Intelligent driving perception system, safety degradation method and equipment thereof and computer medium
By building an independent monitoring layer in the intelligent driving system for real-time verification and data optimization, the safety risks caused by perception algorithm errors are resolved, and safety degradation and continuous optimization are achieved, meeting automotive functional safety standards.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-28
- Publication Date
- 2026-04-21
AI Technical Summary
Existing intelligent driving perception systems lack real-time monitoring and effective countermeasures for perception algorithm errors, leading to safety risks and difficulty in continuous optimization, thus failing to meet automotive functional safety standards.
A monitoring layer independent of the perception algorithm layer is built in the system chip. The perception output is verified in real time by preset threshold range of physical rationality parameters of traffic participants, and differentiated vehicle control permission restriction strategies are executed. Scene source data is uploaded for model optimization.
It achieves ASIL B level requirements without increasing hardware costs, avoids excessive downgrading that affects the driving experience, ensures that the vehicle can quickly enter a safe state in high-risk scenarios, and optimizes perception capabilities through data closed-loop.
Smart Images
Figure CN121893980A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of vehicle control technology, and in particular to an intelligent driving perception system and its safety degradation method, device and computer medium. Background Technology
[0002] With the rapid development of intelligent driving technology, the perception system, as a core component of intelligent driving assistance systems, directly affects vehicle driving safety in terms of its safety and reliability. During operation, the perception system is susceptible to various factors such as sensor malfunctions, interference from complex environments, and algorithm design flaws, leading to perception errors.
[0003] In existing technologies, there are solutions for handling sensor faults. By polling the sensor register, sensor faults can be identified and auxiliary functions can be degraded. However, there is a lack of effective means to deal with perception errors caused by defects in the perception algorithm. Specifically, existing technologies have two major shortcomings: First, they do not monitor the erroneous or unreasonable identification results output by the perception algorithm in real time, which may lead to incorrect control strategies output by critical functions such as Automatic Emergency Braking (AEB), causing safety risks. Second, most systems only record error codes and do not simultaneously collect contextual data when the error occurs, such as the original sensor input and intermediate states of the algorithm. This makes it impossible to incorporate the error scenario into the database for model training, making it difficult to prevent the recurrence of similar problems.
[0004] Therefore, there is an urgent need for a technical solution that can identify and perceive algorithm errors in real time, dynamically execute safety degradation strategies, and support continuous system optimization, so as to reduce implementation costs and meet the requirements of automotive functional safety standards while ensuring the safety of intelligent driving assistance functions. Summary of the Invention
[0005] This disclosure aims to solve at least one of the technical problems existing in the prior art, and proposes an intelligent driving perception system and its safety degradation method, device and computer medium.
[0006] In a first aspect, embodiments of this disclosure provide a method for security downgrading of an intelligent driving perception system, including:
[0007] A monitoring layer, independent of the perception algorithm layer and communicating with the vehicle control unit, is built in the system chip.
[0008] Pre-set threshold ranges for physical rationality parameters for at least one type of traffic participant;
[0009] The monitoring layer obtains the recognition results output by the perception algorithm layer in real time and determines whether there are any abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter.
[0010] If the abnormal parameter exists, a corresponding differentiated vehicle control permission restriction strategy will be executed based on the degree of deviation of the abnormal parameter from the threshold range of the physical rationality parameter.
[0011] At the same time as executing the differentiated vehicle control permission restriction strategy, the scene source data at the time of the abnormal parameter determination is extracted and uploaded to the remote data storage unit for iterative optimization of the perception model.
[0012] In some embodiments, the system chip employs a three-layer software architecture, which includes:
[0013] The system comprises a low-level layer responsible for interacting with hardware devices, an algorithm layer responsible for executing perception algorithms and outputting the recognition results, and a monitoring layer independent of the algorithm layer and responsible for verifying the reasonableness of the recognition results.
[0014] The monitoring layer interacts with the algorithm layer and communicates with the vehicle's MCU module to trigger the execution of the differentiated vehicle control permission restriction strategy.
[0015] In some embodiments, the threshold range of the physical rationality parameter is dynamically configured based on the actual physical attributes of the traffic participant, the road scenario type, and regulatory requirements; wherein...
[0016] The traffic participants include dynamic targets and static obstacles;
[0017] The dynamic targets include pedestrians, non-motorized vehicles, and motorized vehicles;
[0018] The static obstacles include guardrails, curbs, and fixed obstacles;
[0019] The physical plausibility parameters include at least one of the following: the size of the traffic participant object, its speed of movement, and the curvature of its trajectory.
[0020] In some embodiments, the differentiated vehicle control permission restriction strategy specifically includes:
[0021] When the deviation is within the first range, the lateral control command output of the intelligent driving system is restricted, and the driver monitoring system is activated to monitor the driver's attention state and record snapshot data of the current scene.
[0022] When the deviation is in the second range, the lower limit of the second range is greater than the upper limit of the first range, the longitudinal control of the vehicle is transferred to the driver, and the safe parking mechanism is activated to guide the vehicle to park in a safe area.
[0023] The safe parking mechanism includes activating hazard warning lights and controlling the vehicle to decelerate at a constant speed.
[0024] In some embodiments, the degree of deviation is determined in the following manner:
[0025] Calculate the numerical difference between the abnormal parameter and the threshold range boundary value of the physical rationality parameter, then calculate the ratio of the numerical difference to the threshold range of the physical rationality parameter, and use the ratio as the degree of deviation;
[0026] The first interval is where the ratio is greater than 0 and less than or equal to 30%, and the second interval is where the ratio is greater than 30%.
[0027] In some embodiments, the scene source data includes:
[0028] Raw image data captured by the camera, point cloud data captured by the radar, vehicle dynamics status information, and current timestamp;
[0029] The vehicle's dynamic state information includes at least one of the following: vehicle speed, acceleration, steering angle, and brake pedal travel.
[0030] In some embodiments, the method further includes:
[0031] If a misjudgment occurs during the system development and verification phase or the actual operation phase, the system will respond to the externally input marking instructions by extracting the scene source data for the corresponding scenario and the currently effective threshold range of the physical rationality parameter.
[0032] The threshold range of the physical rationality parameter is adaptively adjusted according to the type of misjudgment: if the misjudgment type is that a downgrade should be triggered but was not triggered, the threshold range of the physical rationality parameter is narrowed.
[0033] If the misjudgment type is that a downgrade should not be triggered but is triggered incorrectly, then the threshold range of the physical rationality parameter is expanded;
[0034] The adjusted threshold range of the physical rationality parameter is updated to the system, and the corresponding scene source data is associated with the data and stored in the preset data storage unit for sample labeling during subsequent perception model training.
[0035] Secondly, embodiments of this disclosure provide an intelligent driving perception system, the system comprising:
[0036] The layered architecture unit is configured to build a monitoring layer in the system chip that is independent of the perception algorithm layer and communicates with the vehicle control unit.
[0037] The threshold configuration unit is configured to preset a threshold range for physical rationality parameters for at least one type of traffic participant.
[0038] The anomaly determination unit is configured to obtain the recognition results output by the perception algorithm layer in real time through the monitoring layer, and output an over-limit event when it is determined that there are abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter.
[0039] The hierarchical execution unit is configured to execute a corresponding differentiated vehicle control permission restriction strategy based on the degree of deviation of the over-limit event;
[0040] The data closed-loop unit is configured to automatically extract the scene source data at the time of the abnormal parameter determination and upload it to the remote data storage unit for iterative optimization of the perception model.
[0041] Thirdly, embodiments of this disclosure provide an electronic device, including:
[0042] One or more processors;
[0043] Memory, used to store one or more programs;
[0044] When one or more programs are executed by one or more processors, the one or more processors implement the safety degradation method of the intelligent driving perception system provided in the first aspect.
[0045] Fourthly, embodiments of this disclosure provide a computer-readable medium storing a computer program, which, when executed by a processor, implements the security degradation method for the intelligent driving perception system provided in the first aspect.
[0046] The safety degradation method for the intelligent driving perception system disclosed herein decouples functional safety monitoring logic from perception algorithms through software architecture design, meeting ASIL B level requirements without increasing hardware costs; transforms software algorithm defects into quantifiable and detectable over-limit events through physical rationality threshold verification; achieves refined permission control through deviation degree grading, avoiding excessive degradation that affects the driving experience while ensuring that the vehicle quickly enters a safe state in high-risk scenarios; and establishes a data closed loop between vehicle-side error detection and cloud-based model optimization through automatic extraction and uploading of scenario source data, driving the continuous evolution of perception capabilities. Attached Figure Description
[0047] Figure 1 A flowchart illustrating a safety downgrade method for an intelligent driving perception system provided in this embodiment of the disclosure;
[0048] Figure 2 A flowchart illustrating a differentiated vehicle control permission restriction strategy provided in an embodiment of this disclosure;
[0049] Figure 3A schematic flowchart of a threshold calibration optimization method provided in an embodiment of this disclosure;
[0050] Figure 4 This is a schematic diagram of the structure of an intelligent driving perception system provided in an embodiment of the present disclosure;
[0051] Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation
[0052] To enable those skilled in the art to better understand the technical solutions of this disclosure, exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments of this disclosure to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.
[0053] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.
[0054] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.
[0055] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Words such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.
[0056] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.
[0057] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information in this technical solution comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example, appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely identifying specific individuals.
[0058] As intelligent driving technology evolves from Level 2 to Level 3, the accuracy of the perception system's output, as the core module for environmental understanding, directly determines the safety of regulatory decisions.
[0059] However, existing technologies suffer from two main drawbacks: First, they lack a real-time rationality verification mechanism independent of the algorithm layer to address misidentifications caused by inherent flaws in the perception algorithm itself (such as misidentifying static obstacles as dynamic pedestrians, or distorted estimations of target size or speed). These software-level perception errors cannot be detected through traditional sensor hardware state polling methods, easily leading to unintended vehicle control behaviors in functions such as AEB and Lane Keeping Assist (LKA), posing safety risks at Automotive Safety Integrity Level (ASIL) C or higher. Second, most existing systems only log error codes, failing to fully capture the original sensor inputs, intermediate algorithm states, and vehicle dynamics context data at the time of the error. This results in the inability to reproduce the problem scenario and perform closed-loop training, making it difficult to effectively eradicate similar perception errors in subsequent algorithm iterations.
[0060] In existing technical solutions, the handling of perception errors by autonomous driving assistance systems is limited to sensor fault diagnosis. There is no monitoring layer for algorithm logic defects, nor is a complete closed-loop link from error detection to data feedback established, which limits the system's reliability and continuous optimization capabilities.
[0061] To address at least one or more of the aforementioned technical problems, embodiments of this disclosure provide a method and system for safety degradation of an intelligent driving perception system based on ASIL B functional safety level. This method constructs a monitoring layer independent of the perception algorithm layer within the SoC chip. By pre-setting threshold ranges for the physical rationality parameters of traffic participants, it performs real-time verification of the perception output. Based on the degree of parameter deviation, it executes differentiated vehicle control permission restriction strategies and simultaneously extracts scene source data and uploads it to the cloud, thereby achieving timely containment of perception errors and iterative optimization of the model.
[0062] Figure 1This is a flowchart illustrating a safety downgrade method for an intelligent driving perception system provided in an embodiment of this disclosure, as shown below. Figure 1 As shown, the method includes steps S1-S5:
[0063] Step S1: Construct a monitoring layer in the system chip that is independent of the perception algorithm layer and communicates with the vehicle control unit.
[0064] Step S2: Preset a threshold range for physical rationality parameters for at least one type of traffic participant.
[0065] Step S3: The monitoring layer obtains the recognition results output by the perception algorithm layer in real time and determines whether there are any abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter.
[0066] Step S4: If abnormal parameters exist, execute the corresponding differentiated vehicle control permission restriction strategy according to the degree of deviation of the abnormal parameters from the threshold range of physical rationality parameters.
[0067] Step S5: At the same time as executing the differentiated vehicle control permission restriction strategy, extract the scene source data at the time of abnormal parameter determination, and upload the scene source data to the remote data storage unit for iterative optimization of the perception model.
[0068] The safety downgrade method for the intelligent driving perception system provided in this disclosure decouples the functional safety monitoring logic from the perception algorithm through software architecture design, meeting ASIL B level requirements without increasing hardware costs; it transforms software algorithm defects into quantifiable and detectable over-limit events through physical rationality threshold verification; it achieves refined permission control through deviation degree grading, avoiding excessive downgrades that affect the driving experience while ensuring that the vehicle quickly enters a safe state in high-risk scenarios; and it establishes a data loop between vehicle-side error detection and cloud-based model optimization through automatic extraction and uploading of scene source data, promoting the continuous evolution of perception capabilities.
[0069] In some embodiments, the system chip employs a three-layer software architecture, which includes:
[0070] The system consists of a bottom layer responsible for interacting with hardware devices, an algorithm layer responsible for executing perception algorithms and outputting recognition results, and a monitoring layer independent of the algorithm layer that is responsible for verifying the reasonableness of the recognition results. The monitoring layer interacts with the algorithm layer and communicates with the vehicle's MCU module to trigger the execution of differentiated vehicle control permission restriction policies.
[0071] In other words, the embodiments disclosed herein rely on a three-layer software architecture deployed within the system chip. The bottom layer is responsible for data interaction with hardware devices such as cameras and millimeter-wave radar, completing raw data acquisition and drive management; the algorithm layer runs core algorithm modules such as perception, fusion, and control, outputting recognition results including information such as target type, size, speed, and trajectory; the monitoring layer, independent of the algorithm layer, establishes a direct communication connection with the vehicle's MCU module, specifically responsible for verifying the reasonableness of the recognition results and triggering degradation strategies. This layered design decouples functional safety monitoring logic from core algorithms. The monitoring layer does not undertake perception calculation responsibilities, only performing lightweight threshold comparisons and logical judgments, complying with the independence requirements of ASIL B level safety mechanisms in the ISO 26262 standard, and avoiding common-cause failures.
[0072] Specifically, the monitoring layer and the algorithm layer interact with each other through shared memory or inter-process communication mechanisms to obtain real-time perception output; at the same time, the monitoring layer is connected to the MCU through CAN bus or vehicle Ethernet, and can immediately send a degradation command once an anomaly is detected to ensure that the response delay is less than 50 milliseconds, meeting the time constraints of functional safety.
[0073] Furthermore, to effectively capture defects in perception algorithms, this disclosure pre-sets a threshold range for physical rationality parameters of traffic participants. This range is not a fixed constant, but is dynamically configured based on the actual physical attributes of the target, the current road scene type, and traffic regulations.
[0074] In some embodiments, the threshold range of the physical rationality parameter is dynamically configured based on the actual physical attributes of the traffic participant object, the road scene type, and regulatory requirements; wherein, the traffic participant object includes dynamic targets and static obstacles; dynamic targets include pedestrians, non-motorized vehicles, and motorized vehicles; static obstacles include guardrails, curbs, and fixed obstacles; and the physical rationality parameter includes at least one of the following: the size of the traffic participant object, its moving speed, and the curvature of its trajectory.
[0075] The aforementioned threshold range is dynamically configured based on object attributes, scene type, and regulatory requirements, which improves the adaptability and accuracy of the threshold, reduces the probability of false positives and false negatives, and provides a reliable basis for anomaly identification.
[0076] After receiving the recognition results from the perception algorithm layer, the monitoring layer immediately initiates a rationality verification process. For each identified traffic participant, its physical rationality parameters are extracted and compared with a preset threshold range. If the parameters fall within the threshold range, the perception result is considered reliable, and the recognition result is normally transmitted to the traffic control module; if the parameters exceed the threshold range, they are determined to be abnormal parameters, triggering an over-limit event. After determining the abnormal parameters, the monitoring layer further calculates their deviation degree to determine the severity level of the subsequent degradation strategy.
[0077] In some embodiments, the degree of deviation is determined in the following ways:
[0078] Calculate the numerical difference between the abnormal parameter and the threshold range boundary value of the physical rationality parameter, and then calculate the ratio of the numerical difference to the threshold range of the physical rationality parameter. Use the ratio as the degree of deviation; the first interval is a ratio greater than 0 and less than or equal to 30%, and the second interval is a ratio greater than 30%.
[0079] Furthermore, the monitoring layer implements differentiated vehicle control permission restriction strategies based on the different ranges in which the deviation occurs.
[0080] Figure 2 This is a flowchart illustrating a differentiated vehicle control permission restriction strategy provided in an embodiment of the present disclosure. In some embodiments, such as... Figure 4 As shown, the differentiated vehicle control permission restriction strategy in step S4 specifically includes steps S41-S43:
[0081] Step S41: When the deviation is in the first range, restrict the output of lateral control commands of the intelligent driving system, and at the same time activate the driver monitoring system to monitor the driver's attention state and record the snapshot data of the current scene.
[0082] Step S42: When the deviation is in the second range, the lower limit of the second range is greater than the upper limit of the first range. The longitudinal control of the vehicle is transferred to the driver, and the safe parking mechanism is activated to guide the vehicle to stop in a safe area.
[0083] Step S43, the safe parking mechanism includes turning on the hazard warning lights and controlling the vehicle to decelerate at a constant speed.
[0084] Since different degrees of deviation correspond to different levels of safety risk, differentiated strategies can reduce unnecessary interference with the driving experience while ensuring safety. At the same time, the clearly defined method for calculating the degree of deviation ensures the consistency and accuracy of the downgrade strategy execution. This allows for precise matching of risk levels to execute downgrade operations, effectively avoiding dangerous maneuvers such as unexpected deceleration and steering, reducing the probability of unintended vehicle control by the intelligent driving assistance system, and improving the safety and stability of the system operation.
[0085] At the same time as implementing differentiated vehicle control permission restriction policies, the monitoring layer automatically extracts the scene source data at the time of abnormal parameter determination and uploads it to the remote data storage unit via the vehicle Ethernet network.
[0086] In some embodiments, the scene source data includes:
[0087] The system includes raw image data captured by the camera, point cloud data captured by the radar, vehicle dynamics information, and the current timestamp; the vehicle dynamics information includes at least one of the following: vehicle speed, acceleration, steering angle, and brake pedal travel.
[0088] By uploading these scene source data to the remote data storage unit through the vehicle's existing Ethernet network, without the need for additional communication hardware, the complete collection and low-cost transmission of scene data is achieved, laying a data foundation for subsequent system optimization.
[0089] Furthermore, the embodiments disclosed herein also include a threshold calibration optimization step, which supports the continuous evolution of the system during the development and verification phase and the actual operation phase.
[0090] Figure 3 This is a flowchart illustrating a threshold calibration optimization method provided in an embodiment of the present disclosure. In some embodiments, such as... Figure 3 As shown, the threshold calibration optimization steps include steps S01-S04:
[0091] Step S01: In the system development and verification phase or the actual operation phase, if a misjudgment occurs, the system responds to the externally input marking instruction and extracts the scene source data of the corresponding scenario and the threshold range of the currently effective physical rationality parameters.
[0092] Step S02: Adaptively adjust the threshold range of the physical rationality parameter according to the type of misjudgment: If the misjudgment type is that a downgrade should be triggered but was not triggered, then narrow the threshold range of the physical rationality parameter.
[0093] Step S03: If the misjudgment type is that the downgrade should not be triggered but is triggered incorrectly, then expand the threshold range of the physical rationality parameter.
[0094] Step S04: Update the adjusted physical rationality parameter threshold range to the system and associate the corresponding scene source data to the preset data storage unit for sample labeling during subsequent perception model training.
[0095] Specifically, when a misjudgment occurs, in response to the driver's manual marking command, the system extracts the scene source data for the corresponding scenario and the currently effective threshold range of physical rationality parameters. If the misjudgment type is that a downgrade should have been triggered but was not (i.e., the erroneous function was not downgraded, and the driver believes there is a collision risk), the corresponding threshold range of physical rationality parameters is narrowed. If the misjudgment type is that a downgrade should not have been triggered but was triggered incorrectly (i.e., the erroneous function downgrade was implemented, and the driver believes there is no need to exit the function), the corresponding threshold range of physical rationality parameters is expanded. If the function downgrade judgment is correct, the saved snapshot data is directly incorporated into the perception model training. The adjusted threshold range is updated to the system, and the corresponding scene source data is stored in a preset data storage unit for sample labeling during subsequent perception model training.
[0096] The threshold calibration optimization method described above continuously optimizes the threshold range and perception model through human feedback in actual use scenarios, solves edge scenario problems not covered in the development phase, realizes dynamic optimization of the threshold range and perception model, ensures that the system can adapt to more complex scenarios, avoids the recurrence of similar errors, and improves the robustness and adaptability of the system.
[0097] Based on the same inventive concept, this disclosure also provides an intelligent driving perception system. Figure 4 This is a schematic diagram of the structure of an intelligent driving perception system provided in an embodiment of the present disclosure, such as... Figure 4 As shown, the above system includes:
[0098] The layered architecture unit 10 is configured to build a monitoring layer in the system chip that is independent of the perception algorithm layer and communicates with the vehicle control unit.
[0099] The threshold configuration unit 20 is configured to preset a threshold range for physical rationality parameters for at least one type of traffic participant.
[0100] The anomaly determination unit 30 is configured to obtain the recognition results output by the perception algorithm layer in real time through the monitoring layer, and output an over-limit event when there are abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter.
[0101] The hierarchical execution unit 40 is configured to execute a corresponding differentiated vehicle control permission restriction strategy based on the degree of deviation of the over-limit event;
[0102] The data closed-loop unit 50 is configured to automatically extract scene source data at the time of abnormal parameter determination and upload it to the remote data storage unit for iterative optimization of the perception model.
[0103] The technical effects of the intelligent driving perception system correspond to the methods described above. Through the collaborative work of each unit, it achieves fully automated processing of perception error identification, safety degradation, and system optimization, ensuring the safety, reliability, and scalability of the system operation. These will not be elaborated further here.
[0104] Based on the same inventive concept, this disclosure also provides an electronic device. Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of this disclosure. Figure 5As shown, this disclosure provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, enable the one or more processors to implement a security degradation method as described in any of the intelligent driving perception systems in the above embodiments; the one or more I / O interfaces 103 are connected between the processor and the memory, configured to enable information interaction between the processor and the memory.
[0105] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).
[0106] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.
[0107] In some embodiments, the one or more processors 101 include a field-programmable gate array.
[0108] This disclosure also provides a computer-readable medium. The computer-readable medium stores a computer program, which, when executed by a processor, implements the steps in the security degradation method of any of the intelligent driving perception systems described above. The computer-readable storage medium may be volatile or non-volatile.
[0109] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the security degradation method of the intelligent driving perception system described above.
[0110] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).
[0111] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.
[0112] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.
[0113] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.
[0114] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.
[0115] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.
[0116] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.
[0117] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.
[0118] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.
[0119] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.
Claims
1. A method for safety downgrading of an intelligent driving perception system, characterized in that, include: A monitoring layer, independent of the perception algorithm layer and communicating with the vehicle control unit, is built in the system chip. Pre-set threshold ranges for physical rationality parameters for at least one type of traffic participant; The monitoring layer obtains the recognition results output by the perception algorithm layer in real time and determines whether there are any abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter. If the abnormal parameter exists, a corresponding differentiated vehicle control permission restriction strategy will be executed based on the degree of deviation of the abnormal parameter from the threshold range of the physical rationality parameter. At the same time as executing the differentiated vehicle control permission restriction strategy, the scene source data at the time of the abnormal parameter determination is extracted and uploaded to the remote data storage unit for iterative optimization of the perception model.
2. The method according to claim 1, characterized in that, The system chip adopts a three-layer software architecture, which includes: The system comprises a low-level layer responsible for interacting with hardware devices, an algorithm layer responsible for executing perception algorithms and outputting the recognition results, and a monitoring layer independent of the algorithm layer and responsible for verifying the reasonableness of the recognition results. The monitoring layer interacts with the algorithm layer and communicates with the vehicle's MCU module to trigger the execution of the differentiated vehicle control permission restriction strategy.
3. The method according to claim 2, characterized in that, The threshold range of the physical rationality parameter is dynamically configured based on the actual physical attributes of traffic participants, road scenario type, and regulatory requirements; among which... The traffic participants include dynamic targets and static obstacles; The dynamic targets include pedestrians, non-motorized vehicles, and motorized vehicles; The static obstacles include guardrails, curbs, and fixed obstacles; The physical plausibility parameters include at least one of the following: the size of the traffic participant object, its speed of movement, and the curvature of its trajectory.
4. The method according to claim 1, characterized in that, The differentiated vehicle control permission restriction strategy specifically includes: When the deviation is within the first range, the lateral control command output of the intelligent driving system is restricted, and the driver monitoring system is activated to monitor the driver's attention state and record snapshot data of the current scene. When the deviation is in the second range, the lower limit of the second range is greater than the upper limit of the first range, the longitudinal control of the vehicle is transferred to the driver, and the safe parking mechanism is activated to guide the vehicle to park in a safe area. The safe parking mechanism includes activating hazard warning lights and controlling the vehicle to decelerate at a constant speed.
5. The method according to claim 4, characterized in that, The degree of deviation is determined in the following way: Calculate the numerical difference between the abnormal parameter and the threshold range boundary value of the physical rationality parameter, then calculate the ratio of the numerical difference to the threshold range of the physical rationality parameter, and use the ratio as the degree of deviation; The first interval is where the ratio is greater than 0 and less than or equal to 30%, and the second interval is where the ratio is greater than 30%.
6. The method according to claim 1, characterized in that, The scene source data includes: Raw image data captured by the camera, point cloud data captured by the radar, vehicle dynamics status information, and current timestamp; The vehicle's dynamic state information includes at least one of the following: vehicle speed, acceleration, steering angle, and brake pedal travel.
7. The method according to claim 2, characterized in that, The method further includes: In the system development and verification phase or the actual operation phase, if a misjudgment occurs, in response to the externally input marking instruction, the corresponding scenario source data and the currently effective physical rationality parameter threshold range are extracted. The threshold range of the physical rationality parameter is adaptively adjusted according to the type of misjudgment: if the misjudgment type is that a downgrade should be triggered but was not triggered, the threshold range of the physical rationality parameter is narrowed. If the misjudgment type is that a downgrade should not be triggered but is triggered incorrectly, then the threshold range of the physical rationality parameter is expanded; The adjusted threshold range of the physical rationality parameter is updated to the system, and the corresponding scene source data is associated with it and stored in the preset data storage unit for sample labeling during subsequent perception model training.
8. An intelligent driving perception system, characterized in that, The system includes: The layered architecture unit is configured to build a monitoring layer in the system chip that is independent of the perception algorithm layer and communicates with the vehicle control unit. The threshold configuration unit is configured to preset a threshold range for physical rationality parameters for at least one type of traffic participant. The anomaly determination unit is configured to obtain the recognition results output by the perception algorithm layer in real time through the monitoring layer, and output an over-limit event when it is determined that there are abnormal parameters in the recognition results that exceed the threshold range of the physical rationality parameter. The hierarchical execution unit is configured to execute a corresponding differentiated vehicle control permission restriction strategy based on the degree of deviation of the over-limit event; The data closed-loop unit is configured to automatically extract the scene source data at the time of the abnormal parameter determination and upload it to the remote data storage unit for iterative optimization of the perception model.
9. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the security degradation method of the intelligent driving perception system as described in any one of claims 1 to 7.
10. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps in the security degradation method of the intelligent driving perception system as described in any one of claims 1 to 7.