Vehicle safety processing method and device, program product, storage medium and vehicle
Through the machine learning-based safety evaluation model and the reinforcement learning safety decision model, the problem of poor accuracy in the vehicle safety processing method is solved, and the vehicle safety is improved.
Patent Information
- Application Number
- CN202411131741.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-12-10
AI Technical Summary
The existing vehicle safety treatment methods have poor accuracy in judging expected functional safety events and expected functional safety levels, resulting in inaccurate safety measures provided and affecting vehicle driving safety.
The security evaluation model based on machine learning is used to conduct security assessment of sensor data, combined with the security decision model of reinforcement learning, determine the expected functional safety levels and events, and output the optimal safety measures.
It improves the accuracy of judging the expected functional safety events and safety levels of the vehicle, improves the safety of the entire vehicle, and ensures the accuracy and efficiency of safety measures.
Smart Images
Figure CN119099638A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle safety control, and in particular to a vehicle safety processing method, device, program product, storage medium, and vehicle. Background Art
[0002] With the rapid popularization of intelligent connected vehicles, the domestic and foreign markets have an increasingly strong demand for vehicles with high levels of intelligent driving. Under the existing development environment and supply chain, different levels of intelligent driving systems correspond to different functional combinations and component lists, and the more complex the system, the more safety hazards it represents. Therefore, it is particularly important to develop the expected functional safety for the insufficient functions or performance limitations of these functions and components.
[0003] Safety of the Intended Functionality (SOTIF) is an important part of the vehicle safety system. SOTIF means that the vehicle does not have unreasonable risks caused by hazards due to insufficient design or performance limitations. At present, existing technologies usually use sensor monitoring, fault diagnosis and other methods to monitor the intended functional safety of vehicles. When a safety event is found, control algorithms (such as proportional integral control algorithms and fuzzy control algorithms) are usually used to control the vehicle safely. However, the existing methods have poor accuracy in judging intended functional safety events and intended functional safety levels, which leads to inaccurate safety control measures and affects vehicle driving safety.
[0004] From the above analysis, it can be seen that there is currently no effective solution to the problem that the above-mentioned existing vehicle safety processing methods have poor accuracy in judging expected functional safety events and expected functional safety levels, resulting in inaccurate safety measures provided and poor vehicle driving safety. Summary of the invention
[0005] Embodiments of the present invention provide a vehicle safety processing method, device, program product, storage medium, and vehicle to at least solve the technical problem that the existing vehicle safety processing method has poor accuracy in judging expected functional safety events and expected functional safety levels, resulting in inaccurate safety measures provided and poor vehicle driving safety.
[0006] According to one aspect of an embodiment of the present invention, a vehicle safety processing method is provided, comprising:
[0007] Acquire sensor data of a target vehicle; use a safety assessment model to perform a safety assessment on the sensor data to obtain an assessment result, wherein the assessment result is used to determine an expected functional safety level and expected functional safety events of the target vehicle; use a safety decision model to make a safety decision on the assessment result to obtain a decision result, wherein the decision result is used to determine target safety measures for resolving expected functional safety events.
[0008] Optionally, the above-mentioned vehicle safety processing method also includes: clustering the sensor data according to timestamps to obtain clustering results; obtaining a first processing result of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process different types of functional data; based on the map information and the first processing result, performing a scene analysis on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
[0009] Optionally, the safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and a reasoning model. The classification regression model is used to determine the expected functional safety level, and the reasoning model is used to determine the expected functional safety events.
[0010] Optionally, a safety assessment model is used to perform a safety assessment on the sensor data, and the assessment results obtained include: using a classification regression model to perform classification regression processing on the clustering results, the first processing results, and the analysis results to obtain an expected functional safety level; using an inference model to perform event inference on the first processing results and the analysis results to obtain expected functional safety events.
[0011] Optionally, using the inference model to perform event inference on the first processing result and the analysis result to obtain an expected functional safety event includes: reasoning on the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; reasoning on the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there are target scene features in the traffic scene that meet the event triggering conditions; determining the expected functional safety event based on the first inference result and the second inference result.
[0012] Optionally, a safety decision-making model is used to make a safety decision on the evaluation results, and the decision results include: taking the expected functional safety events and the expected functional safety levels as the states of the safety decision model, taking the whole vehicle decision as the actions of the safety decision model, and taking multiple safety assessment indicators as rewards of the safety decision model, and obtaining the target safety measures through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measures.
[0013] Optionally, the above-mentioned vehicle safety processing method also includes: receiving feedback results on target safety measures, wherein the feedback results are used to evaluate the target safety measures' ability to resolve expected functional safety events; in response to determining that the target safety measures' ability to resolve expected functional safety events is lower than a preset level based on the feedback results, obtaining historical safety measures used to resolve expected functional safety events; adjusting at least one safety assessment indicator of the current safety decision model based on the target safety measures and the historical safety measures to obtain an adjusted safety decision model; retraining the adjusted safety decision model, and using the trained safety decision model as the expected safety decision model for safety decision making.
[0014] According to another aspect of an embodiment of the present invention, a vehicle safety processing device is provided, including:
[0015] An acquisition module is used to acquire sensor data of a target vehicle; an evaluation module is used to perform a safety evaluation on the sensor data using a safety evaluation model to obtain an evaluation result, wherein the evaluation result is used to determine an expected functional safety level and expected functional safety events of the target vehicle; a decision module is used to make a safety decision on the evaluation result using a safety decision model to obtain a decision result, wherein the decision result is used to determine a target safety measure for resolving expected functional safety events.
[0016] Optionally, the above-mentioned vehicle safety processing device also includes: a first processing module, used to cluster the sensor data according to timestamps to obtain clustering results; obtain a first processing result of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process and obtain different types of functional data; based on the map information and the first processing result, perform a scene analysis on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
[0017] Optionally, the safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and a reasoning model. The classification regression model is used to determine the expected functional safety level, and the reasoning model is used to determine the expected functional safety events.
[0018] Optionally, the above-mentioned evaluation module is also used to: use a classification regression model to perform classification regression processing on the clustering results, the first processing results and the analysis results to obtain an expected functional safety level; use an inference model to perform event inference on the first processing results and the analysis results to obtain expected functional safety events.
[0019] Optionally, the above-mentioned evaluation module is also used to: infer the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; infer the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there are target scene features in the traffic scene that meet the event triggering conditions; determine the expected functional safety event based on the first inference result and the second inference result.
[0020] Optionally, the above-mentioned decision module is also used to: use expected functional safety events and expected functional safety levels as the states of the safety decision model, use the whole vehicle decision as the action of the safety decision model, and use multiple safety assessment indicators as rewards of the safety decision model, and obtain target safety measures through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measures.
[0021] Optionally, the above-mentioned vehicle safety processing device also includes: a second processing module, used to receive feedback results on target safety measures, wherein the feedback results are used to evaluate the target safety measures' ability to resolve expected functional safety events; in response to determining that the target safety measures' ability to resolve expected functional safety events is lower than a preset level based on the feedback results, obtaining historical safety measures used to resolve expected functional safety events; adjusting at least one safety assessment indicator of the current safety decision model based on the target safety measures and the historical safety measures to obtain an adjusted safety decision model; retraining the adjusted safety decision model, and using the trained safety decision model as the expected safety decision model for safety decision making.
[0022] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, any one of the aforementioned vehicle safety processing methods is implemented.
[0023] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the aforementioned vehicle safety processing methods.
[0024] According to another aspect of an embodiment of the present invention, a vehicle is provided, comprising an on-board memory and an on-board processor, wherein a computer program is stored in the on-board memory, and the on-board processor is configured to run the computer program to execute any one of the aforementioned vehicle safety processing methods.
[0025] In an embodiment of the present invention, the sensor data of the target vehicle is first obtained, and then the sensor data is security evaluated using a security evaluation model to obtain an evaluation result, wherein the evaluation result is used to determine the expected functional safety level and expected functional safety events of the target vehicle, and finally, the evaluation result is security-determined using a security decision model to obtain a decision result, wherein the decision result is used to determine the target safety measures for solving the expected functional safety events. By using a machine learning-based security evaluation model to accurately evaluate the full range of data related to the driving of the target vehicle (including sensor data and functional data and scene features obtained by processing the sensor data), the expected functional safety events and expected functional safety levels of the target vehicle are obtained, and the optimal safety measures for solving the expected functional safety events are given using a security decision model based on reinforcement learning, the purpose of accurately and efficiently monitoring the expected functional safety of the vehicle is achieved, thereby achieving the technical effect of improving the judgment accuracy of the expected functional safety events and expected functional safety levels of the vehicle and improving the safety of the whole vehicle, thereby solving the technical problem that the existing vehicle safety processing method has poor judgment accuracy for the expected functional safety events and expected functional safety levels, resulting in inaccurate safety measures provided and poor vehicle driving safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:
[0027] Figure 1 is a hardware structure block diagram of an optional vehicle terminal for a vehicle safety processing method according to an embodiment of the present invention;
[0028] Figure 2 is a flow chart of a vehicle safety processing method according to an embodiment of the present invention;
[0029] Figure 3 is a schematic diagram of a vehicle safety processing system according to an embodiment of the present invention;
[0030] Figure 4 4 is a structural block diagram of a vehicle safety processing device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0031] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0033] According to an embodiment of the present invention, a method embodiment of a vehicle safety processing method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0034] Figure 1 is a hardware structure block diagram of a vehicle terminal for a vehicle safety processing method according to an embodiment of the present invention, such as Figure 1 As shown, the vehicle terminal (or mobile device) may include one or more processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor (Microcontroller Unit, MCU) or a programmable logic device (FieldProgrammable Gate Array, FPGA)), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display device 110, an input and output device 108 (i.e., an IO device), a Universal Serial Bus (USB) port (which may be included as one of the ports of a computer bus, not shown in the figure), a network interface (not shown in the figure), a power supply (not shown in the figure) and / or a camera (not shown in the figure). It can be understood by those skilled in the art that Figure 1The structure shown is only for illustration and does not limit the structure of the vehicle terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.
[0035] It should be noted that the one or more processors 102 and / or other data processing circuits may be embodied in whole or in part as software, hardware, firmware or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be integrated in whole or in part into any of the other components in the vehicle terminal (or mobile device).
[0036] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the vehicle safety processing method in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned vehicle safety processing method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the vehicle terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0037] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the vehicle terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.
[0038] Under the above operating environment, the embodiment of the present invention provides the following Figure 2 The vehicle safety handling method shown, Figure 2 is a flow chart of a vehicle safety processing method according to an embodiment of the present invention. Figure 2 As shown, the method includes the following implementation steps:
[0039] Step S201, acquiring sensor data of the target vehicle;
[0040] Step S202, performing a safety assessment on the sensor data using the safety assessment model to obtain an assessment result, wherein the assessment result is used to determine an expected functional safety level and an expected functional safety event of the target vehicle;
[0041] Step S203, using the safety decision model to make a safety decision on the evaluation result to obtain a decision result, wherein the decision result is used to determine the target safety measures to resolve the expected functional safety incident.
[0042] In the present invention, the above sensor data can be collected by devices such as cameras, laser radars, millimeter wave radars, and inertial measurement units (IMU) sensors.
[0043] The above safety assessment model may be a model based on machine learning, and the safety assessment model may include a classification regression model and an inference model. The classification regression model may be used to determine the current expected functional safety level of the target vehicle, and the inference model may be used to determine the current expected functional safety events of the target vehicle. It should be noted here that the machine learning method used in the safety assessment model may include but is not limited to: decision tree, support vector machine, Bayes, convolutional neural network based on attention mechanism, or deep learning method with recurrent neural network suitable for text added.
[0044] The present invention uses a safety assessment model based on machine learning to perform classification, regression processing and reasoning on sensor data, so as to determine the current expected functional safety level and expected functional safety events of the target vehicle. Then, a safety decision model based on reinforcement learning can be used to determine the optimal safety measures (target safety measures) that can resolve the expected functional safety events under different expected functional safety levels in combination with the current expected functional safety events of the target vehicle.
[0045] In an embodiment of the present invention, the sensor data of the target vehicle is first obtained, and then the sensor data is security evaluated using a security evaluation model to obtain an evaluation result, wherein the evaluation result is used to determine the expected functional safety level and expected functional safety events of the target vehicle, and finally, the evaluation result is security-determined using a security decision model to obtain a decision result, wherein the decision result is used to determine the target safety measures for solving the expected functional safety events. By using a machine learning-based security evaluation model to accurately evaluate the full range of data related to the driving of the target vehicle (including sensor data and functional data and scene features obtained by processing the sensor data), the expected functional safety events and expected functional safety levels of the target vehicle are obtained, and the optimal safety measures for solving the expected functional safety events are given using a security decision model based on reinforcement learning, the purpose of accurately and efficiently monitoring the expected functional safety of the vehicle is achieved, thereby achieving the technical effect of improving the judgment accuracy of the expected functional safety events and expected functional safety levels of the vehicle and improving the safety of the whole vehicle, thereby solving the technical problem that the existing vehicle safety processing method has poor judgment accuracy for the expected functional safety events and expected functional safety levels, resulting in inaccurate safety measures provided and poor vehicle driving safety.
[0046] The above method of the embodiment of the present invention is further introduced below.
[0047] In an optional embodiment, the vehicle safety processing method further includes:
[0048] Step S241, clustering the sensor data according to the timestamp to obtain a clustering result;
[0049] Step S242, obtaining a first processing result of the clustering result by multiple functional units in the target vehicle, wherein different functional units are used to process and obtain different types of functional data;
[0050] Step S243, based on the map information and the first processing result, a scene analysis is performed on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
[0051] In the technical solution provided by the present invention, the inconsistency of the timestamps of various sensors can be regarded as a basic event, and the sensor data corresponding to the same timestamp can be clustered into one category to obtain a clustering result including multiple data categories (events). Then, in the example Figure 3 In the vehicle safety processing system shown in FIG. 1 , the clustering results are transmitted to various functional units (such as Figure 3As shown in the figure: positioning unit, perception unit, decision unit, planning unit, control unit), different functional units process the clustering results according to their functions and output different types of functional data. For example, the perception unit can output functional data such as lane lines, targets, road edges, and weather. Furthermore, all functional data related to the scene information are sensed and fused, and the scene characteristics of the current traffic scene of the target vehicle are determined based on the perception fusion results and the map information (such as the vehicle is at a crossroads).
[0052] In an optional embodiment, in step S242, the sensor data is security evaluated using the security evaluation model, and the evaluation result includes:
[0053] Step S2421, using a classification regression model to perform classification regression processing on the clustering result, the first processing result and the analysis result to obtain an expected functional safety level;
[0054] Step S2422: Use the reasoning model to perform event reasoning on the first processing result and the analysis result to obtain an expected functional safety event.
[0055] Among them, in step S2422, the first processing result and the analysis result are subjected to event reasoning using the reasoning model, and the expected functional safety events include:
[0056] Step S24221, inferring the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle;
[0057] Step S24222, inferring the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there is a target scene feature that meets the event triggering condition in the traffic scene;
[0058] Step S24223: Determine an expected functional safety event based on the first reasoning result and the second reasoning result.
[0059] Still Figure 3As shown, after obtaining functional data and scene features according to sensor data processing, the data set consisting of sensor data, functional data, and scene features is input into the safety assessment model based on machine learning. The safety assessment model first processes the expected functional safety level of the target vehicle according to the input data set and the classification and regression capabilities of the model itself. For example, when it is determined after classification and regression processing that extreme road conditions or weather conditions (such as icy and snowy roads, thunderstorms) occur in the traffic scene where the target vehicle is located, the output target vehicle's current expected functional safety level is level 1; when it is determined after classification and regression processing that the sensor data, functional data, or scene features of the target vehicle are wrong, the output target vehicle's current expected functional safety level is level 2; when it is determined after classification and regression processing that the target vehicle meets the triggering conditions of a special event (such as: the target vehicle is driving in a lane near the campus and there are pedestrians passing around), the output target vehicle's current expected functional safety level is level 3. It should be noted here that the lower the value corresponding to the current expected functional safety level of the target vehicle, the higher the current vehicle safety of the target vehicle can be represented.
[0060] In the present invention, the safety assessment model also includes a reasoning model with reasoning ability and grasps the reasoning relationship of expected functional safety events. The reasoning model can perform reasoning based on the input functional data to obtain a first reasoning result for determining the operation status of the driving function of the target vehicle, and the reasoning model can perform reasoning based on the input scene features to obtain a second reasoning result for determining whether there are target scene features that meet the event triggering conditions in the traffic scene. Furthermore, the thrust model outputs the current expected functional safety events of the target vehicle based on the first reasoning result and the second reasoning result (such as: in rainy and snowy weather, lane line detection is unclear; in strong light conditions, the target vehicle's perception of surrounding traffic participants is unclear).
[0061] In an optional embodiment, in step S203, a security decision is made on the evaluation result using a security decision model, and the decision result includes:
[0062] In step S231, the expected functional safety events and the expected functional safety levels are used as the states of the safety decision-making model, the vehicle decision is used as the action of the safety decision-making model, and a variety of safety assessment indicators are used as rewards of the safety decision-making model, and the target safety measures are obtained through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measures.
[0063] Still Figure 3As shown, the safety decision model in the present invention can be a model based on reinforcement learning. The safety decision model can take the expected functional safety level and expected functional safety output by the safety assessment model as the basic state, the whole vehicle decision as the action, and multiple safety assessment indicators as the comprehensive reward to form a Markov chain in reinforcement learning. The safety decision model can output the optimal safety measures (target safety measures) under the current state of the target vehicle under different expected functional safety levels in combination with the current expected functional safety events. Furthermore, the experience of each safety decision process can be stored in the experience playback database to continuously improve the decision-making ability of the safety decision model.
[0064] It should be noted that the various safety assessment indicators may include but are not limited to: confidence, which can be used to evaluate the reliability and accuracy of the safety decision-making model in outputting the target safety measures; decision rationality, which can be used to evaluate whether the target safety measures generated by the safety decision-making model are logical and consistent with the actual situation of the target vehicle; vehicle safety level, which can be used to evaluate the degree of change in the safety performance of the vehicle after the target safety measures are used to resolve expected functional safety incidents.
[0065] As an optional implementation, assuming that the expected functional safety event is the occurrence of extreme road conditions or weather conditions (such as icy and snowy roads, thunderstorms) in the traffic scenario in which the target vehicle is located, the corresponding target safety measure may be to wake up the driver of the target vehicle; assuming that the expected functional safety event is sensor data errors, functional data errors or scene feature errors of the target vehicle, the corresponding target safety measure may be to downgrade the driving function of the target vehicle; assuming that the expected functional safety event is that the target vehicle meets the triggering conditions of a special event (such as the target vehicle is driving in a lane near the campus and there are pedestrians passing by), the corresponding target safety measure may be to send a prompt message to the driver.
[0066] In an optional embodiment, the vehicle safety processing method further includes:
[0067] Step S251, receiving feedback results on target safety measures, wherein the feedback results are used to evaluate the ability of the target safety measures to resolve expected functional safety incidents;
[0068] Step S252, in response to determining that the target safety measure's ability to resolve the expected functional safety incident is lower than a preset level according to the feedback result, obtaining historical safety measures for resolving the expected functional safety incident;
[0069] Step S253, adjusting at least one safety assessment indicator of the current safety decision model according to the target safety measure and the historical safety measure to obtain an adjusted safety decision model;
[0070] Step S254, retraining the adjusted security decision model, and using the trained security decision model as the expected security decision model for security decision making.
[0071] In the present invention, in order to further improve the accuracy of the target safety measures output by the safety decision model, after the safety decision model outputs the target safety measures, the expected functional safety events, the expected functional safety levels, and the target safety measures can be provided to the driver of the target vehicle through an interactive device. The driver can provide feedback on the target safety measures output by the safety decision model based on the current actual situation of the target vehicle and the historical experience of handling expected functional safety events, and return the feedback results to the vehicle safety processing system. The vehicle safety processing system can adjust the safety assessment indicators of the current safety decision model according to the feedback results, and retrain and adjust the adjusted safety decision model until the trained safety decision model can output target safety measures that meet the expectations of the target vehicle, thereby obtaining the expected safety decision model, and then putting the expected safety decision model into use.
[0072] The vehicle safety processing method provided by the above embodiment of the present invention achieves the following technical effects:
[0073] (1) The vehicle safety processing system is applicable to the safety processing requirements of expected functional safety of different models and projects, and has strong generalization;
[0074] (2) The safety assessment model can accurately assess the target vehicle’s current expected functional safety level and expected functional safety events based on all-round data such as sensor data, functional data, and scenario characteristics related to the target vehicle’s driving. The safety decision model can perform reinforcement learning based on expected functional safety events and expected functional safety levels, and output the optimal safety measures for resolving expected functional safety events. The optimal safety measures can then be implemented to resolve the target vehicle’s current expected functional safety events and improve the overall safety of the target vehicle.
[0075] (3) Add model feedback function to make up for the shortcomings of artificial intelligence technologies such as machine learning, further improve the accuracy of target safety measures and enhance the safety of the entire vehicle.
[0076] In this embodiment, a vehicle safety processing device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, a "module" is a combination of software and / or hardware that can implement a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0077] Figure 4 is a structural block diagram of a vehicle safety processing device according to an embodiment of the present invention. Figure 4 As shown, the device comprises:
[0078] An acquisition module 401 is used to acquire sensor data of a target vehicle;
[0079] An evaluation module 402 is used to perform a safety evaluation on the sensor data using a safety evaluation model to obtain an evaluation result, wherein the evaluation result is used to determine an expected functional safety level and an expected functional safety event of the target vehicle;
[0080] The decision module 403 is used to make a safety decision on the evaluation result using the safety decision model to obtain a decision result, wherein the decision result is used to determine the target safety measures to solve the expected functional safety incident.
[0081] Optionally, the above-mentioned vehicle safety processing device also includes: a first processing module 404 (not shown in the figure), which is used to cluster the sensor data according to timestamps to obtain clustering results; obtain a first processing result of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process and obtain different types of functional data; based on the map information and the first processing result, perform a scene analysis on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
[0082] Optionally, the safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and a reasoning model. The classification regression model is used to determine the expected functional safety level, and the reasoning model is used to determine the expected functional safety events.
[0083] Optionally, the above-mentioned evaluation module 402 is also used to: use a classification regression model to perform classification regression processing on the clustering results, the first processing results and the analysis results to obtain an expected functional safety level; use an inference model to perform event inference on the first processing results and the analysis results to obtain an expected functional safety event.
[0084] Optionally, the above-mentioned evaluation module 402 is also used to: infer the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; infer the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there are target scene features in the traffic scene that meet the event triggering conditions; determine the expected functional safety event based on the first inference result and the second inference result.
[0085] Optionally, the above-mentioned decision module 403 is also used to: use the expected functional safety events and the expected functional safety levels as the states of the safety decision model, use the whole vehicle decision as the action of the safety decision model, and use multiple safety assessment indicators as rewards of the safety decision model, and obtain the target safety measures through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measures.
[0086] Optionally, the above-mentioned vehicle safety processing device also includes: a second processing module 405 (not shown in the figure), which is used to receive feedback results on target safety measures, wherein the feedback results are used to evaluate the target safety measures' ability to resolve expected functional safety events; in response to determining that the target safety measures' ability to resolve expected functional safety events is lower than a preset level based on the feedback results, obtaining historical safety measures used to resolve expected functional safety events; adjusting at least one safety assessment indicator of the current safety decision model based on the target safety measures and the historical safety measures to obtain an adjusted safety decision model; retraining the adjusted safety decision model, and using the trained safety decision model as the expected safety decision model for safety decision making.
[0087] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.
[0088] According to another aspect of an embodiment of the present invention, there is further provided a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, any one of the aforementioned vehicle safety processing methods is implemented.
[0089] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the aforementioned vehicle safety processing methods.
[0090] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:
[0091] Step S1, obtaining sensor data of the target vehicle;
[0092] Step S2, performing a safety assessment on the sensor data using the safety assessment model to obtain an assessment result, wherein the assessment result is used to determine the expected functional safety level and expected functional safety events of the target vehicle;
[0093] Step S3, using the safety decision model to make a safety decision on the evaluation results to obtain a decision result, wherein the decision result is used to determine the target safety measures to resolve the expected functional safety incidents.
[0094] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.
[0095] According to another aspect of an embodiment of the present invention, a vehicle is provided, comprising an on-board memory and an on-board processor, wherein a computer program is stored in the on-board memory, and the on-board processor is configured to run the computer program to execute any one of the aforementioned vehicle safety processing methods.
[0096] Optionally, in this embodiment, the vehicle-mounted processor may be configured to perform the following steps through a computer program:
[0097] Step S1, obtaining sensor data of the target vehicle;
[0098] Step S2, performing a safety assessment on the sensor data using the safety assessment model to obtain an assessment result, wherein the assessment result is used to determine the expected functional safety level and expected functional safety events of the target vehicle;
[0099] Step S3, using the safety decision model to make a safety decision on the evaluation results to obtain a decision result, wherein the decision result is used to determine the target safety measures to resolve the expected functional safety incidents.
[0100] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: clustering the sensor data according to timestamps to obtain clustering results; obtaining a first processing result of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process different types of functional data; based on the map information and the first processing result, performing a scene analysis on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
[0101] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: the safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and an inference model. The classification regression model is used to determine the expected functional safety level, and the inference model is used to determine the expected functional safety events.
[0102] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: use a classification regression model to perform classification regression processing on the clustering results, the first processing results and the analysis results to obtain an expected functional safety level; use an inference model to perform event inference on the first processing results and the analysis results to obtain expected functional safety events.
[0103] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: inferring the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; inferring the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there are target scene features in the traffic scene that meet the event triggering conditions; and determining the expected functional safety event based on the first inference result and the second inference result.
[0104] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: taking the expected functional safety events and the expected functional safety levels as the states of the safety decision model, taking the whole vehicle decision as the action of the safety decision model, and taking multiple safety assessment indicators as rewards of the safety decision model, and obtaining the target safety measures through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measures.
[0105] Optionally, in this embodiment, the above-mentioned on-board processor can also be configured to perform the following steps through a computer program: receiving feedback results on the target safety measures, wherein the feedback results are used to evaluate the target safety measures' ability to resolve expected functional safety events; in response to determining, based on the feedback results, that the target safety measures' ability to resolve expected functional safety events is lower than a preset level, obtaining historical safety measures for resolving expected functional safety events; adjusting at least one safety assessment indicator of the current safety decision model based on the target safety measures and the historical safety measures to obtain an adjusted safety decision model; retraining the adjusted safety decision model, and using the trained safety decision model as the expected safety decision model for safety decision making.
[0106] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiment and its optional implementation modes, and this embodiment will not be described in detail here.
[0107] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.
[0108] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0109] In the several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.
[0110] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0111] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0112] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program codes.
[0113] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A vehicle safety processing method, characterized in that: include: Obtain sensor data of the target vehicle; Performing a safety assessment on the sensor data using a safety assessment model to obtain an assessment result, wherein the assessment result is used to determine an expected functional safety level and an expected functional safety event of the target vehicle; A safety decision is made on the evaluation result using a safety decision model to obtain a decision result, wherein the decision result is used to determine a target safety measure to resolve the expected functional safety incident.
2. The vehicle safety processing method according to claim 1, characterized in that: The vehicle safety processing method further includes: Clustering the sensor data according to timestamps to obtain clustering results; Obtaining first processing results of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process and obtain different types of functional data; Based on the map information and the first processing result, a scene analysis is performed on the target vehicle to obtain an analysis result, wherein the analysis result is used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
3. The vehicle safety processing method according to claim 1, characterized in that: The safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and an inference model. The classification regression model is used to determine the expected functional safety level, and the inference model is used to determine the expected functional safety event.
4. The vehicle safety processing method according to claim 3, characterized in that: The sensor data is security evaluated using the security evaluation model, and the evaluation result obtained includes: Using the classification regression model to perform classification regression processing on the clustering result, the first processing result and the analysis result to obtain the expected functional safety level; The reasoning model is used to perform event reasoning on the first processing result and the analysis result to obtain the expected functional safety event.
5. The vehicle safety processing method according to claim 4, characterized in that: Using the reasoning model to perform event reasoning on the first processing result and the analysis result to obtain the expected functional safety event includes: Inferring the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; Inferring the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there is a target scene feature that meets the event triggering condition in the traffic scene; The expected functional safety event is determined according to the first reasoning result and the second reasoning result.
6. The vehicle safety processing method according to claim 1, characterized in that: The security decision model is used to make a security decision on the evaluation result, and the decision result obtained includes: The expected functional safety event and the expected functional safety level are used as the states of the safety decision model, the whole vehicle decision is used as the action of the safety decision model, and a plurality of safety evaluation indicators are used as rewards of the safety decision model, and the target safety measure is obtained through reinforcement learning, wherein different safety evaluation indicators are used to characterize different performances of the target safety measure.
7. The vehicle safety processing method according to claim 6, characterized in that: The vehicle safety processing method further includes: Receiving feedback results on the target safety measures, wherein the feedback results are used to evaluate the ability of the target safety measures to resolve the expected functional safety incidents; In response to determining, according to the feedback result, that the ability of the target safety measure to resolve the expected functional safety event is lower than a preset level, obtaining historical safety measures for resolving the expected functional safety event; Adjusting at least one safety assessment indicator of the current safety decision model according to the target safety measure and the historical safety measure to obtain an adjusted safety decision model; The adjusted security decision model is retrained, and the trained security decision model is used as the expected security decision model for security decision making.
8. A vehicle safety processing device, characterized in that: include: An acquisition module, used to acquire sensor data of a target vehicle; An evaluation module, configured to perform a safety evaluation on the sensor data using a safety evaluation model to obtain an evaluation result, wherein the evaluation result is used to determine an expected functional safety level and an expected functional safety event of the target vehicle; A decision module is used to make a safety decision on the evaluation result by using a safety decision model to obtain a decision result, wherein the decision result is used to determine a target safety measure to resolve the expected functional safety incident.
9. The vehicle safety processing device according to claim 8, characterized in that: The vehicle safety processing device also includes: a first processing module, used to cluster the sensor data according to timestamps to obtain clustering results; obtain a first processing result of the clustering results by multiple functional units in the target vehicle, wherein different functional units are used to process and obtain different types of functional data; based on map information and the first processing result, perform scene analysis on the target vehicle to obtain analysis results, wherein the analysis results are used to determine the scene characteristics of the traffic scene in which the target vehicle is located.
10. The vehicle safety processing device according to claim 8, characterized in that: The safety assessment model is a model based on machine learning, and the safety assessment model includes a classification regression model and an inference model. The classification regression model is used to determine the expected functional safety level, and the inference model is used to determine the expected functional safety event.
11. The vehicle safety processing device according to claim 10, characterized in that: The evaluation module is also used to: use the classification regression model to perform classification regression processing on the clustering results, the first processing results and the analysis results to obtain the expected functional safety level; use the reasoning model to perform event reasoning on the first processing results and the analysis results to obtain the expected functional safety event.
12. The vehicle safety processing device according to claim 11, characterized in that: The evaluation module is also used to: infer the first processing result to obtain a first inference result, wherein the first inference result is used to determine the operation status of the driving function of the target vehicle; infer the analysis result to obtain a second inference result, wherein the second inference result is used to determine whether there are target scene features in the traffic scene that meet the event triggering conditions; determine the expected functional safety event based on the first inference result and the second inference result.
13. The vehicle safety processing device according to claim 8, characterized in that: The decision module is also used to: use the expected functional safety event and the expected functional safety level as the state of the safety decision model, use the whole vehicle decision as the action of the safety decision model, and use multiple safety assessment indicators as rewards of the safety decision model, and obtain the target safety measure through reinforcement learning, wherein different safety assessment indicators are used to characterize the different performances of the target safety measure.
14. The vehicle safety processing device according to claim 13, characterized in that: The vehicle safety processing device also includes: a second processing module, used to receive feedback results on the target safety measures, wherein the feedback results are used to evaluate the target safety measures' ability to resolve the expected functional safety events; in response to determining, based on the feedback results, that the target safety measures' ability to resolve the expected functional safety events is lower than a preset level, obtaining historical safety measures used to resolve the expected functional safety events; adjusting at least one safety assessment indicator of a current safety decision model based on the target safety measures and the historical safety measures to obtain an adjusted safety decision model; retraining the adjusted safety decision model, and using the trained safety decision model as an expected safety decision model for safety decision making.
15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the vehicle safety processing method according to any one of claims 1 to 7 is implemented.
16. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is run, the device where the computer-readable storage medium is located is controlled to execute the vehicle safety processing method of any one of claims 1 to 7.
17. A vehicle, characterized in that: The vehicle safety processing method comprises an on-board memory and an on-board processor, wherein the on-board memory stores a computer program, and the on-board processor is configured to run the computer program to execute the vehicle safety processing method according to any one of claims 1 to 7.