Performance detection method for autonomous driving system, electronic device, and program product

Abstract

The present disclosure provides a performance detection method of an automatic driving system, an electronic device and a program product, and relates to the technical field of intelligent driving. The method comprises the following steps: acquiring driving control information planned by an automatic driving system for a vehicle; acquiring actual response information of the vehicle to the driving control information; determining a driving deviation of the automatic driving system according to the driving control information and the actual response information; and determining a performance detection result of the automatic driving system according to the driving deviation, wherein the performance detection result is used to indicate whether the automatic driving system is abnormal. The present disclosure can accurately monitor the performance state of the automatic driving system.

Description

Technical Field

[0001] This disclosure relates to the field of autonomous driving technology, specifically to a performance testing method, electronic device, and software product for an autonomous driving system. Background Technology

[0002] In the field of safety monitoring for autonomous driving systems, with the implementation of relevant national standards, there is a clear requirement to monitor the continuous operational capability of vehicles equipped with autonomous driving systems and to have a management mechanism to allow for the timely retirement of the system. Therefore, a technical solution is needed that can monitor the performance status or lifespan degradation of various components involved in vehicle operation (such as steering and braking actuators) in real-time or near real-time, so as to provide timely warnings or take safety measures when component performance is insufficient. Summary of the Invention

[0003] In view of this, embodiments of the present disclosure provide a performance detection method, electronic device, and program product for an autonomous driving system, which can accurately monitor the performance status of the autonomous driving system.

[0004] In a first aspect, this disclosure provides a performance testing method for an autonomous driving system, including: acquiring driving control information planned by the autonomous driving system for the vehicle; Obtain the actual response information of the vehicle to the driving control information; The driving deviation of the autonomous driving system is determined based on the driving control information and the actual response information. The performance test result of the autonomous driving system is determined based on the driving deviation, and the performance test result is used to indicate whether the autonomous driving system is abnormal.

[0005] Secondly, this disclosure provides an electronic device, including: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores at least one computer program that can be executed by the at least one processor, the at least one computer program being executed by the at least one processor to enable the at least one processor to perform the performance testing method of the autonomous driving system as described in the first aspect.

[0006] Thirdly, this disclosure provides a computer program product, which includes a computer program that, when run in a processor, implements the performance detection method for the autonomous driving system described in the first aspect.

[0007] The embodiments provided in this disclosure determine the system's driving deviation by acquiring and comparing the driving control information planned by the autonomous driving system with the vehicle's actual response to this information, and judge whether the system performance is abnormal based on this deviation. This scheme utilizes the inherent correlation between system planning instructions and the vehicle's actual response to construct a feedback verification mechanism for indirectly evaluating execution performance. Therefore, it can effectively reveal control deviations caused by actuator attenuation or failure without relying on additional dedicated sensors, achieving continuous health monitoring of the core execution link of the autonomous driving system. This provides timely early warning for addressing potential safety risks caused by performance degradation, improving the overall reliability and operational safety of the autonomous driving system. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0009] Figure 1 The diagram shown is a schematic flowchart of the performance testing method for an autonomous driving system in an embodiment of this disclosure.

[0010] Figure 2 The diagram shown is a schematic representation of the lateral path between vehicles in an embodiment of this disclosure.

[0011] Figure 3 The diagram shown is a schematic representation of the lateral path between vehicles and pedestrians in an embodiment of this disclosure.

[0012] Figure 4 The diagram shown is a schematic diagram of the longitudinal path between vehicles and pedestrians in an embodiment of this disclosure.

[0013] Figure 5 The diagram shown is a schematic representation of the longitudinal path between vehicles in an embodiment of this disclosure.

[0014] Figure 6 The diagram shown is a schematic diagram of the deviation detection process in an embodiment of this disclosure.

[0015] Figure 7 The diagram shown is a schematic flowchart of a two-level fault early warning method in an embodiment of this disclosure.

[0016] Figure 8 The diagram shown is a structural schematic of an electronic device in an embodiment of this disclosure. Detailed Implementation

[0017] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0018] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0019] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0020] 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.

[0021] 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.

[0022] In the field of autonomous driving technology, to ensure the accuracy and safety of vehicle control, it is typically necessary to monitor the health status of the control components (such as steering and braking systems). A common approach in related technologies is to rely on built-in sensors to directly detect the physical state of these components (such as temperature and voltage), or to read preset fault codes through diagnostic protocols. However, this approach is less than ideal when used to assess the gradual performance degradation caused by wear and aging during long-term service. A fundamental contradiction lies in the fact that, in order to optimize the reliability of detecting explicit, sudden faults, this approach inevitably compromises the ability to perceive slow performance degradation processes, and may even mask potential safety risks. Specifically, during daily vehicle operation, the steering mechanism's response may gradually slow down, or there may be a slight delay in the establishment of braking force, but these changes have not yet triggered any sensor thresholds or internal diagnostic fault codes. Due to the lack of quantitative perception of these gradual deviations, unpredictable deviations will gradually occur between the vehicle's actual driving trajectory and the trajectory expected by the control system, and this risk is invisible under existing solutions.

[0023] Through in-depth analysis, the inventors discovered that the root causes of the aforementioned contradictions are multifaceted: At the component level, the performance degradation of many mechanical actuators (such as steering racks and brake calipers) is difficult to measure directly and in real-time using a limited number of built-in sensors. At the system level, there is a lack of a continuous, quantitative comparison and feedback loop between the control commands issued by the autonomous driving controller and the vehicle's final execution effect; existing compensation mechanisms in vehicle control loops may mask initial deviations within a single control cycle. Furthermore, at the data level, related technical solutions focus on the absolute state of the components themselves, failing to construct an indirect observational indicator for evaluating the overall performance of the system.

[0024] To overcome the aforementioned contradictions, this disclosure proposes a performance detection method for an autonomous driving system. Its core concept lies in acquiring driving control information issued by the autonomous driving system and simultaneously acquiring the vehicle's actual response information to this control information. The driving deviation is then determined based on the comparison between the two, and the system's performance status is judged according to this deviation. This concept improves the system performance evaluation process, thereby effectively achieving real-time, quantitative monitoring of gradual performance degradation or indirect faults in the system without relying on direct fault diagnosis within the execution components. Specifically, it provides an indirect detection method based on "command-response" comparison feedback to solve the problem that related technologies cannot quantitatively perceive driving trajectory deviations caused by the slow degradation of autonomous driving system performance in real time, achieving continuous evaluation and safety warning of the operational health status of in-service autonomous driving systems.

[0025] The performance testing method for an autonomous driving system provided in this disclosure is applied to a typical system architecture integrated into a vehicle. Exemplarily, this typical system architecture mainly includes a high-level autonomous driving controller, an onboard perception system, a chassis execution system, and a performance testing module responsible for implementing this testing method. The high-level autonomous driving controller makes decisions and plans based on environmental information provided by the onboard perception system, generates driving control commands, and sends them to the chassis execution system. The chassis execution system includes components such as steering actuators and brake actuators, which respond to control commands and drive vehicle movements, while simultaneously feeding back the execution status to the relevant domain controllers through their own sensors. The performance testing module can be integrated with the high-level autonomous driving controller or function as an independent functional unit. It is configured to acquire driving control information from the high-level autonomous driving controller, acquire actual response information from the onboard perception system and / or the chassis execution system, and perform subsequent comparison, judgment, and processing logic.

[0026] Figure 1 The diagram shown is a schematic flowchart of an autonomous driving system performance testing method provided according to an embodiment of this disclosure. This control method can be executed by the aforementioned performance testing module and includes the following:

[0027] Step 101: Obtain the driving control information planned by the autonomous driving system for the vehicle.

[0028] Driving control information broadly refers to any instructions or data generated by an autonomous driving system to control the vehicle's motion state or path. Examples include, but are not limited to: target steering wheel angle, target yaw rate, target acceleration, target deceleration, a sequence of expected trajectory points, or avoidance instructions containing reference targets and their expected relative positions. Driving control information is generated by the decision-making and planning functions of the autonomous driving system and consists of any digital or analog instructions used to directly or indirectly drive changes in the vehicle's motion state. Its core essence lies in representing the system's "expectation" or "requirement" for vehicle behavior at a specific moment.

[0029] One specific implementation method is to read control command frames issued by the higher-level autonomous driving controller in real time via the controller area network bus. For example, the command frame might contain data fields identified as "target left front wheel steering angle: 2.5 degrees" or "target deceleration: -3.0 m / s²". This method allows for direct and accurate acquisition of the system's original planning output, ensuring the authority of the detection benchmark. It is understandable that, in addition to this, driving control information can be acquired through various other methods. These include, but are not limited to: intercepting data from the output interface of the autonomous driving system's planning and decision-making submodule; listening to relevant control messages on the communication bus between the autonomous driving domain controller and the chassis domain controller; or establishing a data hook function within the autonomous driving system to immediately copy and send a copy of the control command to the performance detection module after its generation.

[0030] In a specific example, the advanced autonomous driving controller generates lateral and longitudinal control commands every control cycle (e.g., 10 milliseconds). A lateral control command could be, for example, a target front wheel steering angle value in degrees. A longitudinal control command could be, for example, a target acceleration or deceleration value in meters per square second. The performance monitoring module periodically reads these command values ​​and their corresponding timestamps via an internal data bus or a dedicated hardware interface. Understandably, the specific content and format of the control commands can vary depending on the design of different autonomous driving systems; for example, they might be given in the form of torque, yaw rate, etc., but their essence is the control setpoints for vehicle motion.

[0031] Step 102: Obtain the vehicle's actual response information to the driving control information.

[0032] In this embodiment of the disclosure, actual response information refers to any observable and measurable physical quantity or derived data that can directly or indirectly reflect changes in the vehicle's motion state or spatial position after being acted upon by driving control information. Actual response information is the manifestation of the "execution result" of driving control information in the real physical world. For example, actual response information includes, but is not limited to: the actual steering wheel angle measured by the steering angle sensor, the actual yaw rate or actual deceleration measured by the vehicle inertial measurement unit, the actual position of the vehicle relative to lane lines or other static references identified by the onboard perception system, or the actual relative distance and orientation of the vehicle to dynamic targets.

[0033] One specific implementation method is to obtain feedback values ​​from the steering and braking systems through the chassis domain controller. For example, the actual steering wheel angle measured by an angle sensor can be read from the steering controller; the fused actual longitudinal deceleration of the vehicle can be obtained from the vehicle controller or inertial measurement unit. This method allows for the acquisition of actuator feedback directly corresponding to driving control information, and the actual response information includes actuator feedback information, facilitating one-to-one comparison and analysis. It is understood that the acquisition of actual response information can also be achieved through various other means. For example, it may include, but is not limited to: using vehicle trajectory data output by a combination of an onboard GPS and inertial navigation system; using lane line information perceived by a forward-facing camera to calculate the vehicle's lateral offset relative to the lane centerline as the actual lateral position response through geometric relationships; and using millimeter-wave radar to continuously track a target vehicle ahead and calculate the actual distance between the vehicle and the vehicle ahead as the response to the longitudinal following effect.

[0034] In a specific example, for steering control, the angle sensor inside the steering actuator measures the actual steering wheel angle in real time and sends it to the chassis domain controller and the higher-order autonomous driving controller via the CAN bus. For braking control, the vehicle's inertial measurement unit measures the vehicle's actual longitudinal acceleration / deceleration, while the braking system's pressure sensors measure the wheel cylinder braking pressure. This data is processed by the chassis domain controller's sensor fusion algorithm to obtain a more accurate estimate of the actual deceleration. The performance monitoring module also periodically obtains these actual feedback values ​​and their corresponding timestamps from the bus or through an internal interface. It is understandable that the sources of actual response information can be diverse and may involve some processing delay; therefore, alignment based on this delay is necessary for comparison.

[0035] Step 103: Determine the driving deviation of the autonomous driving system based on the driving control information and the actual response information.

[0036] In this embodiment of the disclosure, "driving deviation" refers to a quantitative value representing the degree of deviation of the execution effect of the autonomous driving system, calculated by comparing driving control information (or the expected state derived therefrom) with actual response information (or the actual state represented by it). It can be an instantaneous error, a statistic (such as mean, variance), a cumulative quantity, or an equivalent spatial deviation after model conversion. For example, it may include, but is not limited to: lateral path deviation and / or longitudinal path deviation obtained by comparing the actual driving path with the expected driving path point by point or segment; or, equivalent lateral displacement deviation or longitudinal displacement deviation calculated by back-calculating through a preset vehicle dynamics model based on the response delay of the control command.

[0037] As a specific implementation method, the difference between the acquired target turning angle and the actual turning angle at the same moment can be used to obtain the turning angle deviation; the difference between the target deceleration and the actual deceleration can be used to obtain the deceleration deviation. Using this method, the numerical error of the control command at the execution level can be directly calculated, which is simple and effective. It should be noted that determining the driving deviation can be achieved through various comparison logics. For example, it can include, but is not limited to: comparing the time difference between the moment the control command is issued and the moment the actuator feedback reaches stability to obtain the response delay deviation; comparing the spatial distance difference between the vehicle's future position predicted based on the control command and the vehicle's actual future position obtained through environmental perception to obtain the path tracking deviation; or, inputting the control command and feedback value into a preset vehicle dynamics model, the model outputs a predicted vehicle displacement change, and then comparing this predicted amount with the actual displacement change obtained through perception to obtain a comprehensive deviation index.

[0038] In a specific example, after obtaining the target steering angle θ_target and the actual steering angle θ_actual, the performance detection module calculates the instantaneous steering angle deviation Δθ = |θ_target - θ_actual|. To assess the cumulative effect of the deviation, this deviation can be used to estimate, through a preset vehicle kinematics model and combined with the current vehicle speed v, the additional lateral displacement deviation ΔS_lat that this steering angle deviation may cause to the vehicle within a specific time period Δt. For example, a simplified formula ΔS_lat ≈ v can be used. Δt The angular error is estimated using tan(Δθ) / L (where L is the wheelbase), thus converting it into a more intuitive path space error. For longitudinal control, the instantaneous deceleration deviation Δa = |a_target - a_actual| is calculated, and the resulting longitudinal displacement deviation ΔS_lon can be estimated similarly. Understandably, the model used can be a more complex dynamic model; the purpose of the estimation is to map the execution-level error to the overall vehicle driving performance, making the deviation more meaningful for safety.

[0039] Step 104: Determine the performance test results of the autonomous driving system based on the driving deviation. The performance test results are used to indicate whether the autonomous driving system is abnormal.

[0040] In this embodiment of the disclosure, the performance detection result generally refers to a judgment conclusion on the control performance status of the autonomous driving system. It is used to indicate whether the autonomous driving system is in a normal working state, or whether its performance level is acceptable, and is a final judgment output to indicate whether the system needs maintenance, downgrading, or shutdown. For example, it may include, but is not limited to: a binary status indication output by comparing the driving deviation with a preset threshold, such as "normal" or "abnormal"; or, a multi-level fault level indication distinguished according to the different numerical ranges in which the deviation falls, such as "Class A maintenance fault" or "Class S risk fault".

[0041] As a specific implementation method, thresholds can be set for the lateral displacement deviation ΔS_lat and longitudinal displacement deviation ΔS_lon calculated in step 103. For example, the lateral deviation threshold can be set to 60 cm and the longitudinal deviation threshold to 100 cm. If ΔS_lat is greater than 60 cm or ΔS_lon is greater than 100 cm, the performance test result is determined to be "abnormal"; otherwise, the result is "normal". This method can provide a clear judgment, which facilitates triggering subsequent fault handling procedures. It is understood that determining the performance test result is not limited to simple binary judgment. For example, it may include, but is not limited to: outputting a continuous performance score (e.g., 0-100 points) based on the magnitude of the deviation value; mapping the deviation to multiple fault levels (e.g., "minor", "moderate", "severe"); or, combining historical deviation data, determining whether the performance is in a degradation process through trend analysis and outputting a warning level.

[0042] In a specific example, the performance detection module maintains a fault status register. When the calculated ΔS_lat consistently exceeds 20 cm but is less than 60 cm, or ΔS_lon consistently exceeds 50 cm but is less than 100 cm for a certain period of time (e.g., 5 seconds), the flag representing "functional abnormality repair fault" in the register is set. This result corresponds to a Class A error, indicating a performance degradation requiring repair, but allowing the vehicle to limp home. When ΔS_lat exceeds 60 cm at once, or ΔS_lon exceeds 100 cm at once, the flag representing "functional abnormality risk fault" is set directly. This result corresponds to a Class S error, requiring immediate and more stringent measures. The detection results can be sent to the vehicle's instrument panel for display via bus messages, or passed to the autonomous driving controller to trigger a degrade strategy. It is understood that the specific threshold values ​​can be calibrated and adjusted according to vehicle model, regulatory requirements, and road type.

[0043] The performance testing methods described above for autonomous driving systems collectively constitute a complete "command-response" comparison and evaluation closed loop. Information is acquired from both the system input and vehicle output, and through comparison, internal performance issues that are difficult to observe directly are transformed into quantifiable driving deviations. State decisions are then made based on these quantified deviations. The synergy of these features enables autonomous driving systems to indirectly and effectively detect control deviations caused by performance degradation, even in the absence of direct lifespan sensors for actuators. This collectively solves the specific technical problem of real-time monitoring and early warning of in-service performance degradation in autonomous driving systems.

[0044] In the first driving deviation detection method provided in this disclosure embodiment, the driving deviation can be determined based on the time dimension. Specifically, the driving control information includes a driving control command and the time at which the driving control command is issued; the actual response information includes the time at which the feedback of the command is completed. Determining the driving deviation of the autonomous driving system based on the driving control information and the actual response information includes: determining the response delay of the driving control command based on the time at which the driving control command is issued and the time at which the feedback of the driving control command is completed; and determining the driving deviation of the autonomous driving system based on the response delay. The response delay refers to the time lag from when the autonomous driving system issues a control command to when the vehicle actuator provides feedback indicating that the command has been substantially executed, and is used to quantify the timeliness of the system's response to the command. For example, it may include, but is not limited to, the time difference calculated by comparing the time at which the driving control command is issued with the time at which the corresponding feedback of the command is completed (such as the time when the actual turning angle reaches the target turning angle requirement). For steering or braking control, if the response delay is less than a preset time (e.g., 50 milliseconds), the response is considered to have no delay; if it exceeds the preset time, a response delay is determined to exist.

[0045] This method can sensitively detect slowed response caused by actuator wear, poor lubrication, or increased communication delays. These problems may not be numerically out of tolerance, but the time lag can accumulate into path deviation. Understandably, the feedback completion time can be defined in various ways, such as when the feedback value enters and stabilizes within a small neighborhood of the command value, or when the derivative of the feedback value approaches zero.

[0046] Specifically, suppose the advanced autonomous driving controller issues a target steering angle command at time t0. The steering system begins to act, and the actual steering angle gradually changes, reaching and stabilizing near the target value at time t1. Then, the response delay T_delay = t1 - t0. The performance detection module records the delay value at this moment. To determine driving deviation based on the delay, target response delay thresholds can be preset for different types of commands. For example, calibration shows that under normal vehicle conditions, the threshold value T_critical_steer for steering commands is 100 milliseconds, and the threshold value T_critical_brake for braking commands is 150 milliseconds. Then, the driving deviation is calculated as the difference between the response delay value and the target threshold value, i.e., ΔT = T_delay - T_critical. This ΔT is a deviation quantity characterizing the degree of performance degradation. Optionally, the response delay can also be directly compared with the threshold value; if T_delay > T_critical, then a deviation is determined to exist.

[0047] For example, the target response delay threshold is not fixed. A mapping relationship is pre-stored within the system, defining different response delay thresholds corresponding to different control command types (such as small-angle steering, large-angle steering, comfort braking, and emergency braking). Specifically, determining the driving deviation of the autonomous driving system based on the response delay value includes: obtaining the target response delay threshold corresponding to the driving control command according to the mapping relationship between control commands and response delay thresholds; comparing the response delay value with the target response delay threshold to obtain a comparison result; and determining the driving deviation of the autonomous driving system based on the comparison result. The driving deviation includes the difference between the response delay value and the target response delay threshold. This makes the evaluation criteria more targeted and improves the rationality and accuracy of the detection. For example, the timeliness requirement for emergency braking commands is much higher than that for comfort braking, therefore its threshold is smaller. It is understood that this mapping relationship can be obtained through experimental calibration and can be dynamically compensated according to vehicle conditions (such as load).

[0048] In the second driving deviation detection method provided in this disclosure embodiment, the driving deviation can be determined by direct comparison based on the spatial path. Specifically, the driving control information includes a reference target and a expected driving path relative to the reference target; the actual response information includes the actual driving path of the vehicle identified by the onboard perception system; determining the driving deviation of the autonomous driving system based on the driving control information and the actual response information includes: determining a path deviation based on the expected driving path and the actual driving path; and determining the driving deviation of the autonomous driving system based on the path deviation. The path deviation includes: lateral path deviation and / or longitudinal path deviation obtained by directly comparing the actual driving path and the expected driving path.

[0049] Reference targets refer to static or dynamic objects in the vehicle's external environment that can be identified by the onboard perception system and used as spatial references. They provide an external coordinate system or anchor points for determining the vehicle's expected and actual driving paths. Reference targets can be lane lines, road edges, vehicles ahead, pedestrians, etc. Figure 2 The diagram shows the lateral paths between vehicles. Figure 3 The diagram shows the lateral path between vehicles and pedestrians. Figure 4 The diagram shows the longitudinal path between vehicles and pedestrians. Figure 5 The diagram shows the longitudinal path between vehicles.

[0050] The expected driving path is the trajectory or positional relationship that the autonomous driving system should maintain relative to the reference target. For example, it may include, but is not limited to: in steering control, the lateral displacement trajectory of the vehicle over a future period of time calculated based on the target turning angle and vehicle dynamics model; in following or detouring scenarios, the lateral or longitudinal position sequence planned relative to the reference target (such as the vehicle in front) based on motion prediction and safety margin requirements (such as maintaining a following distance of 100 cm from the vehicle in front, or a detouring path with a lateral distance of 30 cm from the pedestrian).

[0051] The actual response information specifically includes the vehicle's actual driving path relative to the same reference target, identified in real time by an onboard perception system (such as cameras or radar). The actual driving path broadly refers to the vehicle's true driving trajectory obtained through continuous observation and calculation of the relative relationship between the vehicle and external reference targets by an onboard environmental perception system. It reflects the vehicle's objective displacement in the environment. For example, it may include, but is not limited to: continuously identifying the distance between the vehicle body and the lane lines on both sides using cameras, and calculating the vehicle's lateral position relative to the lane centerline at a fixed period (step period), thus forming a lateral displacement sequence; or continuously detecting the distance to the target ahead using radar or cameras, recording the resulting longitudinal distance change sequence.

[0052] This approach fully utilizes environmental perception results as a benchmark to evaluate the final comprehensive effect of control command execution, covering the entire link performance from planning to execution. It can identify problems where internal feedback seems normal, but actual path deviations are caused by model errors or external interference.

[0053] In a specific example, the autonomous driving system plans for the vehicle to travel along the center line of the lane (the expected driving path). Onboard cameras continuously sense the distance between the vehicle and the left and right lane lines. Suppose that at a certain moment, the perception system calculates that the vehicle's center is 2.5 meters from the left lane line and 3.5 meters from the right lane line, while the total lane width is 6 meters. This means that the vehicle's actual driving path deviates from the center line by 0.5 meters (to the right). The performance detection module calculates the difference between this actual lateral position and the expected center line position (3 meters from each side), obtaining a lateral path deviation of 0.5 meters. This is a direct spatial path deviation.

[0054] In another specific example, for a dynamic target, such as a system planned to reduce the distance to the vehicle in front from 50 meters to 30 meters within 5 seconds (expected relative distance change path), the actual relative distance change path is formed by real-time vehicle distance data acquired by millimeter-wave radar. By comparing the two paths, the distance deviation at each time point is calculated, and the statistical characteristics of these deviations (such as maximum deviation and root mean square error) can be used as the longitudinal path deviation.

[0055] For example, to analyze the deviation generation process more precisely, especially when the system has real-time compensation control, a periodic accumulation calculation method is further provided. This method divides the continuous path tracking process into multiple step cycles. Within each step cycle, the system obtains the vehicle's actual driving sub-path based on perception. Simultaneously, the autonomous driving control algorithm compensates for deviations upon detection; the moment this compensation action occurs is called the deviation compensation moment. Specifically, the actual driving path includes a stepping cycle, the actual driving sub-path of the vehicle within the stepping cycle, and the deviation compensation time within the stepping cycle; determining the path deviation based on the expected driving path and the actual driving path includes: for each stepping cycle, performing the following processing: determining the uncompensated driving sub-path within the stepping cycle based on the actual driving sub-path and the deviation compensation time; determining the sub-path deviation based on the uncompensated driving sub-path and the expected driving sub-path; the sub-path deviation includes a sub-lateral path deviation and / or a sub-longitudinal path deviation; accumulating N consecutive sub-lateral path deviations to obtain a lateral path deviation; and / or accumulating N consecutive sub-longitudinal path deviations to obtain a longitudinal path deviation; where N is an integer greater than 1; determining that the path deviation includes the lateral path deviation and / or the longitudinal path deviation. Here, for each step cycle, based on the actual driving sub-path and the deviation compensation time within that cycle, the algorithm inversely derives "the sub-path the vehicle would have traveled if no compensation had been applied," i.e., the uncompensated driving sub-path. This uncompensated driving sub-path is then compared with the expected driving sub-path planned at the beginning of the cycle to obtain the sub-path deviation for that cycle. This sub-path deviation eliminates the influence of compensation control and better reflects the original performance state at the beginning of each control cycle.

[0056] For example, suppose the stepping cycle is 300 milliseconds. Within a certain cycle, the vehicle is expected to move 0.1 meters laterally. However, the sensor detects that 200 milliseconds after the start of the cycle (the deviation compensation time), the vehicle has already moved 0.15 meters laterally, resulting in a premature deviation of 0.05 meters. The control algorithm then intervenes to compensate. By the end of the cycle, the actual net lateral movement is corrected to 0.1 meters. The performance detection module, through back-calculation, obtains that the lateral movement corresponding to the uncompensated driving sub-path is 0.15 meters, therefore the sub-lateral path deviation for that cycle is 0.05 meters. Accumulating N such sub-lateral path deviations (e.g., N=10, i.e., 3 seconds) yields a total lateral path deviation that better reflects the systemic deviation trend. The cumulative calculation of longitudinal deviation is similar. This method can keenly capture those persistent, small deviation trends masked by rapid compensation, thus providing early warning before systemic failures occur.

[0057] In a specific example, such as Figure 6In the schematic diagram of the deviation detection process shown, if no abnormality is detected using the first driving deviation detection method, the second deviation detection method will continue to be used for detection.

[0058] In some embodiments, determining the performance detection result of the autonomous driving system based on the driving deviation includes: determining the performance detection result of the autonomous driving system as abnormal when the driving deviation is greater than a deviation threshold; and determining the performance detection result of the autonomous driving system as normal when the driving deviation is less than or equal to the deviation threshold. By setting a clear deviation threshold, continuous driving deviations are transformed into clear and operable binary state outputs (normal / abnormal), thereby simplifying the decision-making logic of the upper-level system. This allows the vehicle control unit or driver to quickly and clearly understand the system performance status, facilitating the triggering of subsequent warnings or processing procedures. The comparison between the driving deviation and the deviation threshold can be a comparison of lateral driving deviation and a lateral deviation threshold, or a comparison of longitudinal driving deviation and a longitudinal deviation threshold.

[0059] For example, after determining the driving deviation, a graded judgment and handling scheme is adopted to balance safety and system availability. Specifically, after determining that the performance detection result of the autonomous driving system is abnormal, the method further includes: if the driving deviation is less than the upper limit of deviation, using a deviation compensation strategy to compensate for the driving deviation and obtaining a compensation value; if the compensation value is greater than or equal to the upper limit of compensation, triggering an emergency avoidance strategy; and if the driving deviation is greater than or equal to the upper limit of deviation, triggering an emergency avoidance strategy. The driving deviation includes lateral driving deviation or longitudinal driving deviation; either lateral driving deviation or longitudinal driving deviation meeting a condition triggers the emergency avoidance strategy.

[0060] If the driving deviation is less than the upper limit, it indicates that the anomaly is not critical. The system first attempts to employ a deviation compensation strategy. For example, if a fixed bias in lateral control is detected, a reverse correction is pre-added to subsequent control commands; if a slow braking response is detected, a braking command is issued earlier or the command value is increased. Simultaneously, the system monitors the magnitude of the compensation value or the residual deviation after compensation. If the compensation value exceeds the preset upper limit, or if the deviation remains uncorrected after compensation, it indicates that the system itself is unable to correct the deviation, and only then is an emergency avoidance strategy triggered.

[0061] If the driving deviation is greater than or equal to the upper limit of deviation at the initial detection, it indicates that the abnormality has directly endangered safety. The system will bypass the compensation attempt and directly trigger the emergency avoidance strategy. The emergency avoidance strategy may include, but is not limited to: activating audible and visual warnings to request the driver to take immediate control, automatically performing smooth deceleration and pulling over to the side of the road, and automatically braking to a stop while maintaining the lane.

[0062] A deviation threshold is a preset critical value used to trigger different subsequent actions for driving deviations. Its function is to map continuous deviation amounts to discrete fault levels or handling strategies. For example, it may include, but is not limited to: a deviation threshold (e.g., 20 cm) and a higher upper limit (e.g., 60 cm) set for lateral path deviations; and a deviation threshold (e.g., 50 cm) and a higher upper limit (e.g., 100 cm) set for longitudinal path deviations. When the driving deviation reaches the deviation threshold but not the upper limit, a maintenance reminder and a downgraded strategy supporting continued driving may be triggered; when the deviation reaches or exceeds the upper limit, an avoidance strategy requiring an emergency pullover is triggered.

[0063] For example, the lateral deviation threshold is set to 20 cm, and the upper limit is 60 cm; the longitudinal deviation threshold is set to 50 cm, and the upper limit is 100 cm. When a lateral deviation of 30 cm is detected (greater than the threshold but less than the upper limit), the system reports a "functional malfunction repair fault" (Class A) and attempts to perform control compensation, while allowing the vehicle to continue driving under functional limitations. When a lateral deviation of 70 cm is detected (greater than or equal to the upper limit), the system directly reports a "functional malfunction risk fault" (Class S) and immediately triggers the pull-over procedure.

[0064] For example, after triggering the emergency avoidance strategy, in order to comprehensively determine whether the anomaly is an occasional failure or a sign of the end of the system's lifespan, this embodiment of the disclosure further adds a lifespan expiration condition determination step. Specifically, after triggering the emergency avoidance strategy, the method further includes: determining whether the lifespan expiration condition of the autonomous driving system is met; if it is met, issuing a lifespan expiration reminder; if it is not met, issuing a detection reminder; the lifespan expiration condition includes at least one of the following: the total mileage of the vehicle reaches a mileage threshold (e.g., 200,000 kilometers); the cumulative running time of the autonomous driving system reaches a duration threshold (e.g., 10,000 hours).

[0065] In a specific example, if any lifespan expiration condition is met, a lifespan expiration reminder is issued, such as displaying on the vehicle's infotainment screen, "The autonomous driving system has reached its design lifespan; it is recommended to contact the service center immediately for a comprehensive inspection and evaluation," and potentially prohibiting the reactivation of advanced autonomous driving functions. If the lifespan expiration condition is not met, a detection reminder is issued, such as displaying, "An abnormality has been detected in the control system; please schedule a repair appointment as soon as possible," and restricting the level or scope of autonomous driving functions until repair is completed. This comprehensive judgment method, combining direct performance deviation detection with indirect lifespan indicators, makes the assessment of system status more comprehensive and reliable, avoiding overly conservative lifespan termination judgments based on a single false alarm or temporary malfunction.

[0066] It is understandable that the aforementioned preferred solutions, such as those based on delay, path, tiered handling, and lifetime assessment, are not mutually exclusive and can be combined within a complete performance monitoring system. For example, the system can run delay-based and path-based detection threads in parallel, triggering a status assessment when either thread reports an exception. Simultaneously, the tiered handling logic and the comprehensive lifetime assessment logic can serve as a shared post-processing module for all detection threads; see [link to relevant documentation]. Figure 7 The diagram illustrates the process flow of a two-level fault early warning system. This combination allows for cross-verification of system performance from multiple dimensions, forming a more robust detection and defense system, which should also fall within the scope of this disclosure.

[0067] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic, and the execution order between steps is not limited to implementation according to step number.

[0068] In addition, this disclosure also provides apparatus, electronic devices, and computer program products, all of which can be used to implement any of the performance testing methods for the autonomous driving system provided in this disclosure. The corresponding technical solutions and descriptions are described in the corresponding descriptions in the method section and will not be repeated here.

[0069] This disclosure provides a performance testing device for an autonomous driving system, which mainly includes: The first acquisition module is used to acquire driving control information planned by the autonomous driving system for the vehicle; The second acquisition module is used to acquire the actual response information of the vehicle to the driving control information; The determination module is used to determine the driving deviation of the autonomous driving system based on the driving control information and the actual response information; The detection module is used to determine the performance detection result of the autonomous driving system based on the driving deviation, and the performance detection result is used to indicate whether the autonomous driving system is abnormal.

[0070] Figure 8 This is a block diagram of an electronic device provided in an embodiment of the present disclosure.

[0071] This disclosure provides an electronic device comprising: at least one processor 801; at least one memory 802; and one or more I / O interfaces 803 connected between the processor 801 and the memory 802; wherein the memory 802 stores one or more computer programs executable by the at least one processor 801, the one or more computer programs being executed by the at least one processor 801 to enable the at least one processor 801 to execute the performance detection method of the above-described autonomous driving system.

[0072] The modules in the aforementioned electronic devices can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0073] This disclosure also provides a computer program product, including a computer program that, when run in a processor, implements the performance testing method of the above-described autonomous driving system.

[0074] The computer program may be stored on a readable storage medium of a computer device or in the cloud; the processor of the computer device reads the computer program from the readable storage medium or in the cloud.

[0075] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically manifested as a computer storage medium; in another optional embodiment, the computer program product is specifically manifested as a software product, such as a software development kit (SDK), etc.

[0076] 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).

[0077] 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.

[0078] 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.

[0079] 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.

[0080] 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.

[0081] 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.

[0082] 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.

[0083] 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.

[0084] 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.

[0085] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Any modifications or equivalent substitutions made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A performance testing method for an autonomous driving system, characterized in that, include: Obtain the driving control information planned by the autonomous driving system for the vehicle; Obtain the actual response information of the vehicle to the driving control information; The driving deviation of the autonomous driving system is determined based on the driving control information and the actual response information. The performance test result of the autonomous driving system is determined based on the driving deviation, and the performance test result is used to indicate whether the autonomous driving system is abnormal.

2. The method according to claim 1, characterized in that, The driving control information includes driving control commands and the time when the driving control commands are issued; the actual response information includes the time when the feedback of the execution of the driving control commands is completed. Determining the driving deviation of the autonomous driving system based on the driving control information and the actual response information includes: The response delay of the driving control command is determined based on the time when the driving control command is issued and the time when the feedback of the driving control command is completed. The driving deviation of the autonomous driving system is determined based on the response delay.

3. The method according to claim 2, characterized in that, Determining the driving deviation of the autonomous driving system based on the response delay value includes: Based on the mapping relationship between control commands and response delay thresholds, the target response delay threshold corresponding to the driving control command is obtained; The response delay value is compared with the target response delay threshold to obtain a comparison result; The driving deviation of the autonomous driving system is determined based on the comparison results; the driving deviation includes the difference between the response delay value and the target response delay threshold value.

4. The method according to claim 1, characterized in that, The driving control information includes a reference target and a expected driving path relative to the reference target; the actual response information includes the actual driving path of the vehicle identified by the onboard perception system. Based on the driving control information and the actual response information, the driving deviation of the autonomous driving system is determined, including: Determine the path deviation based on the expected driving path and the actual driving path; The driving deviation of the autonomous driving system is determined based on the path deviation.

5. The method according to claim 4, characterized in that, The actual driving path includes a stepping cycle, the actual driving sub-path of the vehicle within the stepping cycle, and the deviation compensation time within the stepping cycle. The step of determining the path deviation based on the expected driving path and the actual driving path includes: For each step cycle, the following processing is performed: based on the actual driving sub-path and the deviation compensation time within the step cycle, the uncompensated driving sub-path within the step cycle is determined; based on the uncompensated driving sub-path and the expected driving sub-path, the sub-path deviation is determined; the sub-path deviation includes sub-lateral path deviation and / or sub-longitudinal path deviation. The lateral path deviation is obtained by accumulating N consecutive sub-lateral path deviations; and / or, the longitudinal path deviation is obtained by accumulating N consecutive sub-longitudinal path deviations; where N is an integer greater than 1. The path deviation is determined to include the lateral path deviation and / or the longitudinal path deviation.

6. The method according to claim 1, characterized in that, Determining the performance test result of the autonomous driving system based on the driving deviation includes: If the driving deviation is greater than the deviation threshold, the performance detection result of the autonomous driving system is determined to be abnormal. If the driving deviation is less than or equal to the deviation threshold, the performance test result of the autonomous driving system is determined to be normal.

7. The method according to claim 6, characterized in that, After determining that the performance test result of the autonomous driving system is abnormal, the method further includes: If the driving deviation is less than the upper limit of the deviation, a deviation compensation strategy is adopted to compensate for the driving deviation and a compensation value is obtained; if the compensation value is greater than or equal to the upper limit of the compensation, an emergency avoidance strategy is triggered. If the driving deviation is greater than or equal to the upper limit of the deviation, an emergency avoidance strategy is triggered.

8. The method according to claim 7, characterized in that, After triggering the emergency avoidance strategy, the method further includes: Determine whether the conditions for the expiration of the autonomous driving system's lifespan are met. If they are met, issue an expiration reminder; if not, issue a detection reminder. The lifespan expiration conditions include at least one of the following: the total mileage of the vehicle reaches a mileage threshold; the cumulative runtime of the autonomous driving system reaches a duration threshold.

9. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores at least one computer program that can be executed by the at least one processor, the at least one computer program being executed by the at least one processor to enable the at least one processor to perform the performance testing method of the autonomous driving system as described in any one of claims 1-8.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when run in a processor, implements the performance testing method of the autonomous driving system according to any one of claims 1-8.

Images