Sensor detection method, apparatus, computer device, and computer program product

Abstract

The application relates to a sensor detection method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: when a vehicle terminal detects that a vehicle is in a driving mode, scene data in a vehicle driving process is acquired, real-time sensor data of the vehicle is acquired, the type of a scene in which the vehicle is located is determined according to the scene data in the vehicle driving process, the target sensor data of the vehicle in the scene type is acquired according to the type of the scene in which the vehicle is located, the real-time sensor data and the target sensor data are matched, and whether the sensor of the vehicle is abnormal is determined according to a matching result. Therefore, automatic detection of whether the sensor of the vehicle is abnormal can be realized, and the sensor of the vehicle does not need to be diagnosed for faults by professional personnel through instruments in a professional place. Therefore, the detection efficiency of the sensor of the vehicle is improved, and the detection cost is saved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobiles, in particular to a sensor detection method and device, computer equipment, storage medium and computer program product. BACKGROUND

[0002] With the development of automobile intelligent technology, driving experience is more and more concerned by the majority of car owners. At present, driving experience mainly depends on intelligent hardware, such as various sensors in vehicles.

[0003] In the traditional technology, in order to improve the safety and user experience of driving, the sensors on the vehicle are generally diagnosed by instruments and meters regularly to detect whether the sensors are abnormal. Thus, the safety problems and the problems that the driving function cannot be used due to sensor abnormalities are avoided.

[0004] However, the current way of diagnosing the faults of the sensors on the vehicle by instruments and meters mainly relies on professional personnel to detect in professional places, thus resulting in low detection efficiency. SUMMARY

[0005] Therefore, it is necessary to provide a sensor detection method, device, computer equipment, storage medium and computer program product capable of improving detection efficiency in view of the above technical problems.

[0006] In a first aspect, the present application provides a sensor detection method. The method comprises:

[0007] When it is monitored that the vehicle is in a driving mode, scene data in the driving process of the vehicle is acquired, and real-time data of a sensor of the vehicle is acquired;

[0008] According to the scene data in the driving process of the vehicle, a scene type in which the vehicle is located is determined;

[0009] According to the scene type in which the vehicle is located, target data of the sensor of the vehicle under the scene type is acquired;

[0010] The real-time data of the sensor and the target data of the sensor are matched, and it is determined whether the sensor of the vehicle is abnormal according to a matching result.

[0011] In one of the embodiments, the scene data comprises scene time sequence features and scene point cloud features; the scene data of the vehicle driving process is acquired when the vehicle is in the driving mode, comprising: when the vehicle is in the driving mode, real-time collection of vehicle signals, positioning information, radar calibration parameters of the vehicle and radar point cloud data of the vehicle in the vehicle driving process; time sequence analysis and processing of the vehicle signals and the positioning information in the vehicle driving process to obtain the scene time sequence features in the vehicle driving process; point cloud data analysis and processing of the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle in the vehicle driving process to obtain the scene point cloud features in the vehicle driving process.

[0012] In one of the embodiments, the time sequence analysis and processing is performed by using an attention mechanism-based encoding and decoding model; the time sequence analysis and processing of the vehicle signals and the positioning information in the vehicle driving process to obtain the scene time sequence features in the vehicle driving process comprises: inputting the vehicle signals and the positioning information in the vehicle driving process into the attention mechanism-based encoding and decoding model to obtain the scene time sequence features in the vehicle driving process.

[0013] In one of the embodiments, the point cloud data analysis and processing is performed by using a segmentation model; the point cloud data analysis and processing of the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle in the vehicle driving process to obtain the scene point cloud features in the vehicle driving process comprises: inputting the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle in the vehicle driving process into the segmentation model to obtain the scene point cloud features in the vehicle driving process.

[0014] In one of the embodiments, the scene data comprises scene time sequence features and scene point cloud features; the determination of the scene type in which the vehicle is located according to the scene data in the vehicle driving process comprises: inputting the scene time sequence features and the scene point cloud features in the vehicle driving process into a scene classification model to obtain the scene type in which the vehicle is located.

[0015] In one of the embodiments, the acquisition method of the scene classification model comprises: acquisition of sample data of a sample vehicle in different scene types; wherein, the sample data comprises a scene type in which the sample vehicle is located, a scene time sequence feature sample set and a scene point cloud feature sample set in the scene type; training of a basic classification model by using the sample data to obtain the scene classification model.

[0016] In one of the embodiments, the basic classification model comprises at least one of a neural network classification model, a nearest neighbor classification model, a decision tree classification model, a Bayesian classification model and a linear classification model.

[0017] In one of the embodiments, the sensor target data comprises target parameters of the sensor in a normal state, and parameter value ranges corresponding to the target parameters; the sensor real-time data comprises real-time parameters of the sensor, and parameter values corresponding to the real-time parameters; and the matching of the sensor real-time data and the sensor target data, and the determination of whether the sensor of the vehicle is abnormal according to the matching result comprises: when the real-time parameters of the sensor are inconsistent with the target parameters of the sensor in the scene type, determining that the matching result of the sensor real-time data and the sensor target data is mismatching, and determining that the sensor of the vehicle is abnormal; or when the real-time parameters of the sensor are consistent with the target parameters of the sensor in the scene type, but the parameter values corresponding to the real-time parameters do not satisfy the parameter value ranges of the corresponding target parameters, determining that the matching result of the sensor real-time data and the sensor target data is mismatching, and determining that the sensor of the vehicle is abnormal.

[0018] In one of the embodiments, the sensor target data comprises target parameters of the sensor in a normal state, and parameter value ranges corresponding to the target parameters; the sensor real-time data comprises real-time parameters of the sensor, and parameter values corresponding to the real-time parameters; and the matching of the sensor real-time data and the sensor target data, and the determination of whether the sensor of the vehicle is abnormal according to the matching result comprises: inputting the target parameters of the sensor in the normal state, the parameter value ranges corresponding to the target parameters, the real-time parameters of the sensor, and the parameter values corresponding to the real-time parameters into a similarity calculation model to obtain a similarity between the sensor real-time data and the sensor target data; and when the similarity is less than a similarity threshold, determining that the matching result of the sensor real-time data and the sensor target data is mismatching, and determining that the sensor of the vehicle is abnormal.

[0019] In one of the embodiments, after determining that the sensor of the vehicle is abnormal, the method further comprises: reporting an abnormal result of the sensor, wherein the abnormal result comprises a result that the sensor is abnormal.

[0020] In one of the embodiments, the driving mode comprises any one of a manual driving mode and an automatic driving mode.

[0021] In a second aspect, the present application further provides a sensor detection device. The device comprises:

[0022] a driving monitoring module configured to perform the following when it is monitored that the vehicle is in a driving mode: acquiring scene data in a driving process of the vehicle, and acquiring sensor real-time data of the vehicle;

[0023] a scene type determination module configured to determine a scene type in which the vehicle is located according to scene data in a driving process of the vehicle;

[0024] a data query module configured to obtain sensor target data of the vehicle in the scene type according to the scene type in which the vehicle is located;

[0025] an anomaly detection module configured to match the sensor real-time data and the sensor target data, and determine whether the sensor of the vehicle is abnormal according to a matching result.

[0026] In a third aspect, the present application further provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method in the first aspect when executing the computer program.

[0027] In a fourth aspect, the present application further provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the method in the first aspect.

[0028] In a fifth aspect, the present application further provides a computer program product. The computer program product comprises a computer program, and the computer program is executed by a processor to implement the steps of the method in the first aspect.

[0029] The sensor detection method, device, computer device, storage medium and computer program product can realize automatic detection of whether the sensor of the vehicle is abnormal, and do not rely on professional personnel to diagnose the sensor of the vehicle through instruments in a professional place, thereby improving the detection efficiency of the sensor of the vehicle and saving the detection cost. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 FIG. 1 is a flowchart of a sensor detection method according to an embodiment of the present application;

[0031] Figure 2 FIG. 2 is a flowchart of a step of obtaining scene data according to an embodiment of the present application;

[0032] Figure 3A flowchart of the process of acquiring the scene classification model in an embodiment;

[0033] Figure 4 A flowchart of the process of determining the scene type in an embodiment;

[0034] Figure 5 A flowchart of the process of matching the sensor data in an embodiment;

[0035] Figure 6 A structural block diagram of the sensor detection device in an embodiment;

[0036] Figure 7 An internal structural diagram of the computer device in an embodiment. DETAILED DESCRIPTION

[0037] In order to make the purposes, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0038] In an embodiment, as shown in Figure 1 , a sensor detection method is provided, and the present embodiment takes the method applied to a vehicle terminal as an example, which includes the following steps:

[0039] Step 102, when it is monitored that the vehicle is in a driving mode, scene data in the driving process of the vehicle is acquired, and real-time sensor data of the vehicle is acquired.

[0040] The driving mode can be a mode in which the vehicle is in a driving state, including any one of a manual driving mode and an automatic driving mode. Specifically, the automatic driving mode refers to an operation mode in which the vehicle can autonomously drive without human intervention, and the manual driving mode refers to an operation mode in which the vehicle drives based on a manual control command with human intervention.

[0041] The scene data can be feature data for representing the scene in which the vehicle is located. The real-time sensor data can be real-time data of each sensor collected in the driving process of the vehicle. The sensor includes, but is not limited to, at least one of a light-sensing imaging sensor device, a radar sensor, an inertial measurement unit (IMU), a global positioning system (GPS), a vehicle speed sensor, a wheel speed sensor, and a gear sensor of the vehicle. The real-time data of the sensor refers to the parameter value of the sensor collected in real time.

[0042] In the embodiment, the vehicle can be installed with a vehicle terminal, and the vehicle terminal can monitor the driving mode of the vehicle. Specifically, when the vehicle terminal monitors that the vehicle is in the driving mode, scene data in the driving process of the vehicle can be acquired, and real-time sensor data of the vehicle can be acquired.

[0043] In step 104, a scene type in which the vehicle is located is determined according to the scene data in the driving process of the vehicle.

[0044] The scene type can be a category used to represent the environment in which the vehicle is currently located, and the environment is affected by at least one of light factors, weather factors, road factors, and the like of the location where the vehicle is currently located. That is, when at least one of the light factors, the weather factors, the road factors, and the like of the location where the vehicle is located changes, the corresponding environment category can be different, resulting in different final scene types.

[0045] In the embodiment, the vehicle terminal can determine the scene type in which the vehicle is currently located according to the scene data in the driving process of the vehicle.

[0046] In step 106, sensor target data of the vehicle in the scene type is acquired according to the scene type in which the vehicle is located.

[0047] The sensor target data can be sensor standard data corresponding to the scene type, that is, standard data of the sensor that can be collected under different scene types. In some embodiments of the disclosure, the sensor can have different data collection requirements under different scene types, and the data collection requirements can include pre-set parameters (or parameter types) corresponding to the sensor collected data, a value range of the parameters, and the like. Specifically, the vehicle terminal can acquire the sensor target data of the vehicle in the scene type according to the determined scene type in which the vehicle is located.

[0048] In step 108, the real-time sensor data and the sensor target data are matched, and it is determined whether the sensor of the vehicle is abnormal according to a matching result.

[0049] In the embodiment, the vehicle terminal determines whether the sensor of the vehicle is abnormal by matching the sensor real-time data and the sensor target data. For example, when the sensor real-time data and the sensor target data are compared, if the two are not matched, it is determined that the matching result is not matched, and if the matching result is not matched, it is determined that the sensor of the vehicle is abnormal. When the matching result of the sensor real-time data and the sensor target data is matched, it is determined that the sensor of the vehicle is not abnormal. In some embodiments of the present disclosure, the matching result of matched can include consistent data or a difference range of data within a preset error range. Correspondingly, the matching result of not matched means that the matched data is inconsistent or the difference exceeds the error range. The matching can include value matching of the data itself, and can also include matching of the number of data and / or the number of corresponding parameters (or parameter types) of the data.

[0050] In the above sensor detection method, when the vehicle terminal monitors that the vehicle is in the driving mode, scene data in the driving process of the vehicle is acquired, and sensor real-time data of the vehicle is acquired. The scene type in which the vehicle is located is determined according to the scene data in the driving process of the vehicle, and then the sensor target data of the vehicle in the scene type is acquired according to the scene type in which the vehicle is located. The sensor real-time data and the sensor target data are matched, and whether the sensor of the vehicle is abnormal is determined according to the matching result. Thus, automatic detection of whether the sensor of the vehicle is abnormal can be realized, and the detection of the sensor of the vehicle by professional personnel in a professional place through instruments is not needed. Therefore, the detection efficiency of the sensor of the vehicle is improved, and the detection cost is saved.

[0051] In one embodiment, the scene data includes scene time sequence features and scene point cloud features. As shown in FIG. 2, when it is monitored that the vehicle is in the driving mode, the scene data in the driving process of the vehicle is acquired, which can specifically include: Figure 2

[0052] In step 202, when it is monitored that the vehicle is in the driving mode, the vehicle signals, the positioning information, the radar calibration parameters of the vehicle, and the radar point cloud data of the vehicle in the driving process of the vehicle are collected in real time.

[0053] The vehicle signals can be CAN (Controller Area Network) signals of the vehicle. Optionally, the vehicle signals can include an opening state of a vehicle speed sensor, an opening state of a wheel speed sensor, an opening state of a wiper, and the like. The positioning information can include a positioning result of a current position of the vehicle, and can further include map information to which the positioning is directed.

[0054] ​The radar point cloud data of the vehicle is point cloud information obtained by the vehicle emitting a laser signal through a laser radar and then collecting the reflected laser signal. Specifically, each point in the point cloud information contains at least three-dimensional coordinate information. The radar calibration parameter can be a proportional factor or a radar constant for determining the laser radar reflectivity of a target object and the output power value of the laser radar receiver.

[0055] In this embodiment, when the vehicle terminal monitors that the vehicle is in the driving mode, the vehicle signal, the positioning information, the radar calibration parameter of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle are collected in real time.

[0056] In step 204, the vehicle signal and the positioning information in the driving process of the vehicle are subjected to time sequence analysis processing to obtain the scene time sequence feature in the driving process of the vehicle.

[0057] The scene time sequence feature can be a feature of a corresponding scene described by a time sequence feature. The time sequence analysis processing is a method of analyzing the vehicle signal and the positioning information in the driving process of the vehicle, the development process, direction and trend of the time sequence, and predicting the target that can be reached in the future time domain. Specifically, the time sequence analysis processing can be realized by using a mathematical model, or can be realized by using a mathematical method such as probability statistics.

[0058] In this embodiment, the vehicle terminal obtains the scene time sequence feature in the driving process of the vehicle by performing time sequence analysis processing on the vehicle signal and the positioning information in the driving process of the vehicle.

[0059] In step 206, the radar calibration parameter of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle are subjected to point cloud data analysis processing to obtain the scene point cloud feature in the driving process of the vehicle.

[0060] The scene point cloud feature can be a feature of a corresponding scene described by a point cloud feature. The point cloud data analysis processing is a process of extracting the point cloud feature based on the radar calibration parameter of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle. Specifically, the point cloud data analysis processing can be realized by using a neural network, or can be realized by using a filtering algorithm.

[0061] In this embodiment, the vehicle terminal obtains the scene point cloud feature in the driving process of the vehicle by performing point cloud data analysis processing on the radar calibration parameter of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle.

[0062] In the above embodiment, when the vehicle terminal detects that the vehicle is in the driving mode, the vehicle signals, positioning information, radar calibration parameters of the vehicle, and radar point cloud data of the vehicle during the driving process of the vehicle are collected in real time, the vehicle signals and the positioning information during the driving process of the vehicle are processed by time sequence analysis, the scene time sequence features during the driving process of the vehicle are obtained, the radar calibration parameters of the vehicle during the driving process of the vehicle and the radar point cloud data of the vehicle are processed by point cloud data analysis, and the scene point cloud features during the driving process of the vehicle are obtained. Thus, the scene data during the driving process of the vehicle is obtained, so that the scene type in which the vehicle is located can be determined through the scene data subsequently.

[0063] In one embodiment, a coding and decoding model based on an attention mechanism can be used for time sequence analysis processing. Then, the vehicle signals and the positioning information during the driving process of the vehicle are processed by time sequence analysis, and the scene time sequence features during the driving process of the vehicle are obtained. Specifically, the vehicle signals and the positioning information during the driving process of the vehicle are input into the coding and decoding model based on the attention mechanism, so that the scene time sequence features during the driving process of the vehicle are obtained.

[0064] The coding and decoding model based on the attention mechanism can be a model similar to the Transformer, which can process time sequence signals of indefinite length. Specifically, by inputting the vehicle signals and the positioning information during the driving process of the vehicle into the coding and decoding model based on the attention mechanism, the scene time sequence features during the driving process of the vehicle output by the model can be obtained. In this embodiment, the coding and decoding model based on the attention mechanism is used for time sequence analysis processing to obtain the scene time sequence features during the driving process of the vehicle, so that the processing efficiency can be improved.

[0065] In one embodiment, a segmentation model can be used for point cloud data analysis processing. Then, the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during the driving process of the vehicle are processed by point cloud data analysis, and the scene point cloud features during the driving process of the vehicle are obtained. Specifically, the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during the driving process of the vehicle are input into the segmentation model, so that the scene point cloud features during the driving process of the vehicle are obtained.

[0066] The segmentation model can be a segmentation model similar to Mask R-CNN (Mask Region-Convolutional Neural Networks), which can determine the position and category of each target in the picture and give a pixel-level prediction. In this embodiment, by inputting the radar calibration parameters of the vehicle during vehicle driving and the radar point cloud data of the vehicle into the segmentation model, the scene point cloud features of the vehicle during vehicle driving output by the model can be obtained. This embodiment can improve the processing efficiency by using the segmentation model for time series analysis and processing to obtain the scene time series features of the vehicle during vehicle driving.

[0067] In one embodiment, the scene data can include scene time series features and scene point cloud features. Then, a scene classification model can also be used for classification of the scene type. Therefore, according to the scene data of the vehicle during vehicle driving, the scene type in which the vehicle is located can be determined, which can specifically include: inputting the scene time series features and the scene point cloud features of the vehicle during vehicle driving into the scene classification model, so as to obtain the scene type in which the vehicle is located output by the model.

[0068] The scene classification model can be trained based on a basic classification model. In this embodiment, by inputting the scene time series features and the scene point cloud features of the vehicle during vehicle driving into the scene classification model, the scene type in which the vehicle is located output by the model can be obtained. It can be understood that in some embodiments, the time dimensions of the scene time series features and the scene point cloud features are the same or similar. That is, the scene time series features and the scene point cloud features of the same time are input into the scene classification model, so as to obtain the scene type in which the vehicle is located at the time output by the model. This embodiment can further improve the processing efficiency by using the scene classification model to obtain the scene type in which the vehicle is located.

[0069] In one embodiment, as shown in Figure 3 The method for obtaining the scene classification model can include the following steps:

[0070] Step 302, obtaining sample data of a sample vehicle in different scene types.

[0071] The sample data includes the scene type in which the sample vehicle is located, the scene time series feature sample set and the scene point cloud feature sample set under each scene type. For example, if there are three scene types A, B and C, the sample data includes the scene time series feature sample set and the scene point cloud feature sample set of the sample vehicle in the A scene, the scene time series feature sample set and the scene point cloud feature sample set of the sample vehicle in the B scene, and the scene time series feature sample set and the scene point cloud feature sample set of the sample vehicle in the C scene. That is, the sample data of the sample vehicle in different scene types is obtained.

[0072] Step 304, training the basic classification model with the sample data to obtain the scene classification model.

[0073] The basic classification model includes at least one of a neural network classification model, a nearest neighbor classification model, a decision tree classification model, a Bayesian classification model, and a linear classification model. For example, the basic classification model can be any of the above classification models. By training any classification model with the above sample data, the classification model can learn the characteristics of the vehicle in different scene types to obtain the scene classification model.

[0074] In one scenario, the basic classification model can also be a combination of two or more of the above classification models. By training two or more classification models with the above sample data, each classification model can learn the characteristics of the vehicle in different scene types, and then obtain the scene classification model based on the trained classification models to improve the robustness of the scene classification model.

[0075] In the above embodiments, since the model is trained with sample data of sample vehicles in different scene types, the trained scene classification model can accurately and effectively classify the scene type of the vehicle based on scene data during vehicle driving.

[0076] In one embodiment, the following further describes determining the scene type of the vehicle based on scene data during vehicle driving. Specifically, as shown in Figure 4 When it is monitored that the vehicle is in driving mode, the vehicle signals, positioning information, radar calibration parameters of the vehicle, and radar point cloud data of the vehicle during vehicle driving are collected in real time. By inputting the vehicle signals and positioning information during vehicle driving into the attention mechanism-based encoding and decoding model, the time sequence features of the scene during vehicle driving are obtained by time sequence feature extraction of the attention mechanism-based encoding and decoding model.

[0077] By inputting the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during vehicle driving into the segmentation model, the scene point cloud features during vehicle driving are obtained by point cloud feature extraction of the segmentation model. Then, the scene time sequence features and scene point cloud features during vehicle driving are input into the scene classification model to obtain the scene type of the vehicle.

[0078] In the embodiment, the time sequence feature is extracted by adopting the coding-decoding model based on the attention mechanism, the point cloud feature is extracted by adopting the segmentation model, and the extracted scene time sequence feature and scene point cloud feature are taken as the input of the scene classification model, so that the scene type in which the vehicle is located is obtained from the output of the scene classification model, thereby the scene type in which the vehicle is located can be quickly determined.

[0079] In one embodiment, the sensor target data includes target parameters of the sensor in a normal state, and parameter value ranges corresponding to the target parameters; the sensor real-time data includes real-time parameters of the sensor, and parameter values corresponding to the real-time parameters. Then, the sensor real-time data and the sensor target data are matched, and whether the sensor of the vehicle is abnormal is determined according to the matching result. Specifically, when the real-time parameters of the sensor are inconsistent with the target parameters of the sensor in the scene type, it is determined that the matching result of the sensor real-time data and the sensor target data is mismatched, and it is determined that the sensor of the vehicle is abnormal.

[0080] It can be understood that the sensor target data is also different in different scene types, that is, the scene type and the sensor target data have a certain corresponding relationship. For example, if there are three scene types A, B and C, there are corresponding sensor target data a under the scene type A, corresponding sensor target data b under the scene type B, and corresponding sensor target data c under the scene type C.

[0081] Specifically, the sensor target data can be target parameters of the sensor in a normal state under the corresponding scene type, and parameter value ranges corresponding to the target parameters. The target parameters can be the names of the standard parameters corresponding to the normal state of the sensor under the corresponding scene type, and the parameter value ranges are the ranges of the values of the standard parameters. For example, if the corresponding sensor target data a under the scene type A is [(a1: xo-xp), (a2: yo-yp), (a3: zo-zp)], where a1, a2 and a3 are the names of the standard parameters corresponding to the normal state of the sensor under the scene type A, xo-xp is the range of the value of the standard parameter a1 under the scene type A, yo-yp is the range of the value of the standard parameter a2 under the scene type A, and zo-zp is the range of the value of the standard parameter a3 under the scene type A. Therefore, the sensor target data is different under different scene types, which can be different standard parameter names or different ranges of values of the standard parameters.

[0082] The sensor real-time data can be real-time data of each sensor collected during driving of the vehicle. The sensor real-time data includes real-time parameters of the sensor, and parameter values corresponding to the real-time parameters. The real-time parameters refer to the real-time collected sensor parameter names, and the parameter values are the specific values of the real-time collected sensor parameters.

[0083] For example, if it is determined according to the scene data in the vehicle driving process that the scene type in which the vehicle is located is A, and the real-time data r of the sensors of the vehicle obtained is [a1: x1, a2: y2], where a1 and a2 are the names of the real-time collected sensor parameters, x1 is the value of the real-time collected sensor parameter with the name a1, and y2 is the value of the real-time collected sensor parameter with the name a2. If the corresponding sensor target data a of the scene type A is [(a1: xo-xp), (a2: yo-yp), (a3: zo-zp)] obtained. When matching the real-time data of the sensors and the sensor target data, first, whether the real-time parameters of the sensors collected under the scene type A are consistent with the target parameters of the sensors under the corresponding scene type can be compared. As can be seen from the above, the real-time parameters of the sensors collected under the scene type A include a1 and a2, while the target parameters of the sensors under the scene type A include a1, a2 and a3. By comparing the real-time parameters and the target parameters of the sensors under the scene type A, it can be known that the real-time parameters are missing the real-time parameter with the name a3. Therefore, it can be determined that the real-time parameters of the sensors collected under the scene type A are inconsistent with the target parameters of the sensors under the corresponding scene type. That is, it can be determined that the matching result of the real-time data of the sensors and the sensor target data is not matched, and further, it can be determined that the sensors of the vehicle are abnormal.

[0084] In one scenario, when the real-time parameters of the sensors are consistent with the target parameters of the sensors under the scene type, the real-time data of the sensors and the sensor target data are matched, and whether the sensors of the vehicle are abnormal is determined according to the matching result, which can further include: determining whether the real-time data of the sensors and the sensor target data are matched according to the parameter values corresponding to the real-time parameters and the parameter value range of the target parameters of the sensors under the scene type.

[0085] For example, if it is determined according to the scene data in the driving process of the vehicle that the scene type in which the vehicle is located is A, and the real-time data r of the sensors of the vehicle obtained is [a1: x1, a2: y2, a3: z3]. Wherein, a1, a2 and a3 are the names of the real-time collected sensor parameters, x1 is the value of the real-time collected sensor parameter name a1, y2 is the value of the real-time collected sensor parameter name a2, and z3 is the value of the real-time collected sensor parameter name a3. If the corresponding sensor target data a of the scene type A is obtained as [(a1: xo-xp), (a2: yo-yp), (a3: zo-zp)]. By comparison, it can be known that the real-time parameters a1, a2 and a3 of the sensors collected under the scene type A are consistent with the target parameters a1, a2 and a3 of the sensors under the corresponding scene type, therefore, the parameter values corresponding to each real-time parameter and the parameter value range of the corresponding target parameter are further compared. For example, for the real-time parameter a1, the corresponding parameter value is x1, and the parameter value range of the target parameter a1 is xo-xp, then it is judged whether the parameter value x1 of the real-time parameter a1 falls into the parameter value range xo-xp of the target parameter a1. If the parameter value x1 of the real-time parameter a1 does not fall into the parameter value range xo-xp of the target parameter a1, i.e. x1 is less than xo or x1 is greater than xp, it indicates that the parameter value corresponding to the real-time parameter does not satisfy the parameter value range of the corresponding target parameter. Therefore, it can be determined that the matching result of the sensor real-time data and the sensor target data is not matched, and further it can be determined that the sensor of the vehicle is abnormal.

[0086] If the parameter value x1 of the real-time parameter a1 falls into the parameter value range xo-xp of the target parameter a1, it is further judged whether the parameter value y2 of the real-time parameter a2 falls into the parameter value range yo-yp of the target parameter a2. Similarly, if the parameter value y2 of the real-time parameter a2 falls into the parameter value range yo-yp of the target parameter a2, it is further judged whether the parameter value z3 of the real-time parameter a3 falls into the parameter value range zo-zp of the target parameter a3. If the parameter value y2 of the real-time parameter a2 does not fall into the parameter value range yo-yp of the target parameter a2, or the parameter value z3 of the real-time parameter a3 does not fall into the parameter value range zo-zp of the target parameter a3, it can be determined that the matching result of the sensor real-time data and the sensor target data is not matched, and further it can be determined that the sensor of the vehicle is abnormal.

[0087] In the above embodiment, when the real-time parameter of the sensor is inconsistent with the target parameter of the sensor under the scene type, it is determined that the matching result of the sensor real-time data and the sensor target data is not matched, and it is determined that the sensor of the vehicle is abnormal. When the real-time parameter of the sensor is consistent with the target parameter of the sensor under the scene type, it is further determined whether the sensor real-time data and the sensor target data are matched according to the parameter value corresponding to the real-time parameter and the parameter value range of the target parameter. When the parameter value corresponding to the real-time parameter does not satisfy the parameter value range of the corresponding target parameter, it is determined that the matching result of the sensor real-time data and the sensor target data is not matched, and it is determined that the sensor of the vehicle is abnormal. Since the present embodiment calls the sensor target data under the corresponding scene type for comparison based on different scene types before parameter comparison, the sensor state under various scene types can be detected, which is beneficial to improve the accuracy of detection.

[0088] In one embodiment, the sensor target data includes a target parameter under a normal state of the sensor and a parameter value range corresponding to the target parameter; and the sensor real-time data includes a real-time parameter of the sensor and a parameter value corresponding to the real-time parameter. As shown in Figure 5 The sensor real-time data and the sensor target data are matched, and whether the sensor of the vehicle is abnormal is determined according to the matching result. Specifically, the target parameter under the normal state of the sensor, the parameter value range corresponding to the target parameter, the real-time parameter of the sensor, and the parameter value corresponding to the real-time parameter can be input into a similarity calculation model to obtain the similarity between the sensor real-time data and the sensor target data. When the similarity is less than a similarity threshold, it is determined that the matching result of the sensor real-time data and the sensor target data is not matched, and it is determined that the sensor of the vehicle is abnormal.

[0089] The similarity calculation model can be a mathematical model for calculating the similarity between the sensor real-time data and the sensor target data. Specifically, the similarity calculation model can obtain the similarity between the sensor real-time data and the sensor target data by calculating the distance between the features of the sensor real-time data and the features of the sensor target data. Specifically, the smaller the distance between the two, the greater the corresponding similarity; the greater the distance between the two, the smaller the corresponding similarity.

[0090] The similarity threshold can be pre-set as a basis for determining whether the sensor real-time data matches the sensor target data. Specifically, the similarity threshold can be set according to the actual application scenario. In this embodiment, when the similarity between the sensor real-time data and the sensor target data is greater than or equal to the similarity threshold, it can be determined that the matching result of the sensor real-time data and the sensor target data is matched, and it can be determined that the sensor of the vehicle is normal. When the similarity between the sensor real-time data and the sensor target data is less than the similarity threshold, it can be determined that the matching result of the sensor real-time data and the sensor target data is not matched, and it can be determined that the sensor of the vehicle is abnormal.

[0091] In this embodiment, the vehicle terminal inputs the target parameters of the sensor in the normal state under the same scene type, the parameter value range corresponding to the target parameters, the real-time parameters of the sensor, and the parameter values corresponding to the real-time parameters into the similarity calculation model, to obtain the similarity between the sensor real-time data and the sensor target data output by the model. The vehicle terminal further compares the obtained similarity with the similarity threshold. When the similarity is less than the similarity threshold, it is determined that the matching result of the sensor real-time data and the sensor target data is not matched, and it is determined that the sensor of the vehicle is abnormal. Since this embodiment obtains the similarity between the sensor real-time data and the sensor target data by using the similarity calculation model, the calculation efficiency can be improved.

[0092] In one embodiment, after determining that the sensor of the vehicle is abnormal, the above method can further include: reporting an abnormal result of the sensor, wherein the abnormal result includes the result that the sensor is abnormal.

[0093] Specifically, after determining that the sensor of the vehicle is abnormal, the vehicle terminal can also issue an alarm to report the abnormal result of the sensor to the vehicle owner, so that the vehicle owner can timely handle the abnormality of the sensor.

[0094] In one scenario, after determining that the sensor of the vehicle is abnormal, the vehicle terminal can also report the corresponding abnormal result to a remote server. Thus, the server can effectively monitor the state of the sensors of each vehicle in the network. The server can be implemented by an independent server, a server cluster composed of multiple servers, or a cloud server.

[0095] In one scenario, after determining that the sensor of the vehicle is abnormal, the vehicle terminal can further repair the abnormality of the sensor according to the abnormal result, so as to timely repair the abnormality of the sensor locally and avoid safety problems caused by the abnormality of the sensor.

[0096] In one scenario, after determining that the sensor of the vehicle has an exception, the vehicle terminal can also report the exception result to the remote server, and interact with the server according to the exception result to obtain the corresponding repair method to repair the exception of the sensor, so as to realize timely repair of the exception of the sensor.

[0097] In one scenario, after determining that the sensor of the vehicle has an exception, the vehicle terminal can also report the exception result to the remote server, and interact with the server according to the exception result to obtain the corresponding repair method to repair the exception of the sensor, so as to realize timely repair of the exception of the sensor.

[0098] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise stated herein, the execution of these steps has no strict sequence limitation, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0099] Based on the same inventive concept, the embodiments of the present application also provide a sensor detection device for implementing the above-mentioned sensor detection method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more sensor detection device embodiments provided below can refer to the limitations of the sensor detection method in the foregoing, which will not be repeated here.

[0100] In one embodiment, as shown in Figure 6 A sensor detection device is provided, comprising: a driving monitoring module 602, a scene type determination module 604, a data query module 606 and an exception detection module 608, wherein:

[0101] The driving monitoring module 602 is configured to execute the following steps: when it is monitored that the vehicle is in a driving mode, acquiring scene data in the driving process of the vehicle, and acquiring real-time data of the sensor of the vehicle.

[0102] The scene type determination module 604 is configured to determine a scene type in which the vehicle is located according to scene data in a driving process of the vehicle.

[0103] The data query module 606 is configured to obtain sensor target data of the vehicle in the scene type according to the scene type in which the vehicle is located.

[0104] The anomaly detection module 608 is configured to match the real-time sensor data and the sensor target data, and determine whether there is an anomaly in the sensor of the vehicle according to a matching result.

[0105] In an embodiment, the scene data includes scene time sequence features and scene point cloud features; and the driving monitoring module includes: a data acquisition unit configured to acquire, in real time, vehicle signals, positioning information, radar calibration parameters of the vehicle, and radar point cloud data of the vehicle in a driving process of the vehicle when it is monitored that the vehicle is in a driving mode; a time sequence analysis unit configured to perform time sequence analysis processing on the vehicle signals and the positioning information in the driving process of the vehicle to obtain the scene time sequence features in the driving process of the vehicle; and a point cloud data analysis unit configured to perform point cloud data analysis processing on the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle to obtain the scene point cloud features in the driving process of the vehicle.

[0106] In an embodiment, a coding-decoding model based on an attention mechanism is used for time sequence analysis processing; and the time sequence analysis unit is configured to input the vehicle signals and the positioning information in the driving process of the vehicle into the coding-decoding model based on the attention mechanism to obtain the scene time sequence features in the driving process of the vehicle.

[0107] In an embodiment, a segmentation model is used for point cloud data analysis processing; and the point cloud data analysis unit is configured to input the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle in the driving process of the vehicle into the segmentation model to obtain the scene point cloud features in the driving process of the vehicle.

[0108] In an embodiment, the scene type determination module is configured to input the scene time sequence features and the scene point cloud features in the driving process of the vehicle into a scene classification model to obtain the scene type in which the vehicle is located.

[0109] In an embodiment, the device further comprises a scene classification model obtaining module configured to perform: obtaining sample data of a sample vehicle in different scene types, the sample data comprising a scene type in which the sample vehicle is located, a scene time sequence feature sample set and a scene point cloud feature sample set in the scene type; training a basic classification model using the sample data to obtain the scene classification model.

[0110] In an embodiment, the basic classification model comprises at least one of a neural network classification model, a nearest neighbor classification model, a decision tree classification model, a Bayesian classification model, and a linear classification model.

[0111] In an embodiment, the sensor target data comprises a target parameter in a normal state of the sensor and a parameter value range corresponding to the target parameter; the sensor real-time data comprises a real-time parameter of the sensor and a parameter value corresponding to the real-time parameter; and the anomaly detection module is configured to perform: when the real-time parameter of the sensor is inconsistent with the target parameter of the sensor in the scene type, determining that a matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal; or when the real-time parameter of the sensor is consistent with the target parameter of the sensor in the scene type, but the parameter value corresponding to the real-time parameter does not satisfy the parameter value range of the corresponding target parameter, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0112] In an embodiment, the sensor target data comprises a target parameter in a normal state of the sensor and a parameter value range corresponding to the target parameter; the sensor real-time data comprises a real-time parameter of the sensor and a parameter value corresponding to the real-time parameter; and the anomaly detection module is further configured to perform: inputting the target parameter in the normal state of the sensor, the parameter value range corresponding to the target parameter, the real-time parameter of the sensor, and the parameter value corresponding to the real-time parameter into a similarity calculation model to obtain a similarity between the sensor real-time data and the sensor target data; when the similarity is less than a similarity threshold, determining that a matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0113] In an embodiment, the device further comprises a reporting module configured to perform: reporting an abnormal result of the sensor, the abnormal result comprising a result that the sensor is abnormal.

[0114] In an embodiment, the driving mode comprises any one of a manual driving mode and an automatic driving mode.

[0115] Each module in the above sensor detection device can be implemented by software, hardware and combinations thereof, in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above modules.

[0116] In one embodiment, a computer device is provided, which can be a vehicle terminal, and an internal structure diagram thereof can be as shown in Figure 7 The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement a sensor detection method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0117] Those skilled in the art can understand that Figure 7 The structure shown in the above

[0118] In one embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program. The processor executes the computer program to implement the following steps:

[0119] When it is monitored that the vehicle is in a driving mode, scene data in a driving process of the vehicle is acquired, and real-time sensor data of the vehicle is acquired;

[0120] According to the scene data in the driving process of the vehicle, a scene type in which the vehicle is located is determined;

[0121] According to the scene type in which the vehicle is located, target sensor data of the vehicle in the scene type is acquired;

[0122] The sensor real-time data and the sensor target data are matched, and it is determined whether the sensor of the vehicle is abnormal according to a matching result.

[0123] In one embodiment, the scene data includes scene time sequence features and scene point cloud features; when the processor executes the computer program, the following steps are further implemented: when it is monitored that the vehicle is in a driving mode, vehicle signals, positioning information, radar calibration parameters of the vehicle and radar point cloud data of the vehicle during driving of the vehicle are collected in real time; the vehicle signals and the positioning information during driving of the vehicle are subjected to time sequence analysis processing to obtain scene time sequence features during driving of the vehicle; the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle are subjected to point cloud data analysis processing to obtain scene point cloud features during driving of the vehicle.

[0124] In one embodiment, when the processor executes the computer program, the following steps are further implemented: a coding and decoding model based on an attention mechanism is used for time sequence analysis processing; the vehicle signals and the positioning information during driving of the vehicle are input into the coding and decoding model based on the attention mechanism to obtain the scene time sequence features during driving of the vehicle.

[0125] In one embodiment, when the processor executes the computer program, the following steps are further implemented: a segmentation model is used for point cloud data analysis processing; the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle are input into the segmentation model to obtain the scene point cloud features during driving of the vehicle.

[0126] In one embodiment, when the processor executes the computer program, the following steps are further implemented: the scene time sequence features and the scene point cloud features during driving of the vehicle are input into a scene classification model to obtain a scene type in which the vehicle is located.

[0127] In one embodiment, when the processor executes the computer program, the following steps are further implemented: sample data of a sample vehicle in different scene types is obtained, the sample data including a scene type in which the sample vehicle is located, a scene time sequence feature sample set and a scene point cloud feature sample set in the scene type; a basic classification model is trained using the sample data to obtain the scene classification model.

[0128] In one embodiment, the sensor target data includes target parameters of the sensor in a normal state, and parameter value ranges corresponding to the target parameters; the sensor real-time data includes real-time parameters of the sensor, and parameter values corresponding to the real-time parameters; when the computer program is executed by the processor, the following steps are further implemented: when the real-time parameters of the sensor are inconsistent with the target parameters of the sensor in the scene type, it is determined that the matching result of the sensor real-time data and the sensor target data is mismatched, and it is determined that the sensor of the vehicle is abnormal; or, when the real-time parameters of the sensor are consistent with the target parameters of the sensor in the scene type, but the parameter values corresponding to the real-time parameters do not satisfy the parameter value ranges of the corresponding target parameters, it is determined that the matching result of the sensor real-time data and the sensor target data is mismatched, and it is determined that the sensor of the vehicle is abnormal.

[0129] In one embodiment, the sensor target data includes target parameters of the sensor in a normal state, and parameter value ranges corresponding to the target parameters; the sensor real-time data includes real-time parameters of the sensor, and parameter values corresponding to the real-time parameters; when the computer program is executed by the processor, the following steps are further implemented: the target parameters of the sensor in the normal state, the parameter value ranges corresponding to the target parameters, the real-time parameters of the sensor, and the parameter values corresponding to the real-time parameters are input into a similarity calculation model to obtain a similarity between the sensor real-time data and the sensor target data; when the similarity is less than a similarity threshold, it is determined that the matching result of the sensor real-time data and the sensor target data is mismatched, and it is determined that the sensor of the vehicle is abnormal.

[0130] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: reporting an abnormal result of the sensor, the abnormal result including a result that the sensor is abnormal.

[0131] In one embodiment, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0132] When it is monitored that the vehicle is in a driving mode, scene data in a driving process of the vehicle is acquired, and real-time data of a sensor of the vehicle is acquired;

[0133] According to the scene data in the driving process of the vehicle, a scene type in which the vehicle is located is determined;

[0134] According to the scene type in which the vehicle is located, target data of the sensor of the vehicle in the scene type is acquired;

[0135] The sensor real-time data and the sensor target data are matched, and it is determined whether the sensor of the vehicle is abnormal according to a matching result.

[0136] In one embodiment, the scene data includes scene time sequence features and scene point cloud features; when the computer program is executed by the processor, the following steps are further implemented: when it is monitored that the vehicle is in a driving mode, real-time acquisition of vehicle signals, positioning information, radar calibration parameters of the vehicle and radar point cloud data of the vehicle during driving of the vehicle; time sequence analysis processing of the vehicle signals and the positioning information during driving of the vehicle is performed to obtain scene time sequence features during driving of the vehicle; point cloud data analysis processing of the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle is performed to obtain scene point cloud features during driving of the vehicle.

[0137] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: time sequence analysis processing is performed by using an attention mechanism-based encoding-decoding model; the vehicle signals and the positioning information during driving of the vehicle are input into the attention mechanism-based encoding-decoding model to obtain the scene time sequence features during driving of the vehicle.

[0138] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: point cloud data analysis processing is performed by using a segmentation model; the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle are input into the segmentation model to obtain the scene point cloud features during driving of the vehicle.

[0139] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: the scene time sequence features and the scene point cloud features during driving of the vehicle are input into a scene classification model to obtain a scene type in which the vehicle is located.

[0140] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: sample data of a sample vehicle in different scene types is obtained, the sample data including a scene type in which the sample vehicle is located, a scene time sequence feature sample set and a scene point cloud feature sample set in the scene type; a basic classification model is trained by using the sample data to obtain the scene classification model.

[0141] In one embodiment, the sensor target data includes a target parameter in a normal state of the sensor, and a parameter value range corresponding to the target parameter; the sensor real-time data includes a real-time parameter of the sensor, and a parameter value corresponding to the real-time parameter; and the computer program, when executed by the processor, further implements the following steps: when the real-time parameter of the sensor is inconsistent with the target parameter of the sensor in the scene type, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal; or when the real-time parameter of the sensor is consistent with the target parameter of the sensor in the scene type, but the parameter value corresponding to the real-time parameter does not satisfy the parameter value range of the corresponding target parameter, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0142] In one embodiment, the sensor target data includes a target parameter in a normal state of the sensor, and a parameter value range corresponding to the target parameter; the sensor real-time data includes a real-time parameter of the sensor, and a parameter value corresponding to the real-time parameter; and the computer program, when executed by the processor, further implements the following steps: inputting the target parameter in the normal state of the sensor, the parameter value range corresponding to the target parameter, the real-time parameter of the sensor, and the parameter value corresponding to the real-time parameter into a similarity calculation model to obtain a similarity between the sensor real-time data and the sensor target data; and when the similarity is less than a similarity threshold, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0143] In one embodiment, the computer program, when executed by the processor, further implements the following steps: reporting an abnormal result of the sensor, the abnormal result including a result that the sensor is abnormal.

[0144] In one embodiment, a computer program product is provided, including a computer program which, when executed by a processor, implements the following steps:

[0145] When it is monitored that the vehicle is in a driving mode, acquiring scene data in a driving process of the vehicle, and acquiring sensor real-time data of the vehicle;

[0146] According to the scene data in the driving process of the vehicle, determining a scene type in which the vehicle is located;

[0147] According to the scene type in which the vehicle is located, acquiring sensor target data of the vehicle in the scene type;

[0148] The sensor real-time data and the sensor target data are matched, and it is determined whether the sensor of the vehicle is abnormal according to a matching result.

[0149] In one embodiment, the scene data includes scene time sequence features and scene point cloud features; when executed by the processor, the computer program further implements the following steps: when it is monitored that the vehicle is in a driving mode, real-time acquisition of vehicle signals, positioning information, radar calibration parameters of the vehicle and radar point cloud data of the vehicle during driving of the vehicle; time sequence analysis processing of the vehicle signals and the positioning information during driving of the vehicle to obtain scene time sequence features during driving of the vehicle; point cloud data analysis processing of the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle to obtain scene point cloud features during driving of the vehicle.

[0150] In one embodiment, when executed by the processor, the computer program further implements the following steps: time sequence analysis processing by using an attention mechanism-based encoding-decoding model; input of the vehicle signals and the positioning information during driving of the vehicle into the attention mechanism-based encoding-decoding model to obtain the scene time sequence features during driving of the vehicle.

[0151] In one embodiment, when executed by the processor, the computer program further implements the following steps: point cloud data analysis processing by using a segmentation model; input of the radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during driving of the vehicle into the segmentation model to obtain the scene point cloud features during driving of the vehicle.

[0152] In one embodiment, when executed by the processor, the computer program further implements the following steps: input of the scene time sequence features and the scene point cloud features during driving of the vehicle into a scene classification model to obtain a scene type in which the vehicle is located.

[0153] In one embodiment, when executed by the processor, the computer program further implements the following steps: acquisition of sample data of a sample vehicle in different scene types, the sample data including a scene type in which the sample vehicle is located, a scene time sequence feature sample set and a scene point cloud feature sample set in the scene type; training of a basic classification model by using the sample data to obtain the scene classification model.

[0154] In one embodiment, the sensor target data includes a target parameter in a normal state of the sensor, and a parameter value range corresponding to the target parameter; the sensor real-time data includes a real-time parameter of the sensor, and a parameter value corresponding to the real-time parameter; and the computer program, when executed by the processor, further implements the following steps: when the real-time parameter of the sensor is inconsistent with the target parameter of the sensor in the scene type, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal; or when the real-time parameter of the sensor is consistent with the target parameter of the sensor in the scene type, but the parameter value corresponding to the real-time parameter does not satisfy the parameter value range of the corresponding target parameter, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0155] In one embodiment, the sensor target data includes a target parameter in a normal state of the sensor, and a parameter value range corresponding to the target parameter; the sensor real-time data includes a real-time parameter of the sensor, and a parameter value corresponding to the real-time parameter; and the computer program, when executed by the processor, further implements the following steps: inputting the target parameter in the normal state of the sensor, the parameter value range corresponding to the target parameter, the real-time parameter of the sensor, and the parameter value corresponding to the real-time parameter into a similarity calculation model to obtain a similarity between the sensor real-time data and the sensor target data; and when the similarity is less than a similarity threshold, determining that the matching result of the sensor real-time data and the sensor target data is mismatched, and determining that the sensor of the vehicle is abnormal.

[0156] In one embodiment, the computer program, when executed by the processor, further implements the following steps: reporting an abnormal result of the sensor, the abnormal result including a result that the sensor is abnormal.

[0157] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties.

[0158] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0159] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0160] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A sensor detection method, characterized in that, The method includes: When the vehicle is detected to be in driving mode, the system acquires scene data during the vehicle's driving process, as well as real-time sensor data of the vehicle. Based on the scene data during the vehicle's driving process, the scene type in which the vehicle is located is determined; Based on the scene type in which the vehicle is located, obtain the sensor target data of the vehicle under the scene type; The real-time data from the sensor and the target data from the sensor are matched, and the vehicle's sensors are determined to be abnormal based on the matching results. The scene data includes scene temporal features and scene point cloud features; when the vehicle is detected to be in driving mode, the scene data acquired during the vehicle's driving process includes: When the vehicle is detected to be in driving mode, the vehicle signals, positioning information, radar calibration parameters, and radar point cloud data of the vehicle during the driving process are collected in real time. The vehicle signals and positioning information during the vehicle driving process are subjected to time-series analysis and processing. The time-series analysis and processing is implemented using an attention-based encoding and decoding model. By inputting the vehicle signals and positioning information during the vehicle driving process into the attention-based encoding and decoding model, the scene time-series features during the vehicle driving process are obtained. The radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during the driving process are analyzed and processed to obtain the scene point cloud features during the driving process.

2. The method according to claim 1, characterized in that, A segmentation model is used for point cloud data analysis and processing; the point cloud data analysis and processing of the vehicle's radar calibration parameters and radar point cloud data during the vehicle's driving process yields scene point cloud features during the vehicle's driving process, including: The radar calibration parameters of the vehicle during the driving process and the radar point cloud data of the vehicle are input into the segmentation model to obtain the scene point cloud features during the driving process.

3. The method according to claim 1, characterized in that, The scene data includes scene temporal features and scene point cloud features; Determining the scenario type of the vehicle based on scenario data during the vehicle's driving process includes: The scene temporal features and scene point cloud features during the vehicle driving process are input into the scene classification model to obtain the scene type in which the vehicle is located.

4. The method according to claim 3, characterized in that, The method for obtaining the scene classification model includes: Acquire sample data of the sample vehicle in different scene types; wherein, the sample data includes the scene type of the sample vehicle, the scene temporal feature sample set and the scene point cloud feature sample set under the scene type; The basic classification model is trained using the sample data to obtain the scene classification model.

5. The method according to claim 1, characterized in that, The sensor target data includes the target parameters of the sensor under normal conditions, and the parameter value range corresponding to the target parameters; the sensor real-time data includes the real-time parameters of the sensor and the parameter values ​​corresponding to the real-time parameters. The process of matching the real-time data from the sensor with the target data from the sensor, and determining whether there is an anomaly in the vehicle's sensors based on the matching result, includes: When the real-time parameters of the sensor are inconsistent with the target parameters of the sensor under the scene type, it is determined that the matching result between the real-time data of the sensor and the target data of the sensor is mismatched, and it is determined that the sensor of the vehicle is abnormal. Alternatively, when the real-time parameters of the sensor are consistent with the target parameters of the sensor under the scene type, but the parameter value corresponding to the real-time parameters does not meet the parameter value range of the corresponding target parameters, it is determined that the matching result between the real-time data of the sensor and the target data of the sensor is mismatched, and it is determined that the sensor of the vehicle is abnormal.

6. The method according to claim 1, characterized in that, The sensor target data includes the target parameters of the sensor under normal conditions, and the parameter value range corresponding to the target parameters; the sensor real-time data includes the real-time parameters of the sensor and the parameter values ​​corresponding to the real-time parameters. The process of matching the real-time data from the sensor with the target data from the sensor, and determining whether there is an anomaly in the vehicle's sensors based on the matching result, includes: The target parameters of the sensor under normal conditions, the range of parameter values ​​corresponding to the target parameters, the real-time parameters of the sensor, and the parameter values ​​corresponding to the real-time parameters are input into the similarity calculation model to obtain the similarity between the real-time data of the sensor and the target data of the sensor. When the similarity is less than the similarity threshold, it is determined that the matching result between the real-time data of the sensor and the target data of the sensor is not a match, and it is determined that the sensor of the vehicle is abnormal.

7. A sensor detection device, characterized in that, The device includes: The driving monitoring module is configured to acquire scene data during the vehicle's driving process and acquire real-time sensor data of the vehicle when the vehicle is detected to be in driving mode. The scenario type determination module is configured to determine the scenario type of the vehicle based on scenario data during the vehicle's driving process. The data query module is configured to retrieve the sensor target data of the vehicle under the scene type according to the scene type in which the vehicle is located; An anomaly detection module is configured to perform a match between the real-time data of the sensor and the target data of the sensor, and determine whether there is an anomaly in the vehicle's sensors based on the match result; The scene data includes scene temporal features and scene point cloud features; the driving monitoring module is also configured to execute: When the vehicle is detected to be in driving mode, the vehicle signals, positioning information, radar calibration parameters, and radar point cloud data of the vehicle during the driving process are collected in real time. The vehicle signals and positioning information during the vehicle driving process are subjected to time-series analysis and processing. The time-series analysis and processing is implemented using an attention-based encoding and decoding model. By inputting the vehicle signals and positioning information during the vehicle driving process into the attention-based encoding and decoding model, the scene time-series features during the vehicle driving process are obtained. The radar calibration parameters of the vehicle and the radar point cloud data of the vehicle during the driving process are analyzed and processed to obtain the scene point cloud features during the driving process.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images