A method and system for evaluating the dynamic braking performance of a vehicle
By acquiring and processing vehicle braking micro-pulse timing data, generating and filtering braking micro-pulse timing datasets, identifying and filtering road braking digital parameters, and performing counterfactual full braking simulation, the problem that low-excitation short-time data is difficult to characterize real road braking capabilities in existing technologies is solved, and stable and reliable evaluation under non-extreme conditions is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG TRANSPORT VOCATIONAL COLLEGE
- Filing Date
- 2026-04-20
- Publication Date
- 2026-07-14
AI Technical Summary
Existing dynamic evaluation methods for vehicle braking performance are difficult to effectively characterize real-world braking capabilities under low-excitation and short-time data conditions. They also lack stability in multi-parameter joint identification, and lack a credibility constraint mechanism oriented towards parameter reliability. Consequently, they are unable to achieve counterfactual dynamic verification and evaluation of the complete braking process under non-extreme, low-risk detection conditions.
By acquiring vehicle braking micropulse timing data, a braking micropulse timing dataset is generated, outliers are removed and the timing is reconstructed, the set of road braking digital parameters is identified, parameter confidence labels are generated, counterfactual full braking simulation is performed, and dynamic verification and evaluation results are output.
It enables standardized sampling and effective segment purification of braking response under low-risk, short-window conditions, improves the availability of input data, enhances parameter stability and evaluation consistency, and generates recordable and verifiable braking performance evaluation results.
Smart Images
Figure CN122064977B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle braking data processing technology, specifically to a method and system for dynamic testing and evaluation of vehicle braking performance. Background Technology
[0002] With the development of automotive electronic control, on-board sensing, bus communication, and computational evaluation technologies, automotive braking performance testing has gradually evolved from traditional bench testing, road testing, and manual judgment to dynamic testing and data-driven evaluation technologies oriented towards the overall vehicle operating status. Existing technologies typically collect multi-source time-series information such as wheel speed, longitudinal acceleration, vehicle speed, and braking requests to analyze braking distance, deceleration, braking build-up process, and road surface adhesion. Some schemes also introduce parameter identification and numerical extrapolation methods to improve the objectivity and repeatability of test results. Overall, automotive braking performance testing technology is developing from result-based measurement methods to process modeling and dynamic evaluation methods.
[0003] However, existing automotive braking performance testing technologies still primarily rely on full-braking road tests, bench measurements, or direct calculations for single results, which are significantly insufficient for characterizing the overall vehicle braking capacity under non-extreme, low-risk, and short-term dynamic testing scenarios. The coupling between wheel speed changes, longitudinal acceleration response, and braking request quantities under short-window, low-excitation braking conditions is weak, and factors such as vehicle mass, road gradient, tire condition, and road surface adhesion are easily mixed, making it difficult for existing methods to reliably identify a set of parameters sufficient to reflect real-world braking capacity from limited testing data. Related technologies typically lack component constraints and credibility screening mechanisms for parameter reliability, failing to effectively suppress the amplified impact of communication frame gaps, local anomalies, driving fluctuations, and short-window sampling noise on the evaluation results. Furthermore, most existing technologies are limited to direct measurement or partial estimation of braking results that have already occurred, lacking a technical approach to counterfactually extrapolate the complete braking process based on short-term detection data. Consequently, it is difficult to provide comprehensive verification indicators such as equivalent braking distance, equivalent average fully exerted deceleration, equivalent braking establishment time, and left-right wheel consistency risk without conducting high-risk full braking tests. It is even more difficult to achieve dynamic, reliable, and verifiable evaluation of the vehicle's road braking performance. Summary of the Invention
[0004] In view of the above-mentioned problems, the present invention is proposed.
[0005] Therefore, the technical problem solved by this invention is that existing vehicle braking dynamic evaluation methods have the following problems: low-excitation short-time data are difficult to effectively represent real road braking capabilities, multi-parameter joint identification has insufficient stability, lack of credibility constraint mechanism for parameter reliability, and how to achieve counterfactual dynamic verification and evaluation for the complete braking process under non-extreme, low-risk detection conditions.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for dynamic testing and evaluation of vehicle braking performance, comprising acquiring vehicle braking micropulse timing data and generating a braking micropulse timing dataset.
[0007] Based on the braking micropulse time series dataset, identify the set of digital parameters for road braking and generate parameter confidence labels.
[0008] Based on the set of digital parameters for road braking, a counterfactual full braking simulation is performed, and dynamic verification and evaluation results are output.
[0009] As a preferred embodiment of the vehicle braking performance dynamic testing and evaluation method described in this invention, the acquisition of vehicle braking micro-pulse timing data includes continuously collecting wheel speeds, vehicle longitudinal acceleration, vehicle speed, braking request quantity, and corresponding timestamps when the vehicle is within the vehicle braking dynamic testing speed range. When the braking request quantity enters the vehicle braking micro-pulse excitation range and its duration is within the vehicle braking short-window sampling range, the corresponding sampling point is written into the analysis cache according to a unified sampling period. If a lock-up precursor flag, a communication frame loss flag, or a speed change flag is detected, the corresponding sampling point is removed, and the related sampling points before and after are retained for timing reconstruction.
[0010] As a preferred embodiment of the dynamic testing and evaluation method for vehicle braking performance described in this invention, the generation of the braking micro-pulse time-series dataset includes performing unified time-base resampling, outlier removal, drift correction, and missing point completion on the vehicle braking micro-pulse time-series data; dividing the data into non-braking segments, weak braking segments, and micro-pulse braking segments according to the correspondence between the change in braking request quantity and the vehicle's longitudinal acceleration response; retaining only segments that simultaneously satisfy the vehicle braking micro-pulse excitation interval and the vehicle braking short-window sampling interval and do not detect any lock-up precursor indicators; and arranging these segments in chronological order to form the braking micro-pulse time-series dataset.
[0011] As a preferred embodiment of the dynamic testing and evaluation method for vehicle braking performance described in this invention, the identification of the road braking digital parameter set based on the braking micro-pulse time-series dataset includes: calculating the estimated total vehicle mass and road slope based on non-braking segments and weak braking segments in the braking micro-pulse time-series dataset; and calculating the estimated dynamic tire radius, estimated longitudinal equivalent stiffness, estimated brake build-up delay, and estimated road surface adhesion based on the micro-pulse braking segments. The estimated total vehicle mass and road slope are updated using iterative constraints, and the estimated dynamic tire radius, estimated longitudinal equivalent stiffness, estimated brake build-up delay, and estimated road surface adhesion are identified using time-series correlation, and encapsulated into a road braking digital parameter set according to a preset field order.
[0012] As a preferred embodiment of the dynamic testing and evaluation method for vehicle braking performance described in this invention, the generation of parameter credibility identifiers includes: generating sub-item credibility based on the residual consistency, temporal continuity, parameter update stability, and fragment integrity between the road braking digital parameter set and the braking micropulse time series dataset, and aggregating and calculating the parameter credibility identifier according to preset weights. When the credibility of any sub-item is lower than the corresponding vehicle braking parameter credibility lower limit, updating the parameter corresponding to the sub-item credibility is stopped, and the corresponding parameter from the previous stable moment is called back to the road braking digital parameter set. When the parameter credibility identifier is not lower than the vehicle braking parameter credibility lower limit, the road braking digital parameter set enters the counterfactual full braking simulation.
[0013] As a preferred embodiment of the dynamic testing and evaluation method for vehicle braking performance described in this invention, the step of performing a counterfactual full braking simulation based on a set of road braking digital parameters includes: generating a target braking request different from the braking request amount based on the vehicle speed; and constructing a counterfactual longitudinal solution model for vehicle braking by calling the estimated vehicle total mass, road slope, dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road surface adhesion values from the set of road braking digital parameters after being filtered by parameter confidence indicators. Using the target braking request as input, the speed decay, displacement accumulation, and wheel speed response are recursively calculated according to discrete time steps until the vehicle speed drops to a zero-speed determination value, thus obtaining the counterfactual full braking trajectory of the vehicle.
[0014] As a preferred embodiment of the dynamic testing and evaluation method for vehicle braking performance described in this invention, the output dynamic testing and evaluation results include: extracting features from the vehicle's counterfactual full braking trajectory to form equivalent braking distance, equivalent average fully exerted deceleration, equivalent brake setup time, and a left-right wheel inconsistency risk index. The left-right wheel inconsistency risk index is jointly calculated by the difference in wheel speed response between the left and right wheels in the micro-pulse braking segment and the difference in deceleration between the left and right wheels in the vehicle's counterfactual full braking trajectory. The equivalent braking distance, equivalent average fully exerted deceleration, equivalent brake setup time, left-right wheel inconsistency risk index, and parameter reliability indicators are written into the vehicle braking dynamic testing and evaluation record.
[0015] As a preferred embodiment of the dynamic testing and evaluation system for automotive braking performance described in this invention, it includes a micropulse data construction module, a braking parameter joint identification module, and a counterfactual deduction evaluation module.
[0016] The micro-pulse data construction module is used to acquire automotive braking micro-pulse timing data and generate a braking micro-pulse timing dataset.
[0017] The braking parameter joint identification module is used to identify the set of road braking digital parameters based on the braking micropulse time series dataset and generate parameter confidence identifiers.
[0018] The counterfactual deduction and evaluation module is used to perform counterfactual full braking deduction based on the set of road braking digital parameters and output dynamic verification and evaluation results.
[0019] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a method for dynamic testing and evaluation of vehicle braking performance.
[0020] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of a method for dynamic testing and evaluation of vehicle braking performance.
[0021] The beneficial effects of this invention are:
[0022] By acquiring automotive braking micro-pulse time-series data and generating a braking micro-pulse time-series dataset through anomaly removal, time-series reconstruction, and fragment selection, standardized sampling and effective fragment purification of low-risk, short-window braking responses were achieved. Its purpose is to provide unified, stable, and traceable data input for subsequent parameter identification, preventing raw driving disturbances, missing frames, and invalid fragments from directly entering the computational chain. Ultimately, this achieves the beneficial effects of improving input data availability, reducing subsequent identification ambiguity, and enhancing the consistency of the entire evaluation process.
[0023] By using non-braking segments, weak braking segments, and micro-pulse braking segments to identify vehicle gross mass estimates, road slope estimates, dynamic tire radius estimates, longitudinal equivalent stiffness estimates, brake set-off delay estimates, and road adhesion estimates, and superimposing a component reliability screening and a mechanism to write back parameters from the previous stable moment, robust construction of a road braking digital parameter set was achieved. Its purpose is to transform short-window responses into digital state representations that can be directly used for extrapolation. Ultimately, this achieves the beneficial effects of suppressing local anomalies, improving parameter stability, and enhancing the reliability of subsequent extrapolations.
[0024] By generating target braking requests different from the actual braking requests based on vehicle speed, and constructing a longitudinal counterfactual braking model using a set of road braking digital parameters filtered by parameter credibility identifiers, the system recursively obtains the vehicle's counterfactual full braking trajectory and extracts equivalent braking distance, equivalent average fully applied deceleration, equivalent brake setup time, and a risk index indicating inconsistency between left and right wheels. This achieves extrapolation from short-term, low-excitation detection to a complete braking process assessment. Its purpose is to replace high-risk full braking field tests and generate recordable and verifiable verification results. Ultimately, it achieves the beneficial effect of more closely approximating real-world full braking performance while also considering consistency risk assessment. Attached Figure Description
[0025] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 The above is an overall flowchart of a dynamic testing and evaluation method for automobile braking performance provided in Embodiment 1 of the present invention. Detailed Implementation
[0027] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.
[0028] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for dynamic testing and evaluation of vehicle braking performance is provided, comprising:
[0029] S1: Acquire automotive braking micro-pulse timing data and generate a braking micro-pulse timing dataset.
[0030] While the vehicle is within the speed range for dynamic braking testing, wheel speeds, longitudinal acceleration, vehicle speed, braking request quantities, and corresponding timestamps are continuously collected. When a braking request quantity enters the vehicle braking micro-pulse excitation range and its duration falls within the vehicle braking short-window sampling range, the corresponding sampling point is written to the analysis buffer according to a uniform sampling period. If a lock-up precursor flag, a communication frame loss flag, or a speed change flag is detected, the corresponding sampling point is removed, and the preceding and following related sampling points are retained for time-series reconstruction.
[0031] Furthermore, the vehicle braking dynamic test speed range is set to 30km / h~50km / h. When the vehicle speed enters and remains within the vehicle braking dynamic test speed range for no less than 1 second, the acquisition of vehicle braking micropulse timing data is triggered.
[0032] Furthermore, the braking request quantity is a normalized braking request quantity calculated from the brake pedal opening, master cylinder pressure, or brake-by-wire request value. The braking request quantity is represented by a normalized braking request quantity, with a value range of 0 to 100%. The preferred range for the vehicle's braking micro-pulse excitation is 8% to 18%. The rule for setting this range is to ensure a identifiable response to the vehicle's longitudinal acceleration while avoiding a high-excitation braking state close to brake lockup. The preferred short-window sampling range for vehicle braking is 0.20s to 0.70s. This range is derived from the effective coverage requirement of the initial braking phase. A range shorter than 0.20s makes it difficult to characterize the time delay and establishment slope, while a range longer than 0.70s easily incorporates driving fluctuations and road surface disturbances.
[0033] Furthermore, the uniform sampling period is preferably set to 10ms, corresponding to a sampling frequency of 100Hz, to balance wheel speed change resolution, vehicle longitudinal acceleration response capture capability, and onboard controller computational load. The criteria for determining a lock-up precursor are: any wheel speed deviates from the equivalent wheel speed corresponding to the vehicle speed by more than 8%, and this state occurs consecutively for 3 sampling periods. The criteria for determining a communication frame loss are: any critical signal is missing for 2 or more consecutive sampling periods. The criteria for determining a speed change are: the difference in vehicle speed between two adjacent sampling periods is greater than 2km / h. For rejected sampling points, piecewise linear reconstruction is performed using 3 valid sampling points before and after the rejection point to maintain temporal continuity. The equivalent wheel speed corresponding to the vehicle speed is the wheel speed value converted according to the calibrated tire radius.
[0034] The vehicle braking micropulse time series data is subjected to unified time base resampling, outlier removal, drift correction and missing point filling. According to the correspondence between the change of braking request quantity and the longitudinal acceleration response of the vehicle, it is divided into non-braking segments, weak braking segments and micropulse braking segments. Only segments that simultaneously meet the vehicle braking micropulse excitation interval and vehicle braking short window sampling interval and do not detect the lock-up precursor indicator are retained. The segments are arranged in chronological order to form a braking micropulse time series dataset.
[0035] Furthermore, unified time-base resampling aligns all wheel speeds, vehicle longitudinal acceleration, vehicle speed, braking request quantities, and timestamps to a unified 10ms time base. Outlier removal employs a neighborhood consistency rule; if a sampling point deviates from the mean of its five preceding and following sampling points by more than three times the standard deviation, it is considered an outlier and removed. Drift correction performs zero-bias correction on vehicle longitudinal acceleration, taking the mean of a continuous 1-second non-braking segment as the zero bias and subtracting it. Missing point imputation uses linear interpolation when there are no more than three consecutive missing sampling periods; missing point imputation is performed on isolated missing sampling points after communication missing frame identification removal.
[0036] Furthermore, a non-braking segment is defined as a continuous segment where the normalized braking request amount is less than 2% and the absolute value of the vehicle's longitudinal acceleration is no greater than 0.03g. A weak braking segment is defined as a continuous segment where the normalized braking request amount is between 2% and 8% and the vehicle's longitudinal acceleration is between 0.03g and 0.08g. A micro-pulse braking segment is defined as a continuous segment where the normalized braking request amount is between 8% and 18% and the vehicle's longitudinal acceleration is between 0.08g and 0.20g. These thresholds are derived from the joint constraints of this invention on low-risk, identifiable, and non-lock-up braking states, ensuring that different segments can correspond to the three types of operating conditions: basic driving, weak response, and micro-pulse excitation.
[0037] Furthermore, the braking micropulse time-series dataset is arranged chronologically. This involves sorting qualified segments in ascending order of timestamps and adding a segment category identifier, start time, end time, and number of consecutive valid sampling points to each segment. If a single segment has fewer than 20 consecutive valid sampling points, it is deemed to have no identification value and is discarded. When at least three consecutive braking micropulse segments are retained, and their intervals are all no greater than 2 seconds, a braking micropulse time-series dataset is generated for subsequent identification of road braking digital parameter sets. If these conditions are not met, subsequent segments continue to accumulate in the analysis buffer until the generation conditions are met or the current detection cycle ends.
[0038] It should be noted that this step first collects raw micro-pulse braking data at low-risk vehicle speeds, and then forms a computable dataset through anomaly removal, time-series reconstruction, and segment selection, providing a unified, stable, and traceable input basis for subsequent parameter identification. Raw driving disturbances and invalid segments are eliminated upfront to reduce ambiguity in subsequent calculations.
[0039] S2: Identify the set of digital parameters for road braking based on the braking micropulse time series dataset and generate parameter confidence identifiers.
[0040] Based on the non-braking and weak braking segments in the braking micropulse time-series dataset, the vehicle gross weight and road slope estimates are calculated. Furthermore, based on the micropulse braking segments, dynamic tire radius estimates, longitudinal equivalent stiffness estimates, brake set-off delay estimates, and road adhesion estimates are calculated. Iterative constraints are applied to update the vehicle gross weight and road slope estimates. Temporal correlation identification is used for the dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road adhesion estimates, which are then encapsulated into a set of road braking digital parameters according to a preset field order.
[0041] Furthermore, the calculation of the vehicle gross mass estimate and road slope estimate uses non-braking segments and weak braking segments from the braking micropulse time series dataset as input. The cumulative effective duration of the non-braking segment is no less than 1.5s, and the cumulative effective duration of the weak braking segment is no less than 0.8s, with vehicle speed fluctuations within the segment not exceeding 5km / h. The above duration and fluctuation range settings are based on the fact that this invention needs to ensure the identifiability of longitudinal dynamic information in the short-window test, while avoiding coupling distortion of the vehicle gross mass estimate and road slope estimate due to excessive driving disturbances.
[0042] Furthermore, iterative constraint update refers to using the matching relationship between vehicle speed changes, vehicle longitudinal acceleration changes, and braking request changes as a basis. Initial values for the estimated total vehicle mass and road slope are given first, and then the deviation between the longitudinal response obtained from the current estimates and the actual longitudinal response in the braking micro-pulse time-series dataset is compared. The initial values are calculated jointly by the first non-braking segment and the weak braking segment that meet the conditions. An update is performed when the vehicle longitudinal acceleration deviation is greater than 0.02g. In each update, the change in the estimated total vehicle mass relative to the previous round does not exceed 4%, and the change in the estimated road slope relative to the previous round does not exceed 0.2°. A maximum of 5 rounds of updates are performed to keep the parameter changes under short-term detection within the vehicle's actual reachability range.
[0043] Furthermore, the dynamic tire radius estimate is calculated based on the correspondence between vehicle speed and wheel speed; the longitudinal equivalent stiffness estimate is calculated based on the correspondence between wheel speed lag changes and vehicle longitudinal acceleration establishment process in the micro-pulse braking segment; the brake establishment delay estimate is defined as the time interval from the start of the brake request entering the vehicle's brake micro-pulse excitation interval to the vehicle's longitudinal acceleration reaching 0.05g for three consecutive sampling periods; and the road adhesion estimate is calculated based on the degree of matching between the wheel speed decrease trend and the vehicle longitudinal acceleration increase trend. The road adhesion estimate is dimensionless and its value range is set to 0~1.20. This invention only uses micro-pulse braking segments without detected lock-up warning signs as calculation input.
[0044] Furthermore, temporal correlation identification refers to comparing the changing trends of the dynamic tire radius estimate, longitudinal equivalent stiffness estimate, brake setup delay estimate, and road adhesion estimate of the current segment with the previous segment according to the temporal order of the micropulse braking segments. The parameter update is confirmed to be valid only if the same parameter changes in the same direction in two or more consecutive micropulse braking segments and the change amplitude does not exceed 15% of the corresponding parameter value in the previous segment. Subsequently, the estimated vehicle mass, road gradient, dynamic tire radius, longitudinal equivalent stiffness, brake setup delay, and road adhesion are encapsulated into a road braking digital parameter set in the order of these estimates, for direct use in subsequent counterfactual full braking simulations.
[0045] Based on the residual consistency, temporal continuity, parameter update stability, and fragment integrity between the road braking digital parameter set and the braking micropulse time series dataset, sub-item confidence levels are generated and aggregated according to preset weights to obtain parameter confidence identifiers. Each sub-item confidence level and parameter confidence identifier is normalized to a continuous interval of 0 to 1, and a weighted summation method is used for aggregation calculation. When the confidence level of any sub-item falls below the corresponding vehicle braking parameter confidence lower limit, updating the parameter corresponding to the sub-item confidence level stops, and the corresponding parameter from the previous stable moment is retrieved and written back to the road braking digital parameter set. When the parameter confidence identifier is not lower than the vehicle braking parameter confidence lower limit, the road braking digital parameter set enters the counterfactual full braking simulation.
[0046] Furthermore, the predicted wheel speed changes and predicted vehicle longitudinal acceleration changes are obtained by substituting the current road braking digital parameter set into the longitudinal response reconstruction process of the corresponding segment. The longitudinal response reconstruction process involves inputting the current road braking digital parameter set along with the vehicle speed, braking request quantity, and time series from the corresponding segment into a unified longitudinal state update process, reconstructing the predicted wheel speed changes and predicted vehicle longitudinal acceleration changes step by step in discrete time. Residual consistency is obtained by comparing the predicted wheel speed changes and predicted vehicle longitudinal acceleration changes corresponding to the road braking digital parameter set with the actual wheel speed changes and actual vehicle longitudinal acceleration changes in the braking micropulse time series dataset. When the wheel speed residual is no greater than 0.8 km / h and the vehicle longitudinal acceleration residual is no greater than 0.03g, the residual consistency requirement is met. Temporal continuity is determined by comparing the variation range of the same parameter in two adjacent valid segments. When the change in the estimated total vehicle mass does not exceed 3%, the change in the estimated road slope does not exceed 0.15°, the change in the estimated dynamic tire radius does not exceed 1.5%, the change in the estimated longitudinal equivalent stiffness does not exceed 10%, the change in the estimated brake settling delay does not exceed 15ms, and the change in the estimated road surface adhesion does not exceed 0.08, the temporal continuity is deemed to meet the requirements.
[0047] Furthermore, parameter update stability is determined by statistically analyzing the update direction and fluctuation amplitude of the same parameter across three consecutive valid segments. Parameter update stability is deemed to meet the requirements when the update direction does not repeatedly switch over three consecutive segments and the maximum fluctuation amplitude does not exceed 12% of the current value of the corresponding parameter. Segment integrity is determined by the proportion of valid sampling points and segment coverage. Segment integrity is deemed to meet the requirements when the proportion of valid sampling points in a single segment is not less than 90%, and the number of micropulse braking segments used for identification in this round is not less than three. The above thresholds are set based on the principle that this invention needs to avoid amplifying the impact of local anomalies on parameter reliability identification under short window and small sample conditions.
[0048] Furthermore, in the preset weights, the weight for residual consistency is 0.35, the weight for temporal continuity is 0.20, the weight for parameter update stability is 0.25, and the weight for fragment integrity is 0.20. In the lower confidence limit for vehicle braking parameters, the lower confidence limit for each component is set to 0.60, the lower confidence limit for the parameter confidence identifier is set to 0.70, and the previous stable moment is defined as the moment when the most recent parameter confidence identifier is not lower than 0.75 and the confidence of each component is not lower than 0.60. When the confidence of any component is lower than 0.60, only the parameter corresponding to that component's confidence is frozen, and the corresponding parameter from the previous stable moment is written back. When the parameter confidence identifier is not lower than 0.70, the current set of road braking digital parameters is allowed to enter the counterfactual full braking simulation for subsequent generation of dynamic verification and evaluation results.
[0049] It should be noted that this step, under short-term, low-risk micro-pulse samples, does not directly provide braking conclusions. Instead, it first uses non-braking segments, weak braking segments, and micro-pulse braking segments to identify different parameters, and then filters unstable parameters through sub-item confidence levels, forming a set of road braking digital parameters that can be directly used for subsequent extrapolation. Compared to existing technologies that rely solely on single road test results or single sensor experience scoring, this embodiment places vehicle total mass, road gradient, dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road surface adhesion into the same data link. By freezing low-confidence parameters and writing back parameters from the previous stable moment, it suppresses short-window noise and local anomalies, providing a more stable input basis for subsequent counterfactual full braking extrapolation. This supports braking evaluation that more closely approximates real road conditions under non-extreme detection conditions.
[0050] S3: Performs counterfactual full braking simulation based on the set of road braking digital parameters and outputs dynamic verification and evaluation results.
[0051] A target braking request, different from the braking request amount, is generated based on the vehicle speed. The vehicle's gross mass estimate, road slope estimate, dynamic tire radius estimate, longitudinal equivalent stiffness estimate, brake set-off delay estimate, and road adhesion estimate, filtered from the road braking digital parameter set using parameter confidence indicators, are used to construct a counterfactual longitudinal solution model for vehicle braking. Using the target braking request as input, speed decay, displacement accumulation, and wheel speed response are recursively calculated in discrete time steps until the vehicle speed drops to the zero-speed determination value, yielding the vehicle's counterfactual full braking trajectory.
[0052] Furthermore, the target braking request is represented by a normalized target braking request, with a value range of 0~100%, and is distinguished from the braking request quantity in S1. When the vehicle speed is between 30km / h and 40km / h, the target braking request is set to 40%, and when the vehicle speed is between 40km / h and 50km / h, the target braking request is set to 50%. The above setting rules are derived from the unified constraint of the standard service braking intensity in this invention, that is, without directly reusing the micropulse braking request quantity, a request level sufficient to form a complete deceleration and stopping process and not exceeding the stable operating range of the conventional braking system is selected for counterfactual full braking simulation.
[0053] Furthermore, the counterfactual longitudinal solution model for vehicle braking includes a parameter loading module, a time delay module, a wheel speed response module, a longitudinal state update module, and a termination determination module. The parameter loading module reads the estimated total vehicle mass, road slope, dynamic tire radius, longitudinal equivalent stiffness, brake setup delay, and road adhesion from the road braking digital parameter set. The time delay module maintains the vehicle in an initial inertial decay state, affected only by road slope and driving resistance, until the brake setup delay estimate is reached. After the brake setup delay estimate is reached, it loads the target braking request into the wheel speed response module and the longitudinal state update module, forming the counterfactual solution starting point corresponding to the current vehicle state. Driving resistance includes at least rolling resistance and air resistance.
[0054] Furthermore, the discrete time step adopts the unified sampling period from S1, set to 10ms. Within each discrete time step, the wheel braking intensity at the current moment is first determined based on the target braking request, road surface adhesion estimate, and longitudinal equivalent stiffness estimate. Then, the wheel speed, vehicle speed, and vehicle displacement are updated by combining the dynamic tire radius estimate and road slope estimate. If the predicted tire slip trend at a certain moment exceeds 15%, the wheel braking intensity at that moment is limited to 95% of the previous moment to avoid the counterfactual extrapolation result deviating from the achievable range under the current road adhesion conditions. When the vehicle speed drops to 0.5km / h, the zero-speed threshold is reached, and the recursion ends. If the zero-speed threshold is not reached after 8 seconds of continuous recursion, the 8-second mark is used as the termination point for the extrapolation. The predicted tire slip trend is calculated based on the relative changes between the vehicle speed and the wheel speeds within the same discrete time step.
[0055] Feature extraction is performed on the counterfactual full braking trajectory of a vehicle to generate equivalent braking distance, equivalent average fully exerted deceleration, equivalent brake set-off time, and a left-right wheel inconsistency risk index. The left-right wheel inconsistency risk index is jointly calculated by the difference in wheel speed response between the left and right wheels in the micro-pulse braking segment and the difference in deceleration between the left and right wheels in the counterfactual full braking trajectory of the vehicle. The equivalent braking distance, equivalent average fully exerted deceleration, equivalent brake set-off time, left-right wheel inconsistency risk index, and parameter reliability indicators are written into the vehicle braking dynamic test and evaluation record.
[0056] Furthermore, the equivalent braking distance is defined as the cumulative displacement from the moment the target braking request takes effect until the moment the zero-speed determination value is reached. The equivalent average fully exerted deceleration is extracted from the velocity and displacement changes in the vehicle's counterfactual full braking trajectory between the moment the vehicle speed decreases from the initial vehicle speed to 80% of the initial vehicle speed and the moment it decreases to 10% of the initial vehicle speed. The equivalent brake establishment time is defined as the time interval from the moment the target braking request takes effect to the moment the vehicle's longitudinal deceleration first reaches 90% of the steady-state deceleration for the first three consecutive discrete time steps. The steady-state deceleration is defined as the average value of the earliest stable segment in the counterfactual full braking trajectory after the target braking request takes effect, where the fluctuation of the vehicle's longitudinal deceleration for 20 consecutive discrete time steps does not exceed 5%. The above extraction rules are derived from the present invention's requirement to extract stable features from the counterfactual full braking trajectory that can characterize both braking efficiency and the brake establishment process.
[0057] Furthermore, the differences in wheel speed response between the left and right wheels in the micro-pulse braking segment and the differences in deceleration between the left and right wheels in the counterfactual full braking trajectory of the vehicle are first normalized and then combined with equal weights to form a risk index for wheel inconsistency. A risk level label is then generated based on the joint threshold of the differences in wheel speed response and deceleration between the left and right wheels. The risk index for wheel inconsistency is generated using a tiered joint judgment method. First, the differences in wheel speed response between the left and right wheels within the micro-pulse braking segment are calculated, and then the differences in deceleration between the left and right wheels in the counterfactual full braking trajectory of the vehicle are calculated. When the difference in wheel speed response between the left and right wheels does not exceed 3% and the difference in deceleration between the left and right wheels does not exceed 0.03g, it is judged as low risk. When the difference in wheel speed response between the left and right wheels is greater than 3% but not more than 6%, or the difference in deceleration between the left and right wheels is greater than 0.03g but not more than 0.06g, it is judged as medium risk. When the difference in wheel speed response between the left and right wheels is greater than 6% or the difference in deceleration between the left and right wheels is greater than 0.06g, it is judged as high risk. The above thresholds are set based on the fact that this invention needs to combine the asymmetric characteristics of short-window measured data with the asymmetric trends of counterfactual inference to jointly evaluate the consistency of left and right wheel braking.
[0058] Furthermore, the vehicle braking dynamic inspection and evaluation record should include at least the record number, record timestamp, vehicle speed, target braking request, set of road braking digital parameters, parameter confidence level, equivalent braking distance, equivalent average fully applied deceleration, equivalent braking establishment time, left and right wheel inconsistency risk index, and risk level indicator. When the parameter confidence level is between 0.70 and 0.75, a boundary confidence level indicator is added to the vehicle braking dynamic inspection and evaluation record. When the parameter confidence level is not lower than 0.75, a stable confidence level indicator is added to the vehicle braking dynamic inspection and evaluation record, so that subsequent inspection terminals can call the vehicle braking dynamic inspection and evaluation record to generate inspection conclusions, review prompts, or historical comparison results.
[0059] Furthermore, the difference in wheel speed response and the difference in deceleration of the left and right wheels mentioned in this invention are calculated in units of the left and right wheels of the front axle and the left and right wheels of the rear axle, respectively, and the larger of the two values is taken as the difference value between the left and right wheels of the current vehicle.
[0060] It should be noted that this step, without conducting high-risk full braking tests, uses a set of road braking digital parameters filtered for credibility as input to construct a longitudinal counterfactual solution model of vehicle braking corresponding to the current vehicle state. It recursively generates the vehicle's counterfactual full braking trajectory using a unified discrete time step, and then extracts verification indicators such as braking distance, deceleration, brake set-off time, and risk of inconsistency between left and right wheels from this trajectory, forming a recordable and verifiable dynamic verification and evaluation result. Compared to existing technologies that rely on single road test results, fixed empirical thresholds, or a single deceleration indicator, this embodiment can incorporate vehicle total mass, road gradient, tire condition, brake set-off time delay, and road surface adhesion into the counterfactual inference process under short-window, low-excitation, and non-limit braking conditions, thereby providing an evaluation result closer to real-world full braking performance, while retaining credibility layering information to support verification and historical comparison.
[0061] Example 2, an embodiment of the present invention, provides a dynamic testing and evaluation system for automobile braking performance, including a micropulse data construction module, a braking parameter joint identification module, and a counterfactual deduction evaluation module.
[0062] The micro-pulse data construction module is used to acquire automotive braking micro-pulse timing data and generate a braking micro-pulse timing dataset.
[0063] The brake parameter joint identification module is used to identify the set of road braking digital parameters based on the brake micropulse time series dataset and generate parameter confidence identifiers.
[0064] The counterfactual deduction and evaluation module is used to perform counterfactual full braking deduction based on the set of road braking digital parameters and output dynamic verification and evaluation results.
Claims
1. A method for dynamic testing and evaluation of automotive braking performance, characterized in that, include: When the vehicle speed is within the dynamic test speed range of 30km / h to 50km / h, the wheel speed of each wheel, longitudinal acceleration of the vehicle, vehicle speed, braking request quantity and corresponding timestamp are continuously collected; when the braking request quantity enters the vehicle braking micropulse excitation range of 8% to 18% and the duration is within the vehicle braking short window sampling range of 0.20s to 0.70s, the vehicle braking micropulse time series data is obtained according to a uniform sampling period of 10ms, and a braking micropulse time series dataset is generated. Based on non-braking segments, weak braking segments, and micro-pulse braking segments in the braking micro-pulse time-series dataset, a set of road braking digital parameters is identified. The set of road braking digital parameters includes estimated vehicle gross weight, estimated road slope, estimated dynamic tire radius, estimated longitudinal equivalent stiffness, estimated brake set-off delay, and estimated road surface adhesion, in a preset field order. Based on the residual consistency, temporal continuity, parameter update stability, and segment integrity between the set of road braking digital parameters and the braking micro-pulse time-series dataset, parameter confidence identifiers are generated. When the confidence level of the parameter is not lower than the lower confidence limit of the vehicle braking parameter, a target braking request different from the braking request quantity is generated, and the road braking digital parameter set is called to construct the vehicle braking counterfactual longitudinal solution model. Taking the target braking request as input, the vehicle speed decay, vehicle displacement accumulation and wheel speed response of each wheel are recursively calculated according to the discrete time step until the vehicle speed drops to the zero speed judgment value, and the vehicle braking counterfactual full braking trajectory is obtained. Feature extraction is performed on the counterfactual full braking trajectory of a vehicle, and the output includes the equivalent braking distance, the equivalent average fully exerted deceleration, the equivalent braking establishment time, the risk index of inconsistency between the left and right wheels, and the dynamic verification and evaluation results of parameter credibility indicators.
2. The method for dynamic testing and evaluation of vehicle braking performance as described in claim 1, characterized in that: The acquisition of automotive braking micro-pulse timing data includes... When the vehicle is in the vehicle braking dynamic test speed range, continuously collect wheel speed of each wheel, longitudinal acceleration of the vehicle, vehicle speed, braking request amount and corresponding timestamp; When the braking request quantity enters the vehicle braking micro-pulse excitation interval and the duration is within the vehicle braking short window sampling interval, the corresponding sampling point is written into the cache to be analyzed according to a uniform sampling period. When the wheel speed of any wheel deviates from the equivalent wheel speed corresponding to the vehicle speed by more than 8% and this occurs for 3 consecutive sampling periods, a lock-up warning sign is generated. A communication missing frame flag is generated when any critical signal is missing for two or more consecutive sampling periods. When the difference in vehicle speed between two adjacent sampling periods is greater than 2 km / h, a speed change indicator is generated; Sampling points with signs of impending arrest, missing communication frames, or sudden velocity changes are removed, and piecewise linear reconstruction is performed using three valid sampling points before and after the removed sampling point. The reconstructed vehicle braking micropulse timing data is subjected to unified time base resampling, outlier removal, drift correction and missing point filling, and is divided into non-braking segments, weak braking segments and micropulse braking segments according to the correspondence between the change in braking request amount and the longitudinal acceleration response of the vehicle. Only segments that simultaneously satisfy the vehicle braking micropulse excitation interval and the vehicle braking short window sampling interval, and for which no lock-up warning signs are detected, are retained and arranged in chronological order to form the braking micropulse time series dataset.
3. The method for dynamic testing and evaluation of vehicle braking performance as described in claim 2, characterized in that: The set of road braking digital parameters identified based on the braking micropulse time-series dataset includes: The vehicle's total mass and road gradient are estimated based on the non-braking and weak braking segments in the braking micro-pulse time series dataset. The dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road adhesion are also estimated based on the micro-pulse braking segments. The estimated total vehicle mass and road slope are updated using iterative constraints. An update is performed when the longitudinal acceleration deviation of the vehicle is greater than 0.02g. In each update, the change in the estimated total vehicle mass relative to the previous round does not exceed 4%, and the change in the estimated road slope relative to the previous round does not exceed 0.2°. A maximum of 5 rounds of updates are performed. The dynamic tire radius estimate, longitudinal equivalent stiffness estimate, brake set-off delay estimate, and road adhesion estimate are identified by time-series correlation. The time-series correlation identification includes comparing the change trend of the same parameter in the current segment with that in the previous segment according to the time sequence of the micro-pulse braking segments. The parameter update is confirmed to be valid only when the same parameter changes in the same direction in two or more consecutive micro-pulse braking segments and the change amplitude does not exceed 15% of the corresponding parameter value in the previous segment. The estimated values of vehicle gross weight, road slope, dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road surface adhesion are packaged into a set of road braking digital parameters in that order.
4. The method for dynamic testing and evaluation of vehicle braking performance as described in claim 3, characterized in that: The generated parameter credibility identifier includes, The current set of road braking digital parameters, along with the vehicle speed, braking request quantity, and time series in the corresponding segment, are input into the longitudinal state update process to reconstruct the predicted wheel speed changes and the predicted vehicle longitudinal acceleration changes step by step in discrete time. The predicted wheel speed change and the predicted vehicle longitudinal acceleration change are compared with the actual wheel speed change and the actual vehicle longitudinal acceleration change in the braking micropulse time series dataset, respectively. When the wheel speed residual is no greater than 0.8 km / h and the vehicle longitudinal acceleration residual is no greater than 0.03g, the residual consistency is deemed to meet the requirements. When the changes in the estimated total vehicle mass of two adjacent valid segments do not exceed 3%, the changes in the estimated road slope do not exceed 0.15°, the changes in the estimated dynamic tire radius do not exceed 1.5%, the changes in the estimated longitudinal equivalent stiffness do not exceed 10%, the changes in the estimated brake setup delay do not exceed 15ms, and the changes in the estimated road surface adhesion do not exceed 0.08, the time continuity requirement is met. When the update direction of the same parameter does not switch repeatedly in three consecutive valid segments and the maximum fluctuation does not exceed 12% of the current value of the corresponding parameter, the parameter update stability is deemed to meet the requirements. When the proportion of valid sampling points in a single segment is not less than 90%, and the number of micro-pulse braking segments used for identification in this round is not less than 3, the segment integrity is deemed to meet the requirements. Sub-item confidence scores are generated based on residual consistency, temporal continuity, parameter update stability, and fragment integrity. These scores are then aggregated and calculated according to the following weights: residual consistency weight 0.35, temporal continuity weight 0.20, parameter update stability weight 0.25, and fragment integrity weight 0.20, to obtain the parameter confidence score identifier. When the confidence level of any component falls below 0.60, stop updating the parameter corresponding to the confidence level of the component and call the corresponding parameter from the previous stable moment to write it back to the road braking digital parameter set; The previous stable moment is the moment when the parameter confidence index is not lower than 0.75 and the confidence index of each sub-item is not lower than 0.
60. When the parameter confidence index is not lower than 0.70, the road braking digital parameter set enters the counterfactual full braking simulation.
5. The method for dynamic testing and evaluation of vehicle braking performance as described in claim 4, characterized in that: Performing counterfactual full braking simulations based on road braking digital parameter sets includes, Generate a target braking request that differs from the braking request amount based on the vehicle speed. When the vehicle speed is between 30 km / h and 40 km / h, the target braking request is set to 40%; when the vehicle speed is between 40 km / h and 50 km / h, the target braking request is set to 50%. The vehicle braking counterfactual longitudinal solution model is constructed by calling the estimated total vehicle mass, road slope, dynamic tire radius, longitudinal equivalent stiffness, brake set-off delay, and road surface adhesion values from the road braking digital parameter set after filtering by parameter confidence identifier. The counterfactual longitudinal solution model for vehicle braking includes a parameter loading module, a time delay module, a wheel speed response module, a longitudinal state update module, and a termination determination module. Before the estimated braking establishment delay is reached, the vehicle is placed in an initial inertial decay state affected by road gradient and driving resistance. After the brake establishment delay estimate is reached, the target brake request is loaded into the wheel speed response module and the longitudinal state update module; Based on a discrete time step of 10ms, the wheel braking intensity at the current moment is determined according to the target braking request, road surface adhesion estimate, and longitudinal equivalent stiffness estimate. The wheel speed, vehicle speed, and vehicle displacement are updated by combining the dynamic tire radius estimate and road slope estimate. When the predicted tire slippage trend at a certain moment exceeds 15%, the wheel braking intensity at that moment will be limited to 95% of that at the previous moment. When the vehicle speed drops to 0.5 km / h, the zero-speed threshold is reached and the recursion ends. If the zero-speed judgment value is not reached after 8 seconds of continuous iteration, the 8-second mark is taken as the termination time of the deduction, and the counterfactual full braking trajectory of the vehicle is obtained.
6. The method for dynamic testing and evaluation of vehicle braking performance as described in claim 5, characterized in that: The output of dynamic test and evaluation results includes, Feature extraction is performed on the counterfactual full braking trajectory of a car to form the equivalent braking distance, equivalent average fully exerted deceleration, equivalent braking establishment time, and left and right wheel inconsistency risk index; The risk index of inconsistency between left and right wheels is calculated by combining the difference in wheel speed response between the left and right wheels in the micro-pulse braking segment and the difference in deceleration between the left and right wheels in the vehicle's counterfactual full braking trajectory. The equivalent braking distance, equivalent average fully exerted deceleration, equivalent braking establishment time, left and right wheel inconsistency risk index, and parameter reliability indicators are written into the vehicle braking dynamic inspection and evaluation record.
7. A dynamic testing and evaluation system for vehicle braking performance, employing the dynamic testing and evaluation method for vehicle braking performance as described in any one of claims 1 to 6, characterized in that: It includes a micropulse data construction module, a braking parameter joint identification module, and a counterfactual inference and evaluation module; The micro-pulse data construction module is used to continuously collect wheel speeds, longitudinal acceleration, vehicle speed, braking request quantity, and corresponding timestamps when the vehicle speed is in the dynamic test speed range of vehicle braking (30km / h to 50km / h). When the braking request quantity enters the 8% to 18% micro-pulse excitation range of vehicle braking and the duration is within the short window sampling range of vehicle braking (0.20s to 0.70s), the module acquires vehicle braking micro-pulse time series data according to a uniform sampling period of 10ms, and generates a braking micro-pulse time series dataset. The braking parameter joint identification module is used to identify a set of road braking digital parameters based on non-braking segments, weak braking segments, and micro-pulse braking segments in the braking micro-pulse time-series dataset. The set of road braking digital parameters includes, in a preset field order, estimated vehicle gross weight, estimated road slope, estimated dynamic tire radius, estimated longitudinal equivalent stiffness, estimated brake set-off delay, and estimated road surface adhesion. Based on the residual consistency, temporal continuity, parameter update stability, and segment integrity between the set of road braking digital parameters and the braking micro-pulse time-series dataset, a parameter confidence identifier is generated. The counterfactual inference and evaluation module is used to generate a target braking request different from the braking request quantity when the parameter confidence index is not lower than the lower confidence limit of the vehicle braking parameters. It then calls the road braking digital parameter set to construct a longitudinal counterfactual solution model for vehicle braking. Taking the target braking request as input, it recursively calculates vehicle speed decay, vehicle displacement accumulation, and wheel speed response of each wheel according to the discrete time step until the vehicle speed drops to the zero speed judgment value, thus obtaining the vehicle braking counterfactual full braking trajectory. The module extracts features from the vehicle braking counterfactual full braking trajectory and outputs dynamic verification and evaluation results including equivalent braking distance, equivalent average fully exerted deceleration, equivalent braking establishment time, left and right wheel inconsistency risk index, and parameter confidence index.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the dynamic testing and evaluation method for vehicle braking performance as described in any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the dynamic testing and evaluation method for vehicle braking performance as described in any one of claims 1 to 6.