmental adaptability of V2X collaborative perception and improving the accuracy of detection of the vehicle's surrounding driving environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0017] Figure 1 is a flow chart of a data processing method according to an embodiment of the present invention;

[0018] Figure 2 This is an overall system architecture diagram of a multimodal data fusion system based on dynamic credibility evaluation according to an embodiment of the present invention;

[0019] Figure 3 This is a flow chart of a three-level conflict arbitration mechanism according to an embodiment of the present invention;

[0020] Figure 4 This is a diagram of an optimized architecture of a federated learning model according to an embodiment of the present invention;

[0021] Figure 5 is a structural block diagram of a data processing device provided by an embodiment of the present invention;

[0022] Figure 6 FIG. 1 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0024] In one embodiment, reference Figure 1 , Figure 1 FIG. 1 is a flow chart showing a data processing method according to an embodiment of the present invention. Figure 1 As shown, the data processing method of this embodiment may include the following steps:

[0025] Step S11: Obtain the current Internet of Vehicles data and various vehicle-mounted sensor data of the vehicle.

[0026] In this embodiment, the current vehicle-to-everything (V2X) data of the vehicle (the vehicle itself) can be obtained, as well as various on-board sensor data of the vehicle. The current V2X data refers to the current V2X data (V2X data). The V2X data includes at least one or more of the following: the position, speed, and heading angle of surrounding vehicles carried in a BSM (Basic Safety Message), a CAM (Cooperative Awareness Message), or a DENM (Decentralized Environmental Notification Message). The current V2X data represents the detection results of the driving environment surrounding the vehicle by other vehicles around the vehicle. The various on-board sensor data refers to sensor data collected in real time by the various on-board sensors of the vehicle, representing the detection results of the various on-board sensors of the vehicle regarding the driving environment surrounding the vehicle.

[0027] Step S12: performing a credibility evaluation on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data, and performing a credibility evaluation on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data.

[0028] In this embodiment, independent credibility assessments can be performed on the current Internet of Vehicles data and various on-board sensor data of the vehicle to obtain the credibility score of the current Internet of Vehicles data and the credibility scores of various on-board sensor data, which can be used as the basis for subsequent weight distribution. Among them, the credibility score of this embodiment characterizes the reliability of the current Internet of Vehicles data and / or various on-board sensor data, and is a key indicator for quantifying the reliability of the current Internet of Vehicles data and / or various on-board sensor data. It is used to evaluate the confidence level of communication messages with surrounding vehicles and / or the on-board sensor measurements of this vehicle. In this way, this embodiment establishes a cross-modal, multi-factor fusion credibility quantification system, breaking through the limitations of traditional single indicator evaluation.

[0029] Step S13: using the respective credibility scores of the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data as fusion weights, fusing the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data to obtain fused data.

[0030] In this embodiment, after obtaining the credibility score of the current Internet of Vehicles data and the credibility scores of the various types of vehicle-mounted sensor data, the current Internet of Vehicles data and the various types of vehicle-mounted sensor data can be respectively assigned corresponding fusion weights based on the respective credibility scores of the current Internet of Vehicles data and the various types of vehicle-mounted sensor data. Based on the respective corresponding fusion weights of the current Internet of Vehicles data and the various types of vehicle-mounted sensor data, the current Internet of Vehicles data and the various types of vehicle-mounted sensor data are fused to obtain fused data, which represents the final detection result of the driving environment around the vehicle.

[0031] In this embodiment, by performing real-time quantitative evaluation on the credibility score of the vehicle's multimodal perception data (current Internet of Vehicles data and data from multiple on-board sensors), and fusing the multimodal perception data based on dynamically adjusted fusion weights, collaborative perception fusion of V2X and on-board sensors is achieved, thereby enhancing the environmental adaptability of V2X collaborative perception and improving the detection accuracy of the driving environment around the vehicle.

[0032] In combination with the above embodiments, in one implementation, the above step S11 may specifically include three aspects: data access, data preprocessing and standardization, and initial data quality screening.

[0033] Among them, data access: for the current Internet of Vehicles data of this vehicle, it can support DSRC / C-V2X dual-mode communication, and parse the structured data in the obtained BSM, CAM, DENM and other messages, including vehicle position, speed, heading angle and other information; at the same time, it extracts the communication quality parameters between this vehicle and the Internet of Vehicles in real time, such as signal strength (RSSI), signal-to-noise ratio (SINR) and packet loss rate; for on-board sensor data: when the on-board sensor data includes an on-board camera, the original image and target detection results (bounding box + classification confidence) collected by the on-board camera can be obtained; when the on-board sensor data includes millimeter-wave radar, the millimeter-wave radar point cloud and target tracking list can be obtained; when the on-board sensor data includes lidar, 3D point cloud data can be obtained.

[0034] Data preprocessing and standardization: Format unification: For the vehicle's current Internet of Vehicles data, convert it into a standard JSON format (including timestamp, coordinate system, and data source identifier). For onboard sensor data, convert it into the Autoware universal interface format, for example, point cloud to PCD, image to ROS Image, etc. Data is temporally and spatially aligned: The PTP protocol (e.g., accuracy <1μs) is used to align the time bases of various data sources for time synchronization. A calibration matrix is ​​used to align the coordinate systems of various onboard sensors to the vehicle coordinate system (conversion error <5cm) for coordinate conversion.

[0035] Initial data quality screening: Detect anomalies in the data, such as identifying illegal values ​​in the current Internet of Vehicles data (such as speed > 200 km / h) and identifying jumps in on-board sensor data (such as sudden acceleration > 10 m / s²). Priority marking is also performed, such as marking emergency events (such as accident warnings in the current Internet of Vehicles data) with the highest priority (Level 0).

[0036] In conjunction with the above embodiments, in one embodiment, the present invention further provides a data processing method. In this method, the "performing a credibility assessment on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data" in step S12 may specifically include steps S21 to S24:

[0037] Step S21: obtaining a communication quality parameter, and obtaining a communication quality score of the current Internet of Vehicles data based on the communication quality parameter.

[0038] In this embodiment, the communication quality parameters between the vehicle and the Internet of Vehicles are obtained, and the communication quality score of the current Internet of Vehicles data can be determined based on the communication quality parameters. The communication quality score represents the communication quality of the transmission link of the current Internet of Vehicles data. The higher the communication quality score, the more reliable the transmission link.

[0039] In an optional example, the communication quality parameter includes a signal-to-interference-and-noise ratio (SINR), and the communication quality score of the current Internet of Vehicles data can be determined by the following formula: , Sigmoid() is a Sigmoid function. The signal-to-interference-plus-noise ratio (SINR) reflects the reliability of the current IoV data transmission link. The SINR value can be extracted in real time by physical layer chips (such as C-V2X OBU). In practical applications, the SINR range for IoV data communication typically ranges from -10dB to 30dB, with the specific value depending on the communication environment and application scenario. A Sigmoid function is used to nonlinearly map the SINR (range -10-30dB) to [0,1]. For example, an inflection point is set at the critical value of 10dB (it is generally believed that when SINR is less than 10dB, communication quality is poor and a connection may not be established or maintained. The specific inflection point value can be dynamically adjusted based on the actual communication environment). When SINR is less than 10dB, the weight decays exponentially to reflect the dynamic impact of changes in communication quality on the credibility of V2X messages.

[0040] Step S22: Obtain the time difference between the sending time and the receiving time of the current Internet of Vehicles data, and obtain the time decay score of the current Internet of Vehicles data based on the time difference and the decay rate.

[0041] In this embodiment, the sending time and receiving time of the current Internet of Vehicles data can be determined, thereby determining the time difference between the sending time and the receiving time of the current Internet of Vehicles data, and then based on the time difference and the preset decay rate (for example, different decay rates can be preset based on different scenarios), the time decay score of the current Internet of Vehicles data is determined. The time decay score represents the decay of the "freshness" of the current Internet of Vehicles data over time to avoid interference from expired data.

[0042] In an optional example, the time decay score of the current Internet of Vehicles data can be determined by the following formula: . Δt is the time difference between the time the current IoV data is sent and the time it is received (in seconds). The decay rate is controlled by λ, which is a control parameter. In a typical scenario, λ can be set to 0.3 (Δt = 2 seconds → e^{-0.6}≈0.55), and the credibility weight of the message received 2 seconds ago is reduced by 45%. For special scenarios, the value of λ can be dynamically adjusted to reflect the timeliness requirements of the current IoV data in special scenarios. For example, in a high-speed scenario, low latency is required for high-speed vehicle movement, so λ can be set to 0.5 (faster decay); in a congested scenario, vehicles move slowly, so λ can be set to 0.2 (slower decay).

[0043] Step S23: performing consistency analysis on the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and obtaining a consistency score for the current Internet of Vehicles data based on the result of the consistency analysis.

[0044] In this embodiment, a consistency analysis can be performed on the current Internet of Vehicles data using a variety of vehicle-mounted sensor data to obtain a consistency analysis result, and a consistency score of the current Internet of Vehicles data can be determined based on the consistency analysis result.

[0045] In an optional example, local vehicle-mounted sensor data (for example, point cloud data collected by a lidar) can be used to verify the historical consistency of the current Internet of Vehicles data. For example, consistency analysis can be performed on three consecutive frames of Internet of Vehicles data and its own vehicle-mounted sensor data (for example, when the mean vehicle position error is less than 1.5 meters, the consistency score can be considered to be H=0.9).

[0046] For example, the consistency score of the current Internet of Vehicles data is determined by the following formula: : .

[0047] Where N is the number of statistical frames (usually N=3), δ is the error threshold (δ=1.5 meters), Θ is the indicator function (1 when the error ≤ δ, otherwise 0); k represents the kth frame data; To obtain the target location information through the Internet of Vehicles, For the same target position information obtained by the vehicle-mounted sensor of the own vehicle, the specific mathematical expression can be the (x, y) two-dimensional coordinates of the target in the vehicle coordinate system; The specific calculation method is to calculate the two-dimensional space Euclidean distance error between two position points of the same target obtained by the Internet of Vehicles and the local sensor (the vehicle's onboard sensor). Assume The corresponding position coordinates are (x1, y1), The corresponding position coordinates are (x2, y2), and the calculation formula of the above two-dimensional space Euclidean distance error can be explicitly transformed into: If the error of three consecutive frames of data is less than 1.5 meters, then The value of is 1.0 (indicating complete credibility); if two frames of data meet the standard, then The value of is 0.67.

[0048] Step S24: Obtaining a credibility score of the current Internet of Vehicles data based on the communication quality score, time decay score, and consistency score of the current Internet of Vehicles data.

[0049] In this embodiment, the credibility of the current Internet of Vehicles data can be evaluated based on the communication quality score of the current Internet of Vehicles data, the time attenuation score of the current Internet of Vehicles data, and the consistency score of the current Internet of Vehicles data to determine the credibility score of the current Internet of Vehicles data.

[0050] In an optional embodiment, in order to quantify the dynamic credibility of the current Internet of Vehicles data (such as a continuous value from 0 to 1) to accurately reflect its reliability in a specific scenario, different weights can be assigned to the communication quality score of the current Internet of Vehicles data, the time decay score of the current Internet of Vehicles data, and the consistency score of the current Internet of Vehicles data, so as to perform a weighted sum of the communication quality score of the current Internet of Vehicles data, the time decay score of the current Internet of Vehicles data, and the consistency score of the current Internet of Vehicles data to obtain the credibility score of the current Internet of Vehicles data. , as shown below:

[0051] ;

[0052] Here, α, β, and γ are the weights corresponding to the communication quality score, time decay score, and consistency score of the current IoV data, respectively. α, β, and γ ∈ [0, 1], and α + β + γ = 1. The weight coefficients α, β, and γ can be initially configured based on experience and the application scenario. For example, in highway scenarios, timeliness is critical, so β ​​should be increased; in congested scenarios, communication stability is more critical, so α is prioritized. In vehicles with multiple sensors, γ can be set to [0.3, 0.5], while in vehicles with fewer sensors, γ can be reduced to [0.1, 0.2]. For example, for highway scenarios, the following weight coefficients are typically preset: α = 0.5, β = 0.3, and γ = 0.2. While appropriately increasing the time decay weight β, focusing on monitoring communication quality, α = 0.5 is set. For example, in urban intersection collaboration scenarios, the historical consistency of multi-vehicle trajectories should be emphasized, so α = 0.3, β = 0.3, and γ = 0.5 are recommended. After setting initial values ​​for the three weighting systems of communication quality score, time attenuation score, and historical consistency score based on expert experience and application scenario characteristics, the final weights α, β, and γ can be further obtained through actual vehicle roadside calibration.

[0053] It should be noted that this embodiment does not limit the execution order of the above steps S21 to S23. For example, the execution order can be arbitrarily specified, or they can be executed simultaneously.

[0054] In combination with the above embodiments, in one implementation, a dynamic credibility scoring mechanism is proposed for vehicle-mounted sensors (such as vehicle-mounted cameras, vehicle-mounted millimeter-wave radars, and vehicle-mounted lidars). A multidimensional quantitative model is constructed by combining physical properties and environmental interference to ensure the accuracy and robustness of weight allocation during multimodal fusion.

[0055] In conjunction with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. In this method, the multiple types of vehicle-mounted sensor data include at least image data collected by a vehicle-mounted camera; and the "performing credibility assessments on each of the multiple types of vehicle-mounted sensor data to obtain respective credibility scores for the multiple types of vehicle-mounted sensor data" in step S12 may include at least steps S31 to S34:

[0056] Step S31: performing structural similarity analysis on the image data collected by the vehicle-mounted camera and the reference image data to obtain a structural similarity score of the vehicle-mounted camera.

[0057] The credibility assessment of image data captured by an on-board camera requires a comprehensive consideration of image quality and environmental interference. Based on this, this embodiment performs a structural similarity analysis on the image data captured by the on-board camera and reference image data, quantifying the degree of image distortion and generating a structural similarity score for the on-board camera. The reference image data can be reference background image data, which can be updated using different mechanisms depending on weather conditions. For example, in sunny scenes, the reference image data is updated every five minutes; in foggy and rainy scenes, the reference image data is updated by dynamically removing interference areas through motion target segmentation (e.g., optical flow).

[0058] In an optional embodiment, the structural similarity score of the vehicle-mounted camera can be determined by the following formula:

[0059] ;

[0060] in, is the mean value of the local window of the image, reflecting the brightness; is the variance, reflecting the contrast; is the covariance, reflecting structural similarity; is a constant to prevent the denominator from being zero. Let x be the reference image data Let y be the structural similarity score of the vehicle camera. .

[0061] Furthermore, for the constant and Based on the classic SSIM algorithm constant settings, an adaptive adjustment mechanism for vehicle scenes based on lighting and motion status can be added to achieve physical consistency in credibility assessment. In a specific example: =(k1 * L)^2, =(k2 * L)^2;

[0062] Where L is the dynamic range of pixel value (car cameras are usually 8-bit, L=255); k1 and k2 are empirical coefficients (default k1=0.01, k2=0.03).

[0063] Vehicle scene adaptive adjustment strategy: Optimize constants based on the vehicle's specific optical environment. For example, in tunnels and nighttime scenes, to take into account sudden changes in illumination, increase c1 (adjust k1=0.02); in rainy and foggy weather with low contrast, increase c2 (e.g., k2=0.05); in strong glare environments with local overexposure, reduce c2 (e.g., k2=0.02).

[0064] Step S32: performing edge detection on the image data collected by the vehicle-mounted camera, and obtaining a clarity score of the vehicle-mounted camera according to the edge detection result.

[0065] This embodiment also quantifies the image focus quality of the image data collected by the vehicle-mounted camera to prevent motion blur from affecting the detection results. Based on this, this embodiment performs edge detection on the image data collected by the vehicle-mounted camera to obtain edge detection results, and thus obtains a clarity score of the vehicle-mounted camera based on the edge detection results.

[0066] In an optional embodiment, considering that the Laplacian operator is widely used in edge detection and image enhancement in image processing, it detects edges by calculating the square of the gradient of each pixel in the image. Therefore, the clarity score of the vehicle camera can be determined by the following formula: :

[0067] ;

[0068] in, is the image data collected by the vehicle camera; Laplacian(): Laplacian operator, calculates the second-order derivative of the image, and the edge response of the clear image is higher; τ is the normalization threshold: the empirical value τ=120 (if the Laplacian variance>120, ≈1; if the Laplacian variance is <50, ≤0.4).

[0069] Step S33: Obtain the ambient light brightness of the driving environment around the vehicle, and obtain the lighting score of the vehicle camera based on the ambient light brightness and the light on state of the vehicle.

[0070] In this embodiment, the ambient light brightness of the driving environment around the vehicle can be obtained, and the lighting score of the vehicle camera can be obtained based on the ambient light brightness and the status of the headlights, thereby adaptively adjusting the credibility weight of the vehicle camera according to the ambient brightness.

[0071] In an optional embodiment, the illumination score of the vehicle camera can be determined by the following calculation formula: :

[0072] ;

[0073] Among them, the current illumination and illumination are the ambient light brightness of the driving environment around the vehicle.

[0074] Step S34: obtaining a credibility score of the image data collected by the vehicle-mounted camera based on the structural similarity score, the clarity score, and the illumination score of the vehicle-mounted camera.

[0075] In this embodiment, the credibility of the image data collected by the vehicle-mounted camera can be evaluated based on the structural similarity score of the vehicle-mounted camera, the clarity score of the vehicle-mounted camera, and the lighting score of the vehicle-mounted camera to determine the credibility score of the image data collected by the vehicle-mounted camera.

[0076] In an optional embodiment, the credibility score of the image data collected by the vehicle camera can be obtained by the following formula: , as shown below:

[0077] ;

[0078] in, Score the structural similarity of the car cameras, Score the lighting of the car camera. Rate the clarity of the car's camera.

[0079] It should be noted that this embodiment does not limit the execution order of the above steps S31 to S33. For example, the execution order can be arbitrarily specified, or they can be executed simultaneously.

[0080] In conjunction with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. In this method, the multiple types of vehicle-mounted sensor data include at least current point cloud data collected by the vehicle-mounted millimeter-wave radar; and the "performing credibility assessments on the multiple types of vehicle-mounted sensor data to obtain credibility scores for each of the multiple types of vehicle-mounted sensor data" in step S12 above may include at least steps S41 to S45:

[0081] Step S41: Obtaining a point cloud density score of the on-board millimeter-wave radar according to the number of target point clouds represented by the current point cloud data collected by the on-board millimeter-wave radar and the average number of point clouds represented by multiple first historical point cloud data.

[0082] In this embodiment, considering that vehicle-mounted millimeter-wave radar detects targets by emitting and receiving high-frequency electromagnetic waves, the credibility assessment of the vehicle-mounted millimeter-wave radar focuses on signal quality, environmental interference, and consistency with physical laws. Specifically, the completeness of target detection can be assessed based on the number of target point clouds represented in the current point cloud data collected by the vehicle-mounted millimeter-wave radar and the average number of point clouds represented by multiple first historical point cloud data sets, thereby obtaining a point cloud density score for the vehicle-mounted millimeter-wave radar.

[0083] In an optional embodiment, the integrity of target detection can be evaluated by counting the number of target point clouds detected in the current frame collected by the on-board millimeter-wave radar (i.e., the number of target point clouds represented in the current point cloud data collected by the on-board millimeter-wave radar) and the ratio of the historical average point cloud number of the sliding window (i.e., the average point cloud number represented by the first historical point cloud data). Low density may be due to occlusion or reduced detection capability. The point cloud density score of the on-board millimeter-wave radar is The specific formula is as follows:

[0084] ;in, ,If a sudden increase in point cloud is detected, such as the ,sudden increase in point cloud caused by multipath interference, ,anomaly detection should be triggered.

[0085] Step S42: Obtaining a velocity consistency score of the on-board millimeter-wave radar based on the Doppler velocity variance of the current point cloud data and the plurality of first historical point cloud data of the same target collected by the on-board millimeter-wave radar.

[0086] In this embodiment, the speed consistency score of the on-board millimeter-wave radar can be calculated based on the Doppler speed variance of the current point cloud data and multiple first historical point cloud data of the same target collected by the on-board millimeter-wave radar.

[0087] In an optional embodiment, the Doppler velocity measurement and kinematic law verification data rationality are combined to calculate the speed consistency score of the vehicle-mounted millimeter wave radar. , as shown in the following formula:

[0088] ;

[0089] in, , is the Doppler velocity variance, which is obtained by counting the variance of the Doppler velocity of the same target in the current frame; β is an empirical value, which can be taken as 0.1 here to support sharp weight reduction under high variance.

[0090] Step S43: Obtain the millimeter-wave radar signal strength, and obtain a signal-to-noise ratio score of the vehicle-mounted millimeter-wave radar based on a magnitude relationship between the millimeter-wave radar signal strength and a millimeter-wave radar signal strength threshold.

[0091] In this embodiment, the millimeter-wave radar signal strength of the vehicle-mounted millimeter-wave radar can also be obtained to obtain the signal-to-noise ratio score of the vehicle-mounted millimeter-wave radar based on the size relationship between the millimeter-wave radar signal strength and a preset millimeter-wave radar signal strength threshold.

[0092] In an optional embodiment, the credibility can be dynamically adjusted based on the radar signal strength. Specifically, the signal-to-noise ratio score of the vehicle-mounted millimeter-wave radar It can be calculated by the following formula:

[0093] ;

[0094] in, ; is the millimeter wave radar signal strength, That is, the preset millimeter wave radar signal strength threshold can be set to 15dB under typical circumstances. When the signal level reaches or exceeds 30dB (i.e. strong signal state), The value is 1; when When it is lower than 20dB (weak signal state), The value is less than 0.5.

[0095] Step S44: obtaining a multipath suppression score of the on-board millimeter-wave radar according to the proportion of invalid point clouds in the current point cloud data collected by the on-board millimeter-wave radar.

[0096] In this embodiment, the invalid point clouds in the current point cloud data collected by the vehicle-mounted millimeter-wave radar can be determined, and the proportion of invalid point clouds in the current point cloud data collected by the vehicle-mounted millimeter-wave radar can be determined, so that the multipath suppression score of the vehicle-mounted millimeter-wave radar can be obtained based on the proportion of invalid point clouds in the current point cloud data collected by the vehicle-mounted millimeter-wave radar.

[0097] In an optional embodiment, false echoes can be identified through spatial topological analysis. The algorithm strategy is as follows: after constructing the three-dimensional distribution map of the point cloud of the current frame, if it is observed that the point cloud distribution in a certain area is mirror-symmetrical with the surrounding environment (for example, due to the reflection effect of the tunnel wall), the area is marked as a multipath interference area, the point cloud in the area is determined to be invalid, and the weight of the corresponding invalid point cloud is suppressed accordingly. Specifically, the multipath suppression score of the vehicle-mounted millimeter-wave radar is It can be calculated by the following formula:

[0098] ;

[0099] in, is the proportion of invalid point cloud, and the calculation formula is as follows:

[0100] ;

[0101] in, is the total number of point clouds detected in the current frame, is the number of “invalid point clouds generated by multipath interference”; wherein, this embodiment does not limit the specific determination method of “invalid point clouds generated by multipath interference”, and any determination method of “invalid point clouds generated by multipath interference” can implement the content of this embodiment.

[0102] The above multipath mitigation score The calculation formula describes the inverse relationship between interference intensity and credibility in an exponential form. The coefficient 3 is an empirical value used to adjust the attenuation rate and control the sensitivity of credibility when invalid point clouds increase. Through the exponential attenuation mechanism, while ensuring physical consistency, the value is limited to: , achieving quantifiable suppression of multipath interference.

[0103] Step S45: Based on the point cloud density score, speed consistency score, signal-to-noise ratio score, and multipath suppression score of the on-board millimeter-wave radar, a credibility score of the current point cloud data collected by the on-board millimeter-wave radar is obtained.

[0104] In this embodiment, the credibility of the current point cloud data collected by the on-board millimeter-wave radar can be evaluated based on the point cloud density score of the on-board millimeter-wave radar, the speed consistency score of the on-board millimeter-wave radar, the signal-to-noise ratio score of the on-board millimeter-wave radar, and the multipath suppression score of the on-board millimeter-wave radar, and the credibility score of the current point cloud data collected by the on-board millimeter-wave radar can be determined.

[0105] In an optional embodiment, the credibility score of the current point cloud data collected by the vehicle-mounted millimeter wave radar can be obtained by the following formula: , as shown below:

[0106] ;

[0107] in, Score the point cloud density of the vehicle-mounted millimeter-wave radar, Scoring the speed consistency of vehicle-mounted millimeter-wave radars, Score the signal-to-noise ratio of automotive mmWave radars, and Scoring the multipath suppression of automotive millimeter-wave radar.

[0108] It should be noted that this embodiment does not limit the execution order of the above steps S41 to S44. For example, the execution order can be arbitrarily specified, or they can be executed simultaneously.

[0109] In conjunction with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. In this method, the multiple types of vehicle-mounted sensor data include at least current point cloud data collected by a vehicle-mounted laser radar; and the "performing credibility assessments on each of the multiple types of vehicle-mounted sensor data to obtain credibility scores for each of the multiple types of vehicle-mounted sensor data" in step S12 above may include at least steps S51 to S54:

[0110] Step S51: Obtain a reflectivity score of the vehicle-mounted laser radar based on the target reflectivity and reflectivity threshold of the current point cloud data collected by the vehicle-mounted laser radar.

[0111] Considering that the accuracy of vehicle-mounted LiDAR is significantly affected by weather and target reflectivity, this embodiment requires a comprehensive consideration of the dynamic environment and target characteristics when evaluating the reliability of the vehicle-mounted LiDAR. In this embodiment, the reflectivity score of the vehicle-mounted LiDAR can be determined based on the target reflectivity of the current point cloud data collected by the vehicle-mounted LiDAR and a preset reflectivity threshold.

[0112] In an optional embodiment, the reflectivity score of the vehicle-mounted laser radar can be obtained by the following formula: , to correct the problem of missed detection of low reflectivity targets (such as black vehicles):

[0113] ;

[0114] in, , where the normalized reflectivity threshold is the preset reflectivity threshold: for example, the average reflectivity of a white vehicle body is set to 0.8; for example, when a target with a reflectivity of less than 0.2 (dark object) is detected, = 0.5, and compensate for missed targets by correlating camera data.

[0115] Step S52: Obtain visibility and rainfall intensity of the driving environment around the vehicle, and obtain the environmental attenuation score of the vehicle-mounted laser radar.

[0116] In this embodiment, the visibility and rainfall intensity of the driving environment around the vehicle can be determined, and based on the visibility and rainfall intensity of the driving environment around the vehicle, the environmental attenuation score of the vehicle-mounted laser radar can be determined. In an optional embodiment, the environmental attenuation score of the vehicle-mounted laser radar can be determined by the following formula: :

[0117] ;

[0118] in, The visibility unit is m, which can be obtained through the inversion algorithm of the lidar controller, or real-time visibility can be obtained through other means. The rainfall intensity unit is mm / h, which can be obtained through dedicated sensors or other means. k is an empirical value coefficient, which is preset. In typical cases, k = 0.05. The smaller the k value, the smoother the attenuation response. In actual applications, it can be calibrated according to the performance differences of different lidar hardware.

[0119] Step S53: Obtain a trajectory continuity score of the vehicle-mounted laser radar based on the target position detected by the current point cloud data collected by the vehicle-mounted laser radar and the target position predicted based on the historical trajectory of the target in multiple second historical point cloud data.

[0120] In this embodiment, the credibility of the target trajectory can be verified through the spatiotemporal continuity of the LiDAR to avoid false detections caused by sudden noise or interference from dynamic obstacles. Based on this, the target position detected by the current point cloud data collected by the on-board LiDAR can be determined, and the target position can be predicted based on the historical trajectory of the same target in multiple second historical point cloud data. The trajectory continuity score of the on-board LiDAR is thus determined based on the target position detected by the current point cloud data collected by the on-board LiDAR and the target position predicted based on the historical trajectory of the same target in multiple second historical point cloud data.

[0121] In an optional embodiment, the trajectory continuity score of the vehicle-mounted laser radar can be determined by the following formula: :

[0122] ;

[0123] in, ; is the predicted position based on the historical trajectory (using extended Kalman filter dynamic modeling, taking into account acceleration and yaw rate), that is, the target position predicted based on the historical trajectory of the target in multiple second historical point cloud data; The target position detected in the current frame, that is, the target position of the same target detected in the current point cloud data collected by the vehicle-mounted lidar; is the predicted speed based on historical trajectory; The detected target velocity for the current frame. =0.8m: Position tolerance threshold (e.g., allowable positioning error in urban road scenarios); =2m / s: Speed ​​tolerance threshold (based on target type: vehicle / pedestrian).

[0124] In an optional embodiment, it can be based on and Set up anomaly detection mechanism when predicting (Right now ) and detection (Right now )deviation: When the out-of-sequence correction is triggered: If the target is static (speed variance < 0.1m / s), increase The value is 1.5m, allowing for slight positioning jitter; if the target is a turning vehicle (such as when the turn signal is activated), relax the lateral threshold to 2 .

[0125] Step S54: Based on the reflectivity score, environmental attenuation score, and trajectory continuity score of the vehicle-mounted laser radar, a credibility score of the current point cloud data collected by the vehicle-mounted laser radar is obtained.

[0126] In this embodiment, the credibility of the current point cloud data collected by the vehicle-mounted laser radar can be evaluated based on the reflectivity score of the vehicle-mounted laser radar, the environmental attenuation score of the vehicle-mounted laser radar, and the trajectory continuity score of the vehicle-mounted laser radar to determine the credibility score of the current point cloud data collected by the vehicle-mounted laser radar.

[0127] In an optional embodiment, the credibility score of the current point cloud data collected by the vehicle-mounted laser radar can be obtained by the following formula: , as shown below:

[0128] ;

[0129] in, Score the reflectivity of the vehicle-mounted lidar, Score the environmental attenuation of the vehicle-mounted lidar, and Scoring the trajectory continuity of vehicle-mounted lidar.

[0130] It should be noted that this embodiment does not limit the execution order of the above steps S51 to S53. For example, the execution order can be arbitrarily specified, or they can be executed simultaneously.

[0131] In combination with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. In this method, after the above step S11, steps S61 to S62 may be further included, and the above step S13 may specifically include step S63:

[0132] Step S61: Detect whether there is a conflict between the detection results represented by the current Internet of Vehicles data and the detection results represented by the multiple types of vehicle-mounted sensor data.

[0133] In this embodiment, after obtaining the current IoV data of the vehicle and multiple types of onboard sensor data of the vehicle, it is possible to detect whether there are conflicts between the current IoV data and the detection results represented by each of the multiple types of onboard sensor data. For example, it is possible to detect whether there are conflicts between the detection results represented by the current IoV data and the detection results represented by a first type of onboard sensor data (e.g., image data captured by an onboard camera); detect whether there are conflicts between the detection results represented by the current IoV data and the detection results represented by a second type of onboard sensor data (e.g., current point cloud data captured by an onboard millimeter-wave radar); and detect whether there are conflicts between the detection results represented by the current IoV data and the detection results represented by a third type of onboard sensor data (e.g., current point cloud data captured by an onboard lidar).

[0134] Step S62: When there is a conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data, determine whether to accept the current Internet of Vehicles data based on the Internet of Vehicles data within the first time period and the target vehicle-mounted sensor data within the first time period.

[0135] In this embodiment, when it is determined that the detection result represented by the current Internet of Vehicles data conflicts with the detection result represented by the first vehicle-mounted sensor data, the local target vehicle-mounted sensor can be triggered to perform local verification of the conflicting data, obtain the Internet of Vehicles data within the first time period and the target vehicle-mounted sensor data within the first time period, and determine whether to accept the current Internet of Vehicles data based on the Internet of Vehicles data within the first time period and the target vehicle-mounted sensor data within the first time period. In an optional example, whether there is a conflict can be determined based on whether the difference between the detection results represented by the current Internet of Vehicles data and the multiple vehicle-mounted sensor data exceeds a difference threshold. The difference threshold can be freely set according to demand, such as 15%, and the difference threshold between the detection results represented by the current Internet of Vehicles data and the multiple vehicle-mounted sensor data can be the same threshold, or it can be different depending on the type of data, and there is no restriction on this.

[0136] The first on-board sensor data is any one of a plurality of on-board sensor data of the vehicle. The target on-board sensor data is the detection result of the vehicle's target on-board sensor regarding the driving environment surrounding the vehicle. The target on-board sensor is a high-confidence sensor of the vehicle. The high-confidence sensor can be a preset sensor, such as a frequency-modulated continuous wave (FMCW) radar, or a sensor with the highest confidence level determined based on the real-time confidence levels of multiple on-board sensors and exceeding a confidence threshold. This embodiment does not impose any restrictions on this. The first time period can be a preset time period after the conflict is determined, such as within 10 milliseconds after the conflict is determined. This is not a restriction.

[0137] For example, in an optional specific example, to reduce false alarms and improve the reliability of the vehicle safety system, the difference threshold between the detection results of the current IoV data and the data from multiple onboard sensors can be increased. For example, the difference threshold between the detection results of local onboard sensors (such as cameras) and IoV data can be relaxed from the standard 15% to 20%. If a conflict is determined between the detection results of the current IoV data and the detection results of the first onboard sensor data, a high-confidence sensor (such as a frequency-modulated continuous wave (FMCW) radar) is activated to perform a directional scan of the conflicting target area, with the scanning range limited to a horizontal field of view of ±30°. A pre-loaded fault injection detection algorithm is also used to eliminate sensor anomalies (such as radar mirror reflection interference). Then, within the next 10ms, three frames of FMCW radar data (50Hz sampling rate) and IoV data are collected. The target's true position is calculated using sliding window filtering to determine whether to accept the current IoV data. For example, the following verification logic can be used to determine whether to accept the current IoV data (V2X data):

[0138] ;

[0139] in, is the probability that the detection result represented by the Internet of Vehicles data in the first time period is true, is the probability that the detection result represented by the target vehicle sensor data (such as radar data collected by FMCW radar) in the first time period is true. The denominator is added with 1 to avoid division by zero, and the result is the normalized error.

[0140] In addition, a timeout fuse mechanism is also provided in this embodiment: if verification is not completed within a certain period of time (such as 20ms), the current Internet of Vehicles data is forced to be accepted and executed according to the current Internet of Vehicles data (safety first principle).

[0141] Step S63: When the current Internet of Vehicles data is accepted, the current Internet of Vehicles data and the remaining multiple types of vehicle sensor data except the first vehicle sensor data are fused using their respective credibility scores as fusion weights to obtain the fused data.

[0142] In this embodiment, when it is determined to accept the current vehicle network data, the first vehicle sensor data and its corresponding credibility score are removed, and based on the respective credibility scores of the current vehicle network data and the various vehicle sensor data remaining except the first vehicle sensor data, the current vehicle network data and the various vehicle sensor data remaining except the first vehicle sensor data are respectively assigned corresponding fusion weights, so that based on the respective corresponding fusion weights of the current vehicle network data and the various vehicle sensor data remaining except the first vehicle sensor data, the current vehicle network data and the remaining various vehicle sensor data are fused to obtain fused data, which represents the final detection result of the driving environment around the vehicle.

[0143] In this embodiment, in the event of a data conflict, a first-level arbitration is first performed by quickly reviewing the data through the local target vehicle-mounted sensor, thereby improving vehicle safety.

[0144] In combination with the above embodiments, in one implementation, the present invention further provides a data processing method. In this method, in addition to the above steps, steps S71 to S74 may also be included:

[0145] Step S71: When the target vehicle-mounted sensor fails, or when the vehicle triggers a collaborative perception request, multiple neighboring vehicles are determined based on the communication quality, relative position, and credit scores of the vehicle and multiple surrounding vehicles.

[0146] In this embodiment, it is possible to verify whether the target vehicle's onboard sensor is faulty and determine whether the vehicle itself has triggered a collaborative sensing request, which is used to request V2X sensing data from surrounding vehicles. In an optional example, the protocol design of the collaborative sensing request is as follows:

[0147] Message format: Expand the BSM message fields and add the ConflictResolutionRequest part:

[0148] message ConflictRequest {

[0149] uint32 conflict_id; / / conflict unique identifier

[0150] double timestamp; / / UTC timestamp (μs level)

[0151] GeoPosition target_geo; / / conflict target WGS84 coordinates

[0152] ConflictType type; / / A / B / C type conflict

[0153] repeated SensorType required_sensors; / / Requested sensor type (radar / camera)

[0154] }

[0155] In this embodiment, if a target vehicle's onboard sensor is determined to be faulty, or if the vehicle itself has triggered a collaborative sensing request, multiple neighboring vehicles can be identified based on the vehicle's communication quality, relative position, and credit scores. These neighboring vehicles represent the target vehicle that has returned a collaborative sensing response to the vehicle itself. Communication quality is measured by SINR (Signal to Interference and Noise Ratio), which reflects the reliability of the IoV data transmission link. The credit score can be a credit score determined based on blockchain historical records. The relative position is the straight-line distance between the vehicle itself and the surrounding vehicles.

[0156] In an optional example, neighboring vehicle screening conditions can be set in advance: vehicles with communication quality greater than a first threshold (such as 15db), relative positions within a second threshold range (such as 100m, and a heading angle deviation less than 45°), and credit scores higher than a third threshold are determined as neighboring vehicles, without any restrictions.

[0157] In addition, in a preferred embodiment, multiple initial neighboring vehicles can be determined based on the communication quality, relative position, and credit scores of the vehicle and multiple surrounding vehicles. A greedy algorithm is then used to select multiple neighboring vehicles (i.e., optimal neighboring vehicles) from the multiple initial neighboring vehicles to ensure that their spatial distribution is uniform, thereby avoiding data redundancy.

[0158] Step S72: Send collaborative sensing requests to the multiple neighboring vehicles, and receive collaborative sensing responses returned by the multiple neighboring vehicles.

[0159] In this embodiment, a collaborative sensing request is sent to the multiple neighboring vehicles, and collaborative sensing responses are received from the multiple neighboring vehicles based on the collaborative sensing request. A collaborative sensing response returned by a neighboring vehicle represents the neighboring vehicle's detection results of the driving environment around the vehicle. In an optional example, after receiving the request, the neighboring vehicle can process and return data based on conflict type priority (response period ≤ 30ms).

[0160] In one embodiment, considering possible conflicts, the priority of conflict types is divided into three categories from high to low:

[0161] Class A conflicts (urgent safety incidents): involve sudden braking, collision warnings, etc. (e.g., the deceleration reported by V2X is greater than 6m / s², but the camera does not detect the corresponding target); corresponding response requirement: arbitration completion time is ≤50ms.

[0162] Class B conflicts (inconsistent status): such as differences in traffic light status and lane line recognition (the camera shows a red light, but the V2X reports a green light). The corresponding response requirement is to complete arbitration in ≤200ms.

[0163] Class C conflict (position deviation): The target vehicle's position / speed error exceeds the safety threshold (lateral error >1m, longitudinal error >3m); corresponding response requirement: arbitration is completed in ≤100ms.

[0164] When multiple conflicts occur simultaneously, the arbitration processing priority corresponding to each conflict can be determined, and arbitration decisions can be made in descending order based on the arbitration processing priority. In an optional specific example, the arbitration processing priority is determined as follows: Among them, the conflict type coefficients of A, B, and C conflicts are set from high to low (the highest is 1), and the scene danger level can be dynamically calculated based on the collision time (TTC).

[0165] Step S73: Using the communication quality, relative position and credit score of each neighboring vehicle as the weight of the neighboring vehicle, the collaborative perception responses returned by multiple neighboring vehicles are weighted and fused to obtain the final collaborative perception response.

[0166] In this embodiment, the weight of the neighboring vehicle can be determined based on the communication quality, relative position and credit score of each neighboring vehicle between the vehicle and the neighboring vehicle. Then, based on the weight of each neighboring vehicle, the collaborative perception responses returned by multiple neighboring vehicles are weightedly fused to obtain the final collaborative perception response.

[0167] Step S74: Compare the final collaborative perception response with the current Internet of Vehicles data to determine whether to accept the current Internet of Vehicles data.

[0168] In this embodiment, after obtaining the final collaborative perception response, the final collaborative perception response can be compared with the current Internet of Vehicles data to determine whether to accept the current Internet of Vehicles data.

[0169] In one specific embodiment, if primary arbitration fails to resolve a conflict (e.g., radar failure due to weather), or if the conflict involves multi-target coordination, secondary arbitration (requesting V2V collaborative perception) is triggered. For example, a request for perception data from n neighboring vehicles is broadcast via V2V communication, and this data is fused and verified using a weighted voting mechanism. After receiving collaborative perception responses from multiple neighboring vehicles, the vehicle can first synchronize its clocks using the NTP / PTP protocol to compensate for clock offsets between the neighboring vehicles and the host vehicle. The neighboring vehicles' coordinate systems are then converted to the host vehicle's coordinate system to achieve spatiotemporal alignment. A weighted voting mechanism is then used to fuse multi-source data on target presence and location:

[0170] ;

[0171] in, is the credibility weight of a neighboring vehicle, which represents the credibility of the neighboring vehicle's collaborative perception response. The credibility weight of the neighboring vehicle = 0.3*neighboring vehicle credit score + 0.7*neighboring vehicle sensor quality score. is the probability that the cooperative perception response returned by the neighboring vehicle is true, is the probability that the final collaborative perception response is true.

[0172] In this embodiment, when the first-level arbitration cannot resolve the conflict or the conflict involves multi-target collaboration, the second-level arbitration is triggered, that is, the collaborative perception of the Internet of Vehicles data is triggered to further securely arbitrate multi-source heterogeneous data, solving the problems of low efficiency and insufficient security of traditional single-level conflict processing.

[0173] In combination with the above embodiments, in one implementation, the present invention further provides a data processing method. In this method, in addition to the above steps, steps S81 to S83 may also be included:

[0174] Step S81: When the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the first target conflict type, the actual driving environment around the vehicle is obtained through the traffic management system.

[0175] In this embodiment, when there is a conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data, it is also necessary to determine the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data, and to determine whether the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the first target conflict type. If it is determined that it belongs to the first target conflict type, a request can be sent to the traffic management system to obtain the actual driving environment around the vehicle.

[0176] For example, the first target conflict type may be a traffic light status conflict type. When it is determined that the conflict type belongs to a traffic light conflict type, the real-time traffic management system may be accessed (such as calling a traffic light API through an edge node) to obtain the actual driving environment around the vehicle, such as obtaining the actual traffic light status.

[0177] Step S82: When the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the second target conflict type, the multi-vehicle trajectory data within the conflict time period is retrieved and analyzed by the cloud server to obtain the actual driving environment around the vehicle.

[0178] In this embodiment, when a conflict exists between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data, it is also necessary to determine the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data, so as to determine whether the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the second target conflict type. In the case where it is determined that the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the second target conflict type, the multi-vehicle trajectory data within the conflict time period can be retrieved and analyzed by the cloud server to obtain the actual driving environment around the vehicle.

[0179] For example, the cloud server is requested to retrieve the multi-vehicle trajectory data within the conflict time period (such as within 5 seconds before and after the conflict). The range of the multi-vehicle trajectory data can be the trajectory data of all vehicles within a certain radius (such as within 500m) with the conflict as the center, so as to determine the motion consistency of the target and obtain the actual driving environment around the vehicle.

[0180] Step S83: Based on the actual driving environment around the vehicle and the current Internet of Vehicles data, determine whether to accept the current Internet of Vehicles data.

[0181] In this embodiment, after obtaining the actual driving environment around the vehicle, it can be determined whether to accept the current Internet of Vehicles data based on the actual driving environment around the vehicle and the current Internet of Vehicles data.

[0182] In one example, if a conflict still exists after the second-level arbitration, or if it is a Class B conflict (inconsistent status, such as conflicting traffic light states), authoritative data verification is required, which can trigger a third-level arbitration, namely cloud-based multi-source verification and trajectory backtracking (such as integrating traffic management data and calling high-precision maps for trajectory backtracking). If it is determined that vehicle trajectories exceeding a first threshold (e.g., 80%) point to the same conclusion (such as the detection results of the current Internet of Vehicles data), the current Internet of Vehicles data can be determined to be reliable. In addition, in a preferred embodiment, historical data matching can also be performed: historical decisions for similar scenarios (such as the same time period at the same intersection) within a past period of time (e.g., 30 days) are retrieved. If the similarity of historical decisions exceeds 85%, a corresponding solution is recommended.

[0183] In this embodiment, a three-level arbitration mechanism is proposed to address abnormal conflicts in multimodal data, namely, rapid review by local sensors -> V2V collaborative perception request -> cloud-based multi-source verification and trajectory backtracking. This three-level conflict arbitration mechanism solves the problems of low efficiency and insufficient security in traditional single-level conflict processing, and reduces the theoretical verification conflict resolution delay.

[0184] In conjunction with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. In this method, the credibility assessment of the current Internet of Vehicles data and the credibility assessment of various types of vehicle-mounted sensor data are implemented using the vehicle's credibility assessment model. In addition to the above steps, the method may also include steps S91 to S92, and the "credibility assessment of the current Internet of Vehicles data and the credibility assessment of various types of vehicle-mounted sensor data" in step S12 may specifically include step S93:

[0185] Step S91: Obtain the target geographic grid where the vehicle is currently located.

[0186] In this embodiment, the target geographic grid where the vehicle is currently located can be obtained. A geographic area includes multiple geographic grids of target sizes, and the target geographic grid is the geographic grid where the vehicle is currently located. For example, a city is finely divided into 500m x 500m geographic grids (cells), and each geographic grid (cell) independently manages its corresponding credibility assessment model parameters.

[0187] Step S92: Loading target credibility assessment model parameters corresponding to the target geographic grid.

[0188] In this embodiment, after determining the target geographic grid, the target credibility assessment model parameters corresponding to the target geographic grid can be loaded, and the model parameters of the credibility assessment model can be updated to obtain a credibility assessment model with the target credibility assessment model parameters. The credibility assessment model parameters corresponding to each geographic grid can be deployed in the vehicle controller to enable the loading of the target credibility assessment model parameters. Alternatively, the credibility assessment model parameters corresponding to each geographic grid can be deployed in the cloud system. The cloud system can determine the target credibility assessment model parameters corresponding to the target geographic grid based on the real-time location of the vehicle and send them to the vehicle, thereby enabling the vehicle to load the target credibility assessment model parameters corresponding to the target geographic grid.

[0189] Step S93: Using the credibility evaluation model with the target credibility evaluation model parameters, the credibility evaluation is performed on the current Internet of Vehicles data, and the credibility evaluation is performed on the various types of vehicle-mounted sensor data respectively.

[0190] In this embodiment, a credibility assessment model with target credibility assessment model parameters can be used to perform credibility assessment on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data. In addition, a credibility assessment model with target credibility assessment model parameters can be used to perform credibility assessment on multiple types of vehicle-mounted sensor data to obtain respective credibility scores for the multiple types of vehicle-mounted sensor data.

[0191] In this embodiment, through the embedding mechanism of spatial features (i.e., geographic rasterization model update), the model can be dynamically optimized to adapt to different scenarios, thereby enhancing the generalization ability of the model and improving the regional adaptation accuracy compared to traditional centralized training.

[0192] In conjunction with the above embodiments, in one implementation, an embodiment of the present invention further provides a data processing method. The credibility assessment of the current Internet of Vehicles data is implemented using the Internet of Vehicles credibility assessment model in the credibility assessment model of the vehicle. In addition to the above steps, the method may also include steps S101 and S102:

[0193] Step S101: when the current moment belongs to the first target time period, the weight corresponding to the communication quality item in the Internet of Vehicles credibility evaluation model is increased from the first weight to the second weight.

[0194] In this embodiment, the IoV credibility assessment model is used to perform a credibility assessment on current IoV data to obtain a credibility score for the current IoV data. The IoV credibility assessment model includes at least model parameters for a communication quality item. This embodiment can determine the time period to which the current moment belongs. If it is determined that the current moment belongs to a first target time period, the weight corresponding to the communication quality item in the IoV credibility assessment model can be increased from a first weight to a second weight. For example, the first target time period can be the time period corresponding to morning and evening rush hour. In this case, the weight corresponding to the communication quality item in the IoV credibility assessment model can be assigned a higher weight (e.g., a weight corresponding to the communication quality item of α + 0.2). The first weight can be a preset weight corresponding to the communication quality item.

[0195] Step S102: When the current moment belongs to the second target time period, the weight corresponding to the communication quality item in the Internet of Vehicles credibility evaluation model is restored to the first weight.

[0196] In this embodiment, the second target time period may not be the first target time period. When it is determined that the current moment belongs to the second target time period, the weight corresponding to the communication quality item in the Internet of Vehicles credibility assessment model may be restored to the first weight.

[0197] In addition, in another embodiment, when it is determined that the current moment belongs to the third target time period representing the night time, after determining the fusion weights of the multiple types of vehicle-mounted sensor data, the fusion weights of the multiple types of vehicle-mounted sensor data are added to obtain the updated fusion weights of the multiple types of vehicle-mounted sensor data, and thus the current vehicle-mounted network data and the multiple types of vehicle-mounted sensor data are weightedly fused according to the fusion weights of the current vehicle-mounted network data and the updated fusion weights of the multiple types of vehicle-mounted sensor data to obtain the fused data.

[0198] In this embodiment, the time sliding window mechanism can be used to achieve dynamic optimization of the model to adapt to different scenarios, enhance the generalization ability of the model, and improve the regional adaptation accuracy compared to traditional centralized training.

[0199] In combination with the above embodiments, in one implementation, the present invention further provides a data processing method. In this method, the above step S13 may specifically include step S111 and step S112:

[0200] Step S111: When the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are discrete, the credibility scores of the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are used as fusion weights, and the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are fused through a Bayesian network to obtain the fused data.

[0201] In this embodiment, when the current Internet of Vehicles data and multiple types of vehicle-mounted sensor data are discrete, for example, when it is determined that at least one type of data among the current Internet of Vehicles data and multiple types of vehicle-mounted sensor data is discrete, data fusion processing is performed through a Bayesian network, that is, the fusion weight is determined according to the respective credibility scores of the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data are fused through a Bayesian network based on the fusion weight to obtain fused data.

[0202] This embodiment uses Bayesian networks to resolve discrete data conflicts based on probabilistic relationship modeling for semantic information such as traffic light status recognition and road topology. This not only allows the analysis of non-numerical semantic information, such as the "red light" and "green light" in traffic lights, demonstrating symbolic reasoning capabilities, but the probabilistic form of the Bayesian network output results can also be directly integrated with the autonomous driving decision-making system for seamless connection. It can also flexibly integrate V2X rules (such as traffic light phase information from roadside units (RSUs)) and local sensor data to achieve scalability under dynamic conditions.

[0203] Step S112: When the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are continuous, the credibility scores of the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are used as fusion weights, and the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are fused through weighted Kalman filtering to obtain the fused data.

[0204] In this embodiment, when the current Internet of Vehicles data and multiple types of vehicle-mounted sensor data are of a continuous type, for example, when it is determined that at least one type of data among the current Internet of Vehicles data and multiple types of vehicle-mounted sensor data is of a continuous type, data fusion processing is performed through an improved weighted Kalman filter (WKF), that is, the fusion weight is determined according to the respective credibility scores of the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data are fused through the weighted Kalman filter based on the fusion weight to obtain fused data.

[0205] Considering that traditional Kalman filtering assumes that all input data has constant credibility and cannot adapt to dynamic weighting scenarios of multi-source data such as V2X, this embodiment selects an improved weighted Kalman filter (WKF) to fuse data. This not only dynamically adjusts weights based on the real-time credibility of V2X, sensor, and other data, improving the robustness of state estimation, but also directly modifies the covariance matrix, achieving efficient computation without placing an additional burden on system computing power, thereby improving computational efficiency (fusion efficiency). Furthermore, the weighted Kalman filter is compatible with traditional autonomous driving perception algorithms, facilitating integration with existing systems and achieving compatibility with engineering projects.

[0206] In this embodiment, a modified weighted Kalman filter is used to process continuous spatiotemporal state estimates, such as vehicle trajectory and speed. This approach is suitable for high-frequency data processing scenarios. Furthermore, a Bayesian network is used to efficiently process discrete semantic information, such as traffic rules and events. Despite its low update frequency, this information plays a crucial role in decision-making. This embodiment utilizes a dual "numerical-semantic" validation layer, formed by the WKF and Bayesian network, to address the traditional approach's reliance on a single data source.

[0207] In one embodiment, if Figure 2 As shown, Figure 2 This is an overall system architecture diagram of a multimodal data fusion system based on dynamic credibility evaluation, shown in one embodiment of the present invention. Figure 2 The overall system architecture of the system consists of four layers:

[0208] Multi-source heterogeneous data input layer: As the basic data entry point of the system, it is responsible for accessing, preprocessing, and standardizing multi-source heterogeneous data. Multi-source heterogeneous data includes: (1) the position, speed, and heading angle of surrounding vehicles carried in V2X communication data (BSM / CAM / DENM, etc.); (2) the data of on-board sensors (detection results of cameras, millimeter-wave radars, and lidars). The V2X communication data is then parsed, the on-board sensor data is standardized, and then time-space alignment is performed (including time synchronization and coordinate conversion, with the coordinate conversion error less than 5cm) to obtain preprocessed data, which is then passed to the dynamic credibility assessment layer.

[0209] Dynamic credibility assessment layer: Independently calculates real-time credibility scores for each type of data (such as V2X messages, camera detection frames, radar point clouds, etc.) as the basis for weight allocation.

[0210] Among them, the V2X credibility model in the credibility assessment model can be used to calculate the credibility of the V2X messages in the preprocessed data based on multiple aspects such as SINR, timeliness, and history to obtain a V2X credibility score. The camera model (i.e., the on-board camera credibility model) in the credibility assessment model can be used to calculate the credibility of the image data collected by the on-board camera in the preprocessed data based on multiple aspects such as SSIM, clarity, and illumination to obtain a camera credibility score. The millimeter-wave radar model (i.e., the millimeter-wave radar credibility model) in the credibility assessment model can be used to calculate the credibility of the current point cloud data collected by the millimeter-wave radar in the preprocessed data based on multiple aspects such as point cloud density, velocity consistency, signal-to-noise ratio, and multipath suppression to obtain a millimeter-wave radar credibility score. The lidar model (i.e., the lidar credibility model) in the credibility assessment model can be used to calculate the credibility of the current point cloud data collected by the lidar in the preprocessed data based on multiple aspects such as reflectivity, environmental attenuation, and trajectory continuity to obtain a lidar credibility score. Finally, a real-time credibility matrix is ​​determined based on the V2X credibility score, camera credibility score, millimeter-wave radar credibility score, and lidar credibility score. This real-time credibility matrix is ​​then passed to the fusion decision layer. This dynamic credibility assessment layer enables real-time, multi-factor quantitative evaluation of multimodal perception data, reducing errors compared to traditional static fixed weights or preset weight adjustment rules.

[0211] Fusion decision layer: Adopting an improved weighted Kalman filter (i.e., improved WKF) + Bayesian network fusion framework, the improved WKF performs trajectory fusion based on dynamic weights, and the Bayesian network performs semantic reasoning based on conditional probability tables, thereby dynamically integrating multimodal data and performing confidence threshold judgment to determine whether arbitration triggering is involved. A three-level arbitration mechanism designed for multimodal perception abnormal conflicts (i.e., three-level conflict processing: local -> V2V -> cloud) is used to arbitrate abnormal data and output 360° environmental perception results.

[0212] Model global optimization layer: Continuously optimize the credibility assessment model based on federated learning and enhance the scenario adaptability of the credibility assessment model: each federated client sends local gradients to the edge server, and the edge server implements edge aggregation based on multiple local gradients and sends them to the cloud server. The cloud server updates the gradient based on the processing results of the edge server's security aggregation and sends the incremental update package to the federated client, thereby realizing the update of the credibility assessment model in the federated client.

[0213] Thus, this embodiment provides a multimodal data fusion system based on dynamic credibility assessment and a multimodal data fusion method based on dynamic credibility assessment to address the issues of rigid multimodal data fusion weights, poor adaptability to dynamic environments, and inefficient conflict arbitration in existing V2X collaborative perception systems. The core of this embodiment is to establish a dynamic credibility quantitative assessment model for multimodal data, implement multi-level data conflict arbitration and real-time fusion decision-making, and construct an adaptive credibility model optimization system based on federated learning optimization to address the aforementioned issues in the existing technology.

[0214] In one embodiment, if Figure 3 As shown, Figure 3 This is a flow chart of a three-level conflict arbitration mechanism according to an embodiment of the present invention. Figure 3 When a data conflict is detected, the conflict type is determined to determine whether it is Class A, Class B, or Class C. If the conflict is determined to be Class A, a first-level arbitration (local sensor rapid review) is performed: the FMCW radar is activated to collect three frames of data within 10ms for verification. Based on these three frames, the error is determined to be less than 15%. If the error is less than 15%, a local decision is made. If the error is not less than 15%, a second-level arbitration is triggered, and a V2X collaboration request is made. This first screens neighboring vehicles (for example, those with a SINR ≥ 15dB). Then, a weighted voting mechanism is used to determine the support rate. If the support rate is greater than 80%, the fusion weights are updated, and new data fusion is performed.

[0215] If the conflict is determined to be a Class B conflict, a second-level arbitration (V2V collaborative perception request) is performed: the RSU signal status is requested to determine whether the API response is successful. If the response is successful, the signal phase is forced to synchronize. If the response is unsuccessful, a third-level arbitration is triggered and trajectory backtracking analysis is performed: multi-trajectory verification is requested through the cloud to determine whether the historical matching degree of multiple trajectories is greater than 85%. If the historical matching degree is greater than 85%, a solution is pushed.

[0216] If the conflict is determined to be a Class C conflict, a three-level arbitration (cloud-based multi-source verification and trajectory prediction comparison) is performed: the cloud is requested to retrieve neighboring vehicle data, and the trajectory data of the neighboring vehicle is fitted with a weighted least squares method using a Hampei filter to determine the average error. If the average error is less than 2 meters, the presence of the target (such as an obstacle) is determined.

[0217] In one embodiment, if Figure 4 As shown, Figure 4 This is a diagram of an optimized architecture of a federated learning model shown in one embodiment of the present invention. Figure 4 In the [1], the architecture includes: local clients (such as vehicles / RSU terminals), edge aggregation points (such as MEC servers), and cloud-based global servers.

[0218] The local client stores a local dataset, including local credibility assessment data (such as historical V2X message collision records and sensor false alarm rates). A lightweight subnetwork is deployed, which uses local data to train the credibility model (e.g., updating only the λ and α coefficients, i.e., λ / α updates). During local training, differential privacy (DP) injection is used to add Gaussian noise (σ=0.01) to the gradients, meeting a privacy budget of (ε=1.0, δ=1e-5). Top-K gradient compression is employed, using Top-K sparsification (retaining the top 10% of the maximum gradient values) combined with Huffman coding, achieving a compression rate of 90%. The updated gradients (i.e., encrypted gradients) are then uploaded to the edge aggregation node. (For example, vehicles are allowed to upload gradients during idle periods (such as when parked and charging) to avoid network congestion, thus implementing an asynchronous federation mechanism.)

[0219] In the edge aggregation node, the model gradients uploaded by multiple local clients in the area (such as a radius of 1km) are received and cached (i.e., regional model cache), and preliminary gradient screening (i.e., anomaly detection (3σ principle): outliers are detected based on the L2 norm, and data that deviates from the mean ±3σ is eliminated) is performed. Gradient aggregation is performed through weighted averaging to obtain aggregated gradients and update the regional model. The edge aggregation node synchronizes the updated regional model information to the local clients covered by the edge aggregation node in real time, and uses a secure transmission protocol to transmit the aggregated gradients to the global server in the cloud.

[0220] The global cloud server receives aggregated gradients from multiple edge aggregation nodes, updates the global model in the global model repository based on the aggregated gradients via a secure aggregation protocol, and sends incremental update packages to local clients (which can be pushed to local clients daily). The global cloud server performs spatiotemporal feature embedding: geogridding is performed, dividing the city into 500m x 500m grid cells. Each cell independently manages the parameters of its sub-model. The global cloud server intelligently loads the corresponding parameter configuration (i.e., scenario adaptation parameters) based on the vehicle's real-time location and / or scenario. Furthermore, a temporal sliding window mechanism is used: during morning and evening rush hours, the model assigns a higher weight (α + 0.2) to communication quality, while at night, the weight of sensors is increased (β + 0.15), enabling incremental model updates (i.e., incremental update generation). This approach, through federated learning and spatiotemporal gridding, achieves global optimization of model parameters, enhances model generalization, and improves regional adaptation accuracy compared to traditional centralized training.

[0221] In this way, without leaking raw vehicle / roadside unit data, this embodiment continuously optimizes the credibility assessment model through distributed training. This model automatically adapts to local characteristics for different regions (cities, highways, tunnels, etc.), environmental conditions (day and night, weather), and traffic density, enhancing its generalization capabilities. The conflict case library is also updated: the resolution of the conflict is stored in the federated learning case library for incremental learning by other vehicles. Dynamic model adjustments are also implemented: if a certain type of conflict frequently occurs in a certain area (such as GPS drift in a tunnel), fine-tuning of the edge node model (such as adjusting the failure factor) is triggered.

[0222] In a specific example, multi-vehicle collaborative perception and traffic light status conflict arbitration were implemented at urban intersections:

[0223] The application scenario of this embodiment is: at an intersection on a main road in a city, with dense traffic during the morning rush hour, the following conflict occurs: the RSU sends a green light signal via V2X (confidence 0.92), the on-board camera detects a red light (confidence 0.85), and the lidar point cloud shows that the vehicle ahead is braking suddenly (deceleration > 8m / s²).

[0224] The specific steps of this embodiment are as follows:

[0225] 1) Data input and preprocessing:

[0226] V2X message analysis: extraction of signal light phase (G), SINR = 18dB, timestamp alignment error < 1ms;

[0227] Sensor data obtained: Camera image SSIM = 0.88, light intensity 120,000 Lux (backlight); LiDAR detected three vehicles braking suddenly, trajectory continuity factor = 0.4;

[0228] Perform spatiotemporal alignment: world coordinate system conversion error <3cm.

[0229] 2) Dynamic credibility assessment:

[0230] V2X credibility calculation: ;

[0231] Camera credibility calculation: ;

[0232] LiDAR credibility: Environmental attenuation factor = 0.6 (visibility 500m), trajectory continuity triggers anomaly detection.

[0233] 3) Fusion decision-making and conflict triggering:

[0234] Inference via Bayesian networks:

[0235] Input Evidence: In wireless communications, the signal-to-interference-plus-noise ratio (SINR) is a key metric for measuring communication quality. For example, an SINR of 18dB typically indicates high communication quality, which can result in vehicle-to-everything (V2X) communication accuracy rates as high as 94%. Furthermore, variations in light intensity can affect a camera's false alarm rate. For example, under certain lighting conditions, the false alarm rate can reach 32%.

[0236] Calculate the posterior probability: P (true green light) = 0.94 × (1-0.32) × 0.7 / [0.94 × 0.68 × 0.7 + 0.06 × 0.32 × 0.3] = 0.91

[0237] Conflict determination: Type A conflict (sudden braking event) and Type B conflict (traffic light conflict) are triggered simultaneously.

[0238] 4) Three-level arbitration execution:

[0239] Level 1 Arbitration: Activate FMCW radar directional scanning and confirm the existence of two vehicles braking suddenly within 3 frames.

[0240] Secondary arbitration: V2V requests data from three neighboring vehicles (response time 28ms), and four or five vehicles report a green light status.

[0241] Level 3 arbitration: The traffic signal log is retrieved from the cloud and confirmed that the RSU message is delayed by 2 seconds (the light is actually red).

[0242] 5) Federated Learning Optimization:

[0243] Conflict case upload: Mark the "High-light V2X-Camera Timing Inconsistency" scenario.

[0244] Incremental model update: Dynamically adjust the time decay factor λ in the traffic light CPT from 0.3 to 0.5; edge nodes send update packages, and the vehicle model update time in the area takes less than 15 seconds.

[0245] In another specific example, emergency obstacle detection in foggy sections of highways was implemented:

[0246] The application scenario of this embodiment is: sudden fog on a highway (visibility <50m), and the following data conflicts occur: the V2X message reports a stationary obstacle 100m ahead (confidence 0.82), the millimeter-wave radar detects multipath interference (point cloud density suddenly increases by 300%), and the camera cannot recognize it due to low light (SSIM = 0.45).

[0247] The specific steps of this embodiment are as follows:

[0248] 1) Initial screening of data quality:

[0249] V2X message verification: historical consistency H = 0.95 (error in the past 5 frames < 1m);

[0250] Sensor anomaly flag: radar point cloud density score = 0.2 (compared to the sliding window mean), camera credibility is automatically downgraded to 0.3;

[0251] 2) Dynamic weight fusion:

[0252] Improved WKF parameters: , the state estimation covariance is expanded by 2 times (obstacle area);

[0253] Trajectory prediction correction: The deviation between the extended Kalman filter predicted position and the V2X report is 0.8m, and the dynamic trust threshold is relaxed to 1.2m (bad weather strategy);

[0254] 3) Multi-level conflict handling:

[0255] Level 1 arbitration: The lidar detects a metal reflector through the fog layer (confidence level 0.78);

[0256] Secondary arbitration: V2V collaboration request is triggered to obtain radar data of the rear vehicle (3 / 5 vehicles confirm obstacles);

[0257] Level 3 arbitration: Cloud-based weather data matching, activation of “fog cluster mode” parameters: V2X weight increased to 0.9, radar multipath suppression factor forced to 0.1;

[0258] 4) Model online optimization:

[0259] The edge node aggregates the fog scene gradient data from 10 vehicles;

[0260] Federated learning update: κ in the lidar environment attenuation formula is increased from 0.05 to 0.08, and the dynamic trust threshold module adds a visibility-speed correlation term.

[0261] The above two specific implementation use cases are verified through typical scenarios, demonstrating the benefits of the present invention in terms of dynamic weight accuracy, conflict resolution efficiency, and model generalization capability.

[0262] It should be noted that for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0263] Based on the same inventive concept, an embodiment of the present invention provides a data processing device 500. Figure 5 , Figure 5 FIG. 1 is a structural block diagram of a data processing device provided by an embodiment of the present invention. Figure 5 As shown, the device includes:

[0264] A data acquisition module is used to obtain the current vehicle network data and various vehicle sensor data of the vehicle, wherein the current vehicle network data represents the detection results of other vehicles around the vehicle on the driving environment around the vehicle, and the various vehicle sensor data represents the detection results of various vehicle sensors of the vehicle on the driving environment around the vehicle;

[0265] a scoring determination module, configured to perform a credibility assessment on the current IoV data to obtain a credibility score for the current IoV data, and to perform a credibility assessment on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data;

[0266] The data fusion module is used to fuse the current Internet of Vehicles data and the multiple vehicle-mounted sensor data using their respective credibility scores as fusion weights to obtain fused data, where the fused data represents the final detection result of the driving environment around the vehicle.

[0267] Optionally, the score determination module includes:

[0268] A first determination module is configured to obtain a communication quality parameter, and obtain a communication quality score of the current Internet of Vehicles data based on the communication quality parameter;

[0269] A second determination module is configured to obtain a time difference between a sending time and a receiving time of the current Internet of Vehicles data, and obtain a time decay score of the current Internet of Vehicles data based on the time difference and a decay rate;

[0270] a third determination module, configured to perform consistency analysis on the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and obtain a consistency score for the current Internet of Vehicles data based on a result of the consistency analysis;

[0271] The first evaluation module is used to obtain a credibility score of the current Internet of Vehicles data based on the communication quality score, time decay score and consistency score of the current Internet of Vehicles data.

[0272] Optionally, the multiple types of vehicle-mounted sensor data include at least image data collected by a vehicle-mounted camera; and the scoring determination module includes at least:

[0273] a fourth determination module, configured to perform a structural similarity analysis on the image data collected by the vehicle-mounted camera and the reference image data to obtain a structural similarity score of the vehicle-mounted camera;

[0274] a fifth determination module, configured to perform edge detection on the image data collected by the vehicle-mounted camera, and obtain a clarity score of the vehicle-mounted camera according to the edge detection result;

[0275] A sixth determination module is configured to obtain the ambient light brightness of the driving environment around the vehicle, and obtain a lighting score of the vehicle camera based on the ambient light brightness and the vehicle light on state;

[0276] The second evaluation module is used to obtain a credibility score of the image data collected by the vehicle-mounted camera based on the structural similarity score, clarity score and lighting score of the vehicle-mounted camera.

[0277] Optionally, the multiple types of vehicle-mounted sensor data include at least current point cloud data collected by a vehicle-mounted millimeter-wave radar; and the scoring determination module includes at least:

[0278] a seventh determination module, configured to obtain a point cloud density score of the on-board millimeter-wave radar based on a target point cloud number represented by the current point cloud data collected by the on-board millimeter-wave radar and an average point cloud number represented by a plurality of first historical point cloud data;

[0279] An eighth determination module is configured to obtain a velocity consistency score of the on-board millimeter-wave radar based on the Doppler velocity variance of the current point cloud data and the plurality of first historical point cloud data of the same target collected by the on-board millimeter-wave radar;

[0280] a ninth determination module, configured to obtain a millimeter-wave radar signal strength, and obtain a signal-to-noise ratio score of the vehicle-mounted millimeter-wave radar based on a magnitude relationship between the millimeter-wave radar signal strength and a millimeter-wave radar signal strength threshold;

[0281] a tenth determination module, configured to obtain a multipath suppression score of the on-board millimeter-wave radar based on a proportion of invalid point clouds in the current point cloud data collected by the on-board millimeter-wave radar;

[0282] The third evaluation module is used to obtain a credibility score of the current point cloud data collected by the vehicle-mounted millimeter-wave radar based on the point cloud density score, speed consistency score, signal-to-noise ratio score and multipath suppression score of the vehicle-mounted millimeter-wave radar.

[0283] Optionally, the multiple types of vehicle-mounted sensor data include at least current point cloud data collected by a vehicle-mounted laser radar; and the scoring determination module includes at least:

[0284] an eleventh determination module, configured to obtain a reflectivity score of the vehicle-mounted laser radar based on a target reflectivity and a reflectivity threshold of current point cloud data collected by the vehicle-mounted laser radar;

[0285] The twelfth determination module is used to obtain the visibility and rainfall intensity of the driving environment around the vehicle and obtain the environmental attenuation score of the vehicle-mounted laser radar;

[0286] a thirteenth determination module, configured to obtain a trajectory continuity score of the on-board laser radar based on a target position detected by the current point cloud data collected by the on-board laser radar and a target position predicted based on a historical trajectory of the target in a plurality of second historical point cloud data;

[0287] The fourth evaluation module is used to obtain a credibility score of the current point cloud data collected by the vehicle-mounted laser radar based on the reflectivity score, environmental attenuation score and trajectory continuity score of the vehicle-mounted laser radar.

[0288] Optionally, the device further comprises:

[0289] a detection module configured to, after obtaining current Internet of Vehicles data and multiple types of vehicle-mounted sensor data of the vehicle, detect whether there is a conflict between detection results represented by each pair of the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data;

[0290] a first judgment module, configured to determine whether to accept the current Internet of Vehicles data based on the Internet of Vehicles data within a first time period and target onboard sensor data within the first time period, if a conflict exists between a detection result represented by the current Internet of Vehicles data and a detection result represented by the first onboard sensor data; the target onboard sensor data being a detection result of a target onboard sensor of the vehicle regarding the driving environment surrounding the vehicle, and the first onboard sensor data being any one of the multiple types of onboard sensor data;

[0291] Data fusion module, including:

[0292] The data fusion submodule is used to, when the current Internet of Vehicles data is accepted, fuse the current Internet of Vehicles data with the remaining multiple types of vehicle sensor data except the first vehicle sensor data using their respective credibility scores as fusion weights to obtain the fused data.

[0293] Optionally, the device further comprises:

[0294] a neighboring vehicle determination module, configured to determine multiple neighboring vehicles based on the communication quality, relative position, and credit scores of the vehicle and multiple surrounding vehicles when the target vehicle-mounted sensor fails or when the vehicle triggers a collaborative perception request;

[0295] A result returning module is configured to send collaborative sensing requests to the plurality of neighboring vehicles and receive collaborative sensing responses returned by the plurality of neighboring vehicles, wherein the collaborative sensing response returned by a neighboring vehicle represents a detection result of the neighboring vehicle on the driving environment around the vehicle;

[0296] A result fusion module is used to perform weighted fusion on the collaborative sensing responses returned by multiple neighboring vehicles, using the communication quality, relative position and credit score of the vehicle and each neighboring vehicle as the weight of the neighboring vehicle, to obtain a final collaborative sensing response;

[0297] A comparison module is used to compare the final collaborative perception response with the current Internet of Vehicles data to determine whether to accept the current Internet of Vehicles data.

[0298] Optionally, the device further comprises:

[0299] a first acquisition module, configured to acquire, through a traffic management system, an actual driving environment around the vehicle when the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first on-board sensor data belongs to a first target conflict type;

[0300] a second acquisition module configured to retrieve and analyze multi-vehicle trajectory data within a conflict time period through a cloud server to obtain an actual driving environment around the vehicle when the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to a second target conflict type;

[0301] The second judgment module is used to determine whether to accept the current Internet of Vehicles data based on the actual driving environment around the vehicle and the current Internet of Vehicles data.

[0302] Optionally, the credibility evaluation of the current Internet of Vehicles data and the credibility evaluation of the multiple types of vehicle-mounted sensor data are performed using a credibility evaluation model of the vehicle; the device further includes:

[0303] The third acquisition module is used to obtain the target geographic grid where the vehicle is currently located. A geographic area includes multiple geographic grids of target sizes;

[0304] A loading module, used for loading target credibility assessment model parameters corresponding to the target geographic grid;

[0305] Rating determination module, including:

[0306] The scoring determination submodule is used to use the credibility evaluation model with the target credibility evaluation model parameters to perform credibility evaluation on the current Internet of Vehicles data and to perform credibility evaluation on the multiple types of vehicle-mounted sensor data respectively.

[0307] Optionally, the credibility assessment of the current Internet of Vehicles data is performed by using an Internet of Vehicles credibility assessment model in the credibility assessment model of the vehicle; and the device further includes:

[0308] A first adjustment module is configured to increase the weight corresponding to the communication quality item in the Internet of Vehicles credibility assessment model from the first weight to the second weight when the current moment belongs to the first target time period;

[0309] The second adjustment module is used to restore the weight corresponding to the communication quality item in the Internet of Vehicles credibility assessment model to the first weight when the current moment belongs to the second target time period.

[0310] Optionally, the data fusion module includes:

[0311] a first processing module configured to, when the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data are discrete, fuse the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data using a Bayesian network, using the respective credibility scores of the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data as fusion weights, to obtain the fused data;

[0312] The second processing module is used when the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are continuous. The credibility scores of the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are used as fusion weights, and the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are fused through weighted Kalman filtering to obtain the fused data.

[0313] Based on the same inventive concept, another embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps in the data processing method described in any of the above embodiments of the present invention are implemented.

[0314] Based on the same inventive concept, another embodiment of the present invention provides an electronic device, such as Figure 6 shown. Figure 6 This is a schematic diagram of an electronic device according to an embodiment of the present invention. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When executed by the processor, the computer program implements the steps of the data processing method according to any of the above embodiments of the present invention.

[0315] Based on the same inventive concept, another embodiment of the present invention provides a vehicle, which includes at least a controller, and the controller executes the steps of the data processing method described in any of the above embodiments of the present invention.

[0316] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0317] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0318] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0319] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0320] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0321] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0322] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0323] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0324] The above is a detailed introduction to the data processing method, device, equipment, medium and vehicle provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A data processing method, characterized in that: The method comprises: Obtaining current Internet of Vehicles data and multiple on-board sensor data of the vehicle, wherein the current Internet of Vehicles data represents detection results of the driving environment around the vehicle by other vehicles around the vehicle, and the multiple on-board sensor data represents detection results of the driving environment around the vehicle by multiple on-board sensors of the vehicle; Performing a credibility evaluation on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data, and performing a credibility evaluation on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data; fusing the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data using their respective credibility scores as fusion weights to obtain fused data, where the fused data represents a final detection result of the driving environment surrounding the vehicle; Performing a credibility assessment on the current Internet of Vehicles data to obtain a credibility score for the current Internet of Vehicles data includes: Obtaining a communication quality parameter, and obtaining a communication quality score of the current Internet of Vehicles data based on the communication quality parameter; Obtaining a time difference between a sending time and a receiving time of the current Internet of Vehicles data, and obtaining a time decay score of the current Internet of Vehicles data based on the time difference and a decay rate; Performing consistency analysis on the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and obtaining a consistency score for the current Internet of Vehicles data based on a result of the consistency analysis; Based on the communication quality score, time decay score and consistency score of the current Internet of Vehicles data, a credibility score of the current Internet of Vehicles data is obtained.

2. The data processing method according to claim 1, wherein: The multiple types of vehicle-mounted sensor data include at least image data collected by a vehicle-mounted camera; The credibility of the plurality of vehicle-mounted sensor data is evaluated respectively to obtain a credibility score for each of the plurality of vehicle-mounted sensor data, including at least: Performing structural similarity analysis on the image data collected by the vehicle-mounted camera and the reference image data to obtain a structural similarity score of the vehicle-mounted camera; Performing edge detection on the image data collected by the vehicle-mounted camera, and obtaining a clarity score of the vehicle-mounted camera according to the edge detection result; Obtaining the ambient light brightness of the driving environment around the vehicle, and obtaining a lighting score for the vehicle camera based on the ambient light brightness and the vehicle light on status; Based on the structural similarity score, clarity score and lighting score of the vehicle-mounted camera, a credibility score of the image data collected by the vehicle-mounted camera is obtained.

3. The data processing method according to claim 1, wherein: The multiple types of vehicle-mounted sensor data include at least current point cloud data collected by the vehicle-mounted millimeter-wave radar; The credibility of the plurality of vehicle-mounted sensor data is evaluated respectively to obtain a credibility score for each of the plurality of vehicle-mounted sensor data, including at least: Obtaining a point cloud density score of the on-board millimeter-wave radar based on the number of target point clouds represented by the current point cloud data collected by the on-board millimeter-wave radar and the average number of point clouds represented by the plurality of first historical point cloud data; Obtaining a velocity consistency score of the on-board millimeter-wave radar based on a Doppler velocity variance of the current point cloud data and the plurality of first historical point cloud data of the same target; Obtaining millimeter-wave radar signal strength, and obtaining a signal-to-noise ratio score of the vehicle-mounted millimeter-wave radar based on a magnitude relationship between the millimeter-wave radar signal strength and a millimeter-wave radar signal strength threshold; Obtaining a multipath suppression score of the on-board millimeter-wave radar according to a proportion of invalid point clouds in the current point cloud data collected by the on-board millimeter-wave radar; Based on the point cloud density score, speed consistency score, signal-to-noise ratio score, and multipath suppression score of the on-board millimeter-wave radar, a credibility score of the current point cloud data collected by the on-board millimeter-wave radar is obtained.

4. The data processing method according to claim 1, wherein: The multiple types of vehicle-mounted sensor data include at least current point cloud data collected by a vehicle-mounted laser radar; and credibility assessments are performed on the multiple types of vehicle-mounted sensor data to obtain respective credibility scores of the multiple types of vehicle-mounted sensor data, including at least: Obtaining a reflectivity score of the vehicle-mounted laser radar based on a target reflectivity and a reflectivity threshold of current point cloud data collected by the vehicle-mounted laser radar; Obtain visibility and rainfall intensity of the driving environment around the vehicle, and obtain the environmental attenuation score of the vehicle-mounted lidar; Obtaining a trajectory continuity score for the on-board laser radar based on a target position detected by the current point cloud data collected by the on-board laser radar and a target position predicted based on a historical trajectory of the target in a plurality of second historical point cloud data; Based on the reflectivity score, environmental attenuation score, and trajectory continuity score of the on-board laser radar, a credibility score of the current point cloud data collected by the on-board laser radar is obtained.

5. The data processing method according to claim 1, wherein: After obtaining the current Internet of Vehicles data and multiple vehicle sensor data of the vehicle, the method further includes: Detecting whether there is a conflict between the current Internet of Vehicles data and the detection results represented by each of the multiple types of vehicle-mounted sensor data; In the event that a detection result represented by the current Internet of Vehicles data conflicts with a detection result represented by the first vehicle-mounted sensor data, determining whether to accept the current Internet of Vehicles data based on the Internet of Vehicles data within a first time period and target vehicle-mounted sensor data within the first time period; the target vehicle-mounted sensor data is a detection result of a target vehicle-mounted sensor of the vehicle with respect to the driving environment surrounding the vehicle, and the first vehicle-mounted sensor data is any one of the multiple types of vehicle-mounted sensor data; The current Internet of Vehicles data and the multiple vehicle sensor data are fused using their respective credibility scores as fusion weights to obtain fused data, including: When the current Internet of Vehicles data is accepted, the credibility scores of the current Internet of Vehicles data and the remaining multiple types of vehicle sensor data except the first vehicle sensor data are used as fusion weights, and the current Internet of Vehicles data and the remaining multiple types of vehicle sensor data are fused to obtain the fused data.

6. The data processing method according to claim 5, characterized in that: The method further comprises: In the event that the target vehicle's onboard sensor fails, or in the event that the vehicle triggers a collaborative perception request, determining multiple neighboring vehicles based on the vehicle's communication quality and relative positions with multiple surrounding vehicles and the credit scores of the multiple surrounding vehicles; Sending collaborative sensing requests to the multiple neighboring vehicles and receiving collaborative sensing responses returned by the multiple neighboring vehicles, wherein the collaborative sensing response returned by a neighboring vehicle represents a detection result of the neighboring vehicle on the driving environment around the vehicle; The collaborative sensing responses returned by multiple neighboring vehicles are weighted and fused using the communication quality, relative position, and credit score of each neighboring vehicle as the weight of the neighboring vehicle to obtain the final collaborative sensing response. Compare the final collaborative perception response with the current Internet of Vehicles data to determine whether to accept the current Internet of Vehicles data.

7. The data processing method according to claim 6, characterized in that: The method further comprises: When the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first on-vehicle sensor data belongs to a first target conflict type, obtaining an actual driving environment around the vehicle through a traffic management system; If the type of conflict between the detection result represented by the current Internet of Vehicles data and the detection result represented by the first vehicle-mounted sensor data belongs to the second target conflict type, retrieving and analyzing multi-vehicle trajectory data within the conflict time period through a cloud server to obtain an actual driving environment around the vehicle; Based on the actual driving environment around the vehicle and the current Internet of Vehicles data, determine whether to accept the current Internet of Vehicles data.

8. The data processing method according to any one of claims 1 to 7, characterized in that: The credibility evaluation of the current Internet of Vehicles data and the credibility evaluation of the multiple types of vehicle-mounted sensor data are performed using a credibility evaluation model of the vehicle. The method further includes: Get the target geographic grid where the vehicle is currently located. A geographic area includes multiple geographic grids of target sizes. Loading target credibility assessment model parameters corresponding to the target geographic grid; Performing a credibility assessment on the current Internet of Vehicles data, and performing a credibility assessment on each of the multiple types of vehicle-mounted sensor data, including: The credibility evaluation model having the target credibility evaluation model parameters is used to perform credibility evaluation on the current Internet of Vehicles data, and credibility evaluation is performed on the multiple types of vehicle-mounted sensor data respectively.

9. The data processing method according to any one of claims 1 to 7, characterized in that: The credibility assessment of the current Internet of Vehicles data is performed using an Internet of Vehicles credibility assessment model in the credibility assessment model of the vehicle; the method further includes: When the current moment belongs to the first target time period, the weight corresponding to the communication quality item in the Internet of Vehicles credibility assessment model is increased from the first weight to the second weight; When the current moment belongs to the second target time period, the weight corresponding to the communication quality item in the Internet of Vehicles credibility evaluation model is restored to the first weight.

10. The data processing method according to any one of claims 1 to 7, characterized in that: The current Internet of Vehicles data and the multiple vehicle sensor data are fused using their respective credibility scores as fusion weights to obtain fused data, including: When the current Internet of Vehicles data and the multiple types of vehicle sensor data are discrete, using the respective credibility scores of the current Internet of Vehicles data and the multiple types of vehicle sensor data as fusion weights, the current Internet of Vehicles data and the multiple types of vehicle sensor data are fused through a Bayesian network to obtain the fused data; In the case that the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are continuous, the credibility scores of the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are used as fusion weights, and the current Internet of Vehicles data and the multiple vehicle-mounted sensor data are fused through weighted Kalman filtering to obtain the fused data.

11. A data processing device, characterized in that: The device comprises: A data acquisition module is used to obtain the current vehicle network data and various vehicle sensor data of the vehicle, wherein the current vehicle network data represents the detection results of other vehicles around the vehicle on the driving environment around the vehicle, and the various vehicle sensor data represents the detection results of various vehicle sensors of the vehicle on the driving environment around the vehicle; a scoring determination module, configured to perform a credibility assessment on the current IoV data to obtain a credibility score for the current IoV data, and to perform a credibility assessment on each of the multiple types of vehicle-mounted sensor data to obtain a credibility score for each of the multiple types of vehicle-mounted sensor data; a data fusion module, configured to fuse the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data using their respective credibility scores as fusion weights to obtain fused data, wherein the fused data represents a final detection result of the driving environment around the vehicle; The scoring determination module includes: A first determination module is configured to obtain a communication quality parameter, and obtain a communication quality score of the current Internet of Vehicles data based on the communication quality parameter; A second determination module is configured to obtain a time difference between a sending time and a receiving time of the current Internet of Vehicles data, and obtain a time decay score of the current Internet of Vehicles data based on the time difference and a decay rate; a third determination module, configured to perform consistency analysis on the current Internet of Vehicles data and the multiple types of vehicle-mounted sensor data, and obtain a consistency score for the current Internet of Vehicles data based on a result of the consistency analysis; The first evaluation module is used to obtain a credibility score of the current Internet of Vehicles data based on the communication quality score, time decay score and consistency score of the current Internet of Vehicles data.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is executed by the processor, the data processing method according to any one of claims 1 to 10 is implemented.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the data processing method according to any one of claims 1 to 10 is implemented.

14. A vehicle comprising at least a controller, characterized in that: When executed, the controller implements the data processing method according to any one of claims 1 to 10.

Images