Brake performance degradation prediction method and system based on cloud big data

Abstract

The application relates to the technical field of big data resource services, in particular to a brake performance decay prediction method and system based on cloud big data, which comprises the following steps: calculating brake pedal stroke, brake master cylinder pressure, four-wheel angular velocity, brake disc temperature and longitudinal deceleration uploaded by a vehicle-mounted brake controller, cutting a brake event according to pedal lifting and falling time, generating a brake event sequence, a brake energy decay track and a dynamic judgment boundary, generating a decay level and a brake force distribution instruction according to continuous overrunning coordinates, and forming a vehicle maintenance judgment basis. In the application, the brake pressure, wheel speed and temperature change relationship of the same vehicle are constrained by the cloud event sequence, the temperature rise lag and brake distance drift are uniformly mapped, the isolated fluctuation of emergency braking does not directly change the prediction conclusion, the identification gap caused by fixed period detection is reduced, and the continuity of brake performance decay trend judgment is improved.

Description

Technical Field

[0001] This invention relates to the field of big data resource service technology, and in particular to a method and system for predicting braking performance degradation based on cloud-based big data. Background Technology

[0002] The field of big data resource service technology involves data aggregation, indexing, correlation calculation, and result distribution for vehicle operation records, braking execution records, and cloud storage resources. It falls under the technical category of organizing and utilizing continuously generated industrial operation data within internet data services. Traditional methods for predicting brake performance degradation involve recording brake pressure, wheel speed changes, braking distance, and brake disc temperature at points such as the vehicle's brake pedal, brake lines, wheel speed components, and brake disc temperature measurement points. Data from the onboard recording device is periodically exported to a computer, and brake test tables are created according to a fixed testing cycle. Maintenance personnel then combine this data with brake pad thickness, brake fluid condition, and test bench curves to determine the trend of brake performance changes.

[0003] Traditional braking performance assessment relies on fixed-cycle test bench testing and manual record export. During continuous vehicle operation, brake pressure fluctuations, wheel end temperature rise lags, and braking distance drift are difficult to correlate within the same time axis. Cloud-based historical records are stacked according to vehicle number, lacking coupling constraints for brake pad wear and brake fluid thermal decay. Brake disc temperature drop intervals and wheel speed decrease slopes are recorded in a scattered manner and are difficult to calibrate. When the number of abnormal decay samples is limited, the judgment boundary is easily offset by a single emergency braking record. Maintenance reminders lag behind the performance degradation process. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method and system for predicting braking performance degradation based on cloud-based big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a braking performance degradation prediction method based on cloud-based big data, comprising the following steps: The system calculates the brake pedal travel, master cylinder pressure, four-wheel angular velocity, brake disc temperature, and longitudinal deceleration uploaded by the vehicle brake controller. The braking process is segmented according to the pedal lifting and lowering moments. The braking process is written into the braking event sequence built in the cloud, generating pressure change curves, wheel speed change curves, and temperature drop curves. The pressure build-up slope is calculated based on the pressure change curve, the wheel end deceleration slope is calculated based on the wheel speed change curve, and the temperature rise peak drop interval is calculated based on the temperature drop curve. The slope difference and interval difference of adjacent braking processes of the same vehicle are compared to generate a braking energy decay trajectory. The temperature rise lag value and braking distance drift value are calculated based on the braking energy decay trajectory. The temperature rise lag value and braking distance drift value are mapped to the same coordinate plane. A dynamic judgment boundary is constructed according to the median offset and discrete width of the coordinate distribution within the continuous driving mileage. The braking energy attenuation trajectory is input into the dynamic judgment boundary for judgment. Coordinate points that are located outside the dynamic judgment boundary and appear continuously are selected to generate an attenuation level. The braking force ratio of the front and rear axles and the hydraulic pressure build-up rate are adjusted according to the attenuation level to generate a braking force distribution command.

[0006] As a further aspect of the present invention, the generation process of the pressure change curve, the wheel speed change curve, and the temperature drop curve is specifically as follows: Acquire brake pedal travel, master cylinder pressure, four-wheel angular velocity, brake disc temperature and longitudinal deceleration. Align all data points coaxially according to the sampling timestamps. Delete data segments corresponding to missing timestamps, reverse pressure jumps and sudden wheel speed increases to create the original brake segment. The system detects the continuous rising edge and continuous falling edge of the brake pedal travel, determines the time corresponding to the continuous rising edge as the pedal lifting moment, determines the time corresponding to the continuous falling edge as the falling moment, segments the original braking process based on the pedal lifting moment and falling moment, and generates the braking process. The braking process is correlated with vehicle identification, mileage, and ambient temperature. The braking event sequence is written according to the occurrence time, and the temporal changes of master cylinder pressure, four-wheel angular velocity, and brake disc temperature are extracted to generate pressure change curves, wheel speed change curves, and temperature drop curves.

[0007] As a further aspect of the present invention, the process of generating the braking energy attenuation trajectory is specifically as follows: Obtain the pressure change curve and the wheel speed change curve, calculate the rate of increase of the brake master cylinder pressure relative to the sampling time according to the continuous pressure increase interval during the braking process, and calculate the rate of change of the wheel end deceleration relative to the sampling time according to the continuous decrease interval of the four wheel angular velocities, and establish a slope record. The temperature drop curve is obtained, the sampling time when the brake disc temperature reaches the peak temperature rise and the sampling time when it drops back to the stable temperature range are identified, the time interval between the two sampling times is calculated, and the time interval is bound to the vehicle identification and the time of the braking process to obtain the interval record. By comparing the slope records and interval records corresponding to adjacent braking processes of the same vehicle, the difference in pressure build-up slope, the difference in wheel-end deceleration slope, and the difference in temperature rise peak drop interval are extracted and continuously correlated according to the time of braking process to generate a braking energy decay trajectory.

[0008] As a further aspect of the present invention, the process of constructing the dynamic judgment boundary is specifically as follows: The braking energy attenuation trajectory is obtained, and the extension direction and extension amplitude of the temperature rise peak drop interval difference during continuous braking are analyzed. The extension amplitude is matched with the position of the brake disc temperature peak, the temperature rise lag value is calculated, and a temperature rise lag record is established. Obtain the difference in pressure build-up slope and wheel-end deceleration slope in the braking energy decay trajectory, calculate the braking distance change trend according to the correspondence between the initial braking velocity and longitudinal deceleration, compare the braking distance change trend with the historical stable braking distance, calculate the braking distance drift value, and obtain the drift calculation record. Map the temperature rise hysteresis value and braking distance drift value to the same coordinate plane, calculate the median offset and discrete width of the coordinate points relative to the stable coordinate center within the continuous driving mileage, and construct a dynamic judgment boundary by expanding the coordinate judgment range according to the median offset and discrete width.

[0009] As a further aspect of the present invention, the process of generating the braking force distribution command is as follows: Obtain the braking energy attenuation trajectory and the dynamic judgment boundary, input the coordinate points formed by the temperature rise hysteresis value and the braking distance drift value into the dynamic judgment boundary one by one, compare the positional relationship between the coordinate points and the dynamic judgment boundary, and filter the outer coordinate records. The number of occurrences, deviation distance, and deviation duration of the outer coordinate record during continuous braking are determined. The number of occurrences, deviation distance, and deviation duration are matched with preset level thresholds step by step, and the degree of braking performance degradation is determined according to the matching results to generate a degradation level. Calculate the front and rear axle braking force ratio correction and hydraulic pressure build-up rate correction corresponding to the attenuation level, write the front and rear axle braking force ratio correction and hydraulic pressure build-up rate correction into the control field that the vehicle brake controller can recognize, and generate a braking force distribution command.

[0010] As a further aspect of the present invention, the process for determining the timing of the pedal's lifting and lowering specifically includes: Calculate the brake pedal travel increment and travel drop within a continuous sampling window, eliminate small fluctuations below the pedal static vibration amplitude, and mark the sampling window where the travel increment continuously exceeds the lifting judgment threshold as a candidate lifting window to generate a lifting candidate record; Compare the pressure growth trend of the brake master cylinder pressure with the longitudinal deceleration response trend within the candidate lifting window, retain the candidate lifting windows where the pressure growth trend and longitudinal deceleration response trend continue in the same direction over time, extract the time corresponding to the start of the window, and obtain the pedal lifting time. The braking process boundary is generated by detecting the duration of the brake pedal travel decrease and the duration of the brake master cylinder pressure relief after the pedal rises. The starting point of the sampling window that simultaneously satisfies the drop judgment threshold is determined as the drop moment.

[0011] As a further aspect of the present invention, the comparison process of the slope difference and interval difference between adjacent braking processes of the same vehicle specifically includes: Obtain the pressure build-up slope, wheel end deceleration slope, and temperature rise peak drop interval for vehicles with the same identification, arranged by occurrence time. Filter out process pairs whose operating condition span exceeds the calibration threshold based on the difference in mileage and ambient temperature between adjacent braking processes, and generate adjacent comparison records. Calculate the decrease in built-in pressure slope, the decrease in wheel end deceleration slope, and the extension of the temperature rise peak drop interval in adjacent comparison records. Then, based on the braking initial velocity range, merge the decrease and extension intervals within the same range to obtain the attenuation difference record. The attenuation difference records are analyzed in terms of the direction of change and the number of times they last in continuous driving mileage. Attenuation difference records with the same direction of change and the number of times they last reach the trajectory formation threshold are associated by the time of occurrence to generate trajectory segments.

[0012] As a further aspect of the present invention, the calculation process for the median offset and the discrete width specifically includes: Obtain the temperature rise hysteresis value and braking distance drift value of the same vehicle within a continuous driving mileage, divide the coordinate points into multiple continuous mileage windows according to the direction of mileage increase, delete the continuous mileage windows with fewer coordinate points than the window stability threshold, and generate window coordinate records. Compare the offset distances of each coordinate point in the window coordinate record relative to the reference coordinate center, extract the representative offset value according to the median position of the offset distance sorting, and associate the representative offset value with the start and end mileage of the window to obtain the median offset; The dispersion of each coordinate point in the calculation window coordinate record around the representative offset value is corrected by combining the brake disc temperature fluctuation amplitude, and the corrected dispersion is mapped to the continuous mileage window to generate discrete width.

[0013] As a further aspect of the present invention, the process of associating the attenuation level with the braking force distribution command specifically includes: The number of coordinate points that appear consecutively outside the dynamic judgment boundary, the distance of the coordinate points from the outside, and the duration of the outside deviation are obtained. The level intervals are divided according to the joint sorting relationship of increasing number of coordinate points, increasing distance from the outside, and increasing duration of the outside deviation, and the attenuation level is generated. The correction direction of the attenuation level and the braking force ratio of the front and rear axles is matched, and the correction range is limited according to the current response speed of the brake master cylinder pressure and the current drop speed of the brake disc temperature to obtain the braking force ratio correction amount. The system correlates the braking force ratio correction, hydraulic pressure build-up rate correction, and the vehicle's current load status to perform consistency verification on the corrected front and rear axle braking force ratio and hydraulic pressure build-up rate, and then outputs a braking force distribution command.

[0014] A braking performance degradation prediction system based on cloud-based big data, the system being used to implement the aforementioned braking performance degradation prediction method based on cloud-based big data, the system comprising: The cloud sequence module calculates the brake pedal travel, brake master cylinder pressure, four-wheel angular velocity, brake disc temperature and longitudinal deceleration uploaded by the vehicle brake controller. It divides the braking process according to the moment the pedal rises and falls, writes the braking process into the braking event sequence built in the cloud, and generates pressure change curve, wheel speed change curve and temperature drop curve. The trajectory generation module calculates the pressure build-up slope based on the pressure change curve, the wheel-end deceleration slope based on the wheel speed change curve, and the temperature rise peak drop interval based on the temperature drop curve. It compares the slope difference and interval difference of adjacent braking processes of the same vehicle to generate a braking energy decay trajectory. The boundary construction module calculates the temperature rise lag value and braking distance drift value based on the braking energy decay trajectory, maps the temperature rise lag value and braking distance drift value to the same coordinate plane, and constructs a dynamic judgment boundary according to the median offset and discrete width of the coordinate distribution within the continuous driving mileage. The instruction generation module inputs the braking energy attenuation trajectory into the dynamic judgment boundary for judgment, filters out coordinate points that appear continuously outside the dynamic judgment boundary, generates attenuation level, adjusts the front and rear axle braking force ratio and hydraulic pressure build-up rate according to the attenuation level, and generates braking force distribution instruction.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, braking pressure, wheel angular velocity, brake disc temperature, and pedal travel directly collected from the vehicle are written into a cloud event sequence as braking events. A braking energy attenuation trajectory is constructed based on the historical sequence of the same vehicle. The wheel end temperature rise lag and braking distance drift are mapped into performance loss coordinates. Then, the wear slope of adjacent events is used to form a dynamic judgment boundary, so that isolated fluctuations caused by emergency braking do not directly change the attenuation conclusion. Based on the attenuation level, a braking force distribution correction command that can be executed by the vehicle controller is generated, reducing the time gap caused by fixed period detection and improving the ability of cloud recording to continuously identify the braking performance degradation process. Attached Figure Description

[0016] Figure 1This is the main flowchart for predicting braking performance degradation in this invention; Figure 2 This is a schematic diagram of the cloud-based braking data processing architecture of the present invention; Figure 3 This is a schematic diagram illustrating the braking event sequence and curve generation effect of the present invention; Figure 4 This is a schematic diagram illustrating the dynamic determination of boundaries and attenuation levels in this invention; Figure 5 This is a schematic diagram illustrating the feedback effect of the braking force distribution command of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the software-based technical solution is described in detail below with reference to system architecture diagrams and embodiments. It should be understood that the specific embodiments described herein are only for explaining the technical solutions of this invention and do not constitute a limitation on the scope of protection.

[0018] In the description of this invention, the system architecture relationships or data processing flows indicated by terms such as "layer," "module," "interface," "data flow," "client," and "server" are all defined based on the architecture diagram or flowchart corresponding to the embodiments. This way of describing is only used to clearly illustrate the logical relationships between the elements in the technical solution, and not to limit the physical deployment form. The term "multiple" includes two or more technical units, including but not limited to multiple data nodes, processing threads, service instances, or functional components and other scalable elements. The specific number is determined according to the actual business scenario and needs to be specifically specified.

[0019] Please see Figures 1 to 5 This embodiment provides a method for predicting brake performance degradation based on cloud-based big data. In practical applications, such as in a cloud-based brake monitoring environment where connected vehicles continuously upload brake control data, the vehicle-side brake controller generates data related to the pedal, hydraulic pressure, wheel speed, temperature, and deceleration when braking occurs. After receiving the above data, the cloud processing process segments, sorts, compares, performs boundary judgments, and transmits control commands back to the braking event, including the following steps: S1: Calculates the brake pedal travel, master cylinder pressure, wheel angular velocities, brake disc temperature, and longitudinal deceleration uploaded by the onboard brake controller. The braking process is segmented by the pedal lift-off and lowering moments, and written into a cloud-based braking event sequence, generating pressure change curves, wheel speed change curves, and temperature drop curves. Brake pedal travel is the pedal displacement change data output by the pedal position sensor; master cylinder pressure is the master cylinder hydraulic response data output by the hydraulic pressure sensor; wheel angular velocities are the wheel rotation state data output by the wheel speed sensors; brake disc temperature is the thermal state data output by the brake disc temperature acquisition channel; and longitudinal deceleration is the vehicle longitudinal deceleration data output by the vehicle inertial measurement channel or the brake controller fusion channel. Before entering the cloud processing, the above data carries a vehicle identifier, sampling timestamp, brake controller status identifier, and communication integrity identifier. The sampling timestamp is used to establish the timing alignment between channels, and the communication integrity identifier is used to identify transmission interruptions, duplicate uploads, and missing fields.

[0020] The braking event sequence is a time-series data object established in the cloud according to vehicle identification and the order of braking events. It contains braking process boundaries, brake pedal travel segments, master cylinder pressure segments, wheel angular velocity segments, brake disc temperature segments, longitudinal deceleration segments, mileage fields, and ambient temperature fields. The pressure change curve is a data structure derived from the master cylinder pressure segment showing pressure changes over time. The wheel speed change curve is a data structure derived from the wheel angular velocity segments showing wheel-end speed changes. The temperature drop curve is a data structure derived from the brake disc temperature segment showing temperature peaks and subsequent drop-offs. All of these curves are bound to vehicle identification and braking process boundaries and are incorporated into subsequent slope, interval, and trajectory generation processes.

[0021] S101: Acquire brake pedal travel, master cylinder pressure, wheel angular velocities, brake disc temperature, and longitudinal deceleration. Align all data points coaxially according to the sampling timestamps, deleting data segments corresponding to missing timestamps, reverse pressure jumps, and sudden wheel speed increases to create the original braking segment. The sampling timestamps are derived from the timing field of the message uploaded by the onboard brake controller. The coaxial alignment process uses the vehicle identifier and sampling timestamps as indexes to group pedal, pressure, wheel speed, temperature, and deceleration data into the same braking candidate interval. Missing timestamp fields are flagged as missing data, duplicate timestamp fields retain data records with higher communication integrity, and inconsistent format fields are flagged as parsing anomalies and saved in the anomaly log.

[0022] A reverse pressure jump is a sudden change in brake master cylinder pressure that occurs before the pedal travel returns to normal. A sudden increase in wheel speed is a short-term abnormality that occurs when wheel angular velocities are continuously reduced in the longitudinal direction. These abnormalities are identified by pressure continuity rules and wheel speed continuity rules in a cloud-based rule table. This rule table is derived from the vehicle brake controller's channel definitions, sensor range boundaries, and historical stable braking process statistics. Deleted data segments are not included in the curve generation process. The deletion location, field category, anomaly identifier, and vehicle identifier are written to the brake data quality log. The quality log is used to subsequently trace the input integrity of the brake event sequence. The cleaned data is synthesized into original brake segments in chronological order of occurrence. These original brake segments then enter the pedal boundary detection process.

[0023] S102: Detect the continuously rising and falling edges of the brake pedal travel. The moment corresponding to the continuously rising edge is determined as the pedal lift-off moment, and the moment corresponding to the continuously falling edge is determined as the fall-off moment. Based on the pedal lift-off and fall-off moments, the original braking segment is segmented to generate the braking process. A continuously rising edge is the edge data where the brake pedal travel increases progressively between adjacent sampling records, and the brake master cylinder pressure synchronously enters an increasing state. A continuously falling edge is the edge data where the brake pedal travel enters a falling state, and the brake master cylinder pressure synchronously enters a depressurization state. The edge detection process uses the direction of pedal travel change, the direction of pressure response, and the direction of longitudinal deceleration response as joint judgment objects to avoid segmenting invalid braking processes solely based on pedal signal jitter.

[0024] Both the pedal lift-off and lower-off moments are written as boundary fields of the braking process into the original braking segment. When the vehicle is in the initial startup phase and historical braking processes are insufficient, the stable stationary pedal state within the current sampling window uploaded by the onboard brake controller is used as the initial benchmark. After historical braking processes have been established, the pedal stationary vibration amplitude, pressure stability, and wheel speed stability of the most recent effective braking process are used as boundary determination references. If pedal travel is missing but the brake master cylinder pressure and longitudinal deceleration both form a continuous braking response, a boundary verification flag is generated and a candidate braking process is temporarily stored. It will enter the braking event sequence only after the pedal channel data is supplemented or the boundary is confirmed by the continuous pressure relief state.

[0025] S103: The braking process is correlated with vehicle identification, mileage, and ambient temperature. This information is written into the braking event sequence according to the time of occurrence. The temporal changes in master cylinder pressure, wheel angular velocity, and brake disc temperature are extracted to generate pressure change curves, wheel speed change curves, and temperature drop curves. The vehicle identification is obtained from the brake controller's identity field, the mileage from the onboard mileage record field, and the ambient temperature from the vehicle environment acquisition channel or the vehicle environment field received from the cloud. The correlation process uses the vehicle identification as the primary index and the occurrence time as the sorting index, writing the braking process of the same vehicle into the cloud braking event sequence in chronological order.

[0026] The pressure change curve is extracted from the brake master cylinder pressure field during the braking process, and the curve records the temporal relationship between the pressure increase segment, the pressure holding segment, and the pressure relief segment. The wheel speed change curve is extracted from the wheel angular velocity fields, and the curve records the temporal relationship between the wheel speed decrease segment and the recovery segment. The temperature drop curve is extracted from the brake disc temperature field, and the curve records the temporal relationship of the temperature rising to the peak and then falling back to a stable thermal state. If there is a short-term missing temperature field, a temperature to be supplemented marker is formed based on the adjacent effective temperature states within the same braking process. The conclusion of post-peak drop is not generated based on the missing field to avoid interference from abnormal segments in the subsequent temperature rise peak drop interval.

[0027] S104: Calculate the brake pedal travel increment and travel drop within a continuous sampling window, eliminate minor fluctuations below the pedal stationary vibration amplitude, and mark sampling windows whose travel increment continuously exceeds the lifting judgment threshold as candidate lifting windows, generating candidate lifting records. The pedal stationary vibration amplitude is derived from the stable range record of the pedal position sensor under the same vehicle's non-braking state, and the lifting judgment threshold is derived from the vehicle brake controller calibration table and historical effective braking process boundary records. The candidate lifting window incorporates the pedal travel change direction, travel change continuity, pressure response pre-state, and longitudinal deceleration initial state, serving as input for the pedal lifting moment confirmation process.

[0028] S105: Compare the pressure growth trend of the brake master cylinder pressure with the longitudinal deceleration response trend within the candidate lifting window. Retain candidate lifting windows where the pressure growth trend and longitudinal deceleration response trend continue in the same direction over time. Extract the time corresponding to the window's starting point to obtain the pedal lifting time. The pressure growth trend is determined by the continuous growth segment in the pressure change curve, and the longitudinal deceleration response trend is determined by the braking response segment in the vehicle deceleration data. If a pressure growth trend exists but a longitudinal deceleration response trend has not formed, the candidate lifting window is marked as a hydraulic preload state and is not temporarily used as the starting point of the complete braking process. If a longitudinal deceleration response trend exists but the pressure growth trend field is missing, enter the pressure channel anomaly verification state and retain the original segment traceability record.

[0029] S106: Detect the duration of the brake pedal travel descent and the duration of the brake master cylinder pressure relief after the pedal lift-off moment. The starting point of the sampling window where both the duration of the travel descent and the duration of pressure relief simultaneously meet the drop-off judgment threshold is determined as the drop-off moment, generating the braking process boundary. The drop-off judgment threshold is derived from the vehicle brake controller's pedal return calibration, hydraulic pressure relief calibration, and historical stable braking process records in the cloud. The braking process boundary includes the pedal lift-off moment, drop-off moment, boundary confirmation state, and anomaly verification state. The boundary confirmation state is entered into the braking event sequence sorting, and the anomaly verification state is entered into the data quality log and restricted from subsequent trajectory generation process calls.

[0030] S2: Calculate the pressure build-up slope based on the pressure change curve, the wheel-end deceleration slope based on the wheel speed change curve, and the temperature rise peak drop interval based on the temperature drop curve. Compare the slope differences and interval differences of adjacent braking processes for the same vehicle to generate a braking energy decay trajectory. The pressure build-up slope is the rate of increase of the brake master cylinder pressure during the braking process relative to the sampling order; the wheel-end deceleration slope is the rate of decrease of each wheel angular velocity during the braking response phase relative to the sampling order; and the temperature rise peak drop interval is the time interval corresponding to the brake disc temperature dropping from the peak thermal state to the stable thermal state. All of the above fields are derived from the pressure change curve, wheel speed change curve, and temperature drop curve generated in S1, and include vehicle identification, braking process boundaries, mileage, and ambient temperature.

[0031] The braking energy decay trajectory is a data object formed by associating the differences in pressure build-up slope, wheel-end deceleration slope, and temperature rise peak drop interval in sequence during continuous braking of the same vehicle. This trajectory data object does not directly use the original single braking curve as the basis for decay judgment; instead, it is constructed based on the direction of change, duration of change, and consistency of operating conditions between adjacent braking processes. If the span of driving state between adjacent braking processes exceeds the allowable boundary of the rule table, that process will not be included in the continuous trajectory association and will only be retained as a single braking record.

[0032] S201: Acquire pressure change curves and wheel speed change curves. Calculate the rate of increase of the master cylinder pressure relative to the sampling time based on the continuous pressure increase intervals during braking. Calculate the rate of change of wheel-end deceleration relative to the sampling time based on the continuous decrease intervals of each wheel angular velocity, and establish slope records. The continuous pressure increase interval is identified from the pressure change curve, based on the pressure field continuously entering an increasing state without any pressure relief indicator. The continuous decrease intervals of each wheel angular velocity are identified from the wheel speed change curve, based on each wheel angular velocity entering a decreasing state with the braking response and the longitudinal deceleration existing synchronously.

[0033] The slope record includes vehicle identification, braking process boundaries, pressure build-up slope field, wheel-end deceleration slope field, pressure increase status, and wheel speed decrease status. If there is an abnormal jump in the angular velocity of a single wheel, the abnormal wheel end is marked as abnormal in wheel speed; other wheel ends still participate in the formation of wheel-end deceleration slope, and the abnormal wheel end is not used as the primary source for determining the attenuation trajectory. The slope record is included in the comparison process of adjacent braking processes and is bound to the temperature rise peak drop interval record for the same braking process.

[0034] S202: Obtain the temperature drop curve, identify the sampling time when the brake disc temperature reaches the peak temperature rise and the sampling time when it drops back to the stable temperature range, calculate the time interval between the sampling times, and bind the time interval with the vehicle identifier and the braking process occurrence time to obtain the interval record. The peak temperature rise is the position field of the brake disc temperature entering the peak thermal state in the temperature drop curve, and the stable temperature range is a lookup table field formed by the historical stable braking process and brake disc cooling record of the same vehicle. When the temperature drop curve has missing, repeated, or abnormal peaks, the temperature quality state is first generated by the temperature continuity rule. Braking processes whose temperature quality state does not meet the validity boundary are not generated as interval records.

[0035] The interval record contains fields for the peak temperature rise time, the stable thermal state time, temperature quality status, vehicle identification, and the braking process occurrence time. The interval record is then used for comparison of adjacent braking processes, forming trajectory point inputs together with the pressure build-up slope field and the wheel-end deceleration slope field. The cloud storage process retains the source curve index and quality log index for the interval record, allowing for retrieval of the original segment of the temperature drop curve when abnormal outward deviations occur during subsequent dynamic boundary checks.

[0036] S203: Compare the slope records and interval records corresponding to adjacent braking processes of the same vehicle, extract the difference in pressure build-up slope, the difference in wheel-end deceleration slope, and the difference in temperature rise peak drop interval, and continuously correlate them according to the braking process occurrence time to generate a braking energy decay trajectory. Adjacent braking processes are jointly determined by the vehicle identification, occurrence time, and mileage fields. The comparison process first verifies whether the ambient temperature, initial braking speed range, and driving state category belong to comparable operating conditions, and then sorts out the differences in the pressure build-up slope field, wheel-end deceleration slope field, and temperature rise peak drop interval field.

[0037] The pressure build-up slope difference indicates the change in master cylinder pressure build-up response between adjacent braking processes; the wheel-end deceleration slope difference indicates the change in wheel-end speed decrease response between adjacent braking processes; and the temperature rise peak drop interval difference indicates the change in brake disc thermal drop between adjacent braking processes. The continuous correlation process writes these differences into trajectory segments according to their occurrence time. Once the trajectory segments reach a stable and continuous state, a braking energy decay trajectory is formed. If there are mismatched operating conditions, insufficient field quality, or vehicle identifier conflicts in the comparison records of adjacent braking processes, a trajectory interruption status is output, and the reason for the interruption is retained. Subsequent dynamic boundary judgments do not accept the coordinate points corresponding to the interruption status.

[0038] S204: Obtain the pressure build-up slope, wheel-end deceleration slope, and temperature rise peak drop interval for vehicles with the same identification, arranged by occurrence time. Filter out process pairs whose operating condition span exceeds the calibration threshold based on the difference in mileage and ambient temperature between adjacent braking processes, generating adjacent comparison records. The calibration threshold is derived from the vehicle braking test calibration table, the cloud-based historical stable braking event database, and ambient temperature stratification rules. Adjacent comparison records carry the comparable operating condition status, the quality status of the slope field, the quality status of the interval field, and the reason for operating condition filtering, serving as input for subsequent attenuation difference processing.

[0039] S205: Calculate the decrease in the built-up pressure slope, the decrease in the wheel-end deceleration slope, and the extension of the temperature rise peak drop interval in adjacent comparison records. Then, based on the initial braking velocity range, merge the decrease and extension intervals within the same range to obtain the attenuation difference record. The initial braking velocity range originates from the wheel speed change curve and longitudinal deceleration response state near the starting boundary of the braking process. The interval merging process organizes the differences between different initial braking states to the same working condition caliber. The attenuation difference record carries changes in the built-up pressure response, wheel-end deceleration response, temperature drop, and working condition merging identifiers, and enters the trajectory segment formation process.

[0040] S206: Analyze the direction and duration of the brake attenuation difference records over continuous driving mileage. Attenuation difference records with consistent direction of change and duration meeting the trajectory formation rules are associated by occurrence time to generate trajectory segments. The trajectory formation rules are derived from a cloud-based brake attenuation determination rule table, which is updated jointly by vehicle calibration data, historical stable braking events, and controller fault diagnosis boundaries. After trajectory segments are generated, they are written to a brake energy attenuation trajectory cache. The cache is configured with vehicle identification access permissions, log write permissions, and version status. The dynamic boundary construction process only reads trajectory segments with valid version status.

[0041] S3: Calculate the temperature rise lag value and braking distance drift value based on the braking energy decay trajectory. Map these values ​​to the same coordinate plane and construct a dynamic judgment boundary based on the median offset and discrete width of the coordinate distribution within the continuous driving mileage. The temperature rise lag value is a thermal response delay field formed by combining the temperature rise peak drop interval difference, the brake disc temperature peak position, and the continuous state of the braking process. The braking distance drift value is a distance response offset field formed by combining the pressure build-up slope difference, the wheel end deceleration slope difference, the initial braking speed state, the longitudinal deceleration response state, and historical stable braking distance records.

[0042] The dynamically determined boundary is a boundary data object formed in the cloud based on the distribution changes of coordinate points of the same vehicle within a continuous driving mileage. It carries the stable coordinate center, median offset, discrete width, boundary version status, and callable status. The median offset is a representative offset field of the coordinate point relative to the stable coordinate center, and the discrete width is a field representing the degree of dispersion of the coordinate point around the offset state. The dynamically determined boundary enters the coordinate point outer filtering process of S4 and is updated with the valid trajectory segments within the continuous driving mileage.

[0043] S301: Obtain the braking energy decay trajectory, analyze the extension direction and magnitude of the temperature rise peak drop interval difference during continuous braking, match the extension magnitude with the brake disc temperature peak position, calculate the temperature rise lag value, and establish a temperature rise lag record. The temperature rise peak drop interval difference comes from the braking energy decay trajectory generated in S2, and the brake disc temperature peak position comes from the peak thermal state field in the temperature drop curve. The extension direction indicates whether the temperature drop continues to change towards the lag state, and the extension magnitude indicates the degree of deviation of the temperature drop process relative to the historical stable thermal state.

[0044] The temperature rise hysteresis record contains the temperature rise hysteresis value, temperature quality status, peak position status, vehicle identification, and mileage fields. If the brake disc temperature peak position is missing or there is a missing marker on the temperature drop curve, the temperature rise hysteresis record enters a pending confirmation state. In this pending confirmation state, the record is not included in the dynamic boundary calculation. The temperature rise hysteresis record is written to the cloud-based trajectory-derived data area and bound to the drift calculation record of the same braking process.

[0045] S302: Obtain the difference in pressure build-up slope and wheel-end deceleration slope in the braking energy decay trajectory. Calculate the braking distance change trend based on the correspondence between initial braking velocity and longitudinal deceleration. Compare the braking distance change trend with historical stable braking distances to calculate the braking distance drift value and obtain the drift calculation record. The initial braking velocity comes from the speed state of the wheel speed change curve at the braking initiation boundary. The longitudinal deceleration comes from the vehicle body deceleration response field during the braking process. The historical stable braking distance comes from the stable braking records formed by the same vehicle under effective operating conditions. The braking distance change trend is formed by organizing the textual rules between pressure response changes, wheel-end deceleration changes, and vehicle body deceleration responses, and does not use a single channel for independent judgment.

[0046] The drift calculation record contains the braking distance drift value, initial braking velocity state, longitudinal deceleration state, slope mass state, and a version of the historical stable record. If the historical stable braking distance record is insufficient at the initial stage of vehicle startup, the currently confirmed effective braking process is used to form the initial stable record; as effective braking processes gradually accumulate, the initial stable record is replaced by the stable operation record. The drift calculation record enters the coordinate mapping process and serves as one of the lateral or longitudinal distribution fields for dynamic judgment boundaries.

[0047] S303: Maps the temperature rise hysteresis value and braking distance drift value to the same coordinate plane, calculates the median offset and discrete width of the coordinate points relative to the stable coordinate center within a continuous driving mileage, and expands the coordinate judgment range according to the median offset and discrete width to construct a dynamic judgment boundary. The same coordinate plane is a dimensionless state plane established by the cloud for the temperature rise hysteresis value and braking distance drift value. Before entering the coordinate plane, both fields are normalized according to the same vehicle's historical stable state and current operating condition state to avoid direct cross-dimensional comparison between the thermal response field and the distance response field.

[0048] The stable coordinate center is derived from the temperature rise hysteresis record and drift calculation record of the same vehicle's historical stable braking process. The median offset is derived from the representative position of the coordinate point's offset state within a continuous driving mileage window. The discrete width is derived from the dispersion state of the coordinate points around the representative position. The dynamic boundary judgment forms a boundary version based on the stable coordinate center, median offset, and discrete width, and writes it to the cloud boundary data area. The boundary version carries the vehicle identifier, update status, input record index, and quality status. S4 only calls the valid boundary version.

[0049] S304: Obtain the temperature rise hysteresis and braking distance drift values ​​of the same vehicle within a continuous driving mileage. Divide the coordinate points into continuous mileage windows according to the direction of increasing mileage. Delete continuous mileage windows with fewer coordinate points than the window stability threshold, and generate window coordinate records. The window stability threshold is derived from the cloud boundary judgment rule table and the statistical caliber of vehicle braking event validity. The threshold field is stored in the rule configuration area and is jointly limited by vehicle type, braking system calibration status, and historical stability record version. The window coordinate record carries the coordinate point set status, window validity status, and deletion reason.

[0050] S305: Compare the offset distances of each coordinate point in the window coordinate record relative to the reference coordinate center. Extract the representative offset value according to the median position of the sorted offset distances, and associate the representative offset value with the start and end mileages of the window to obtain the median offset. The reference coordinate center is derived from the stable distribution state of the coordinate points corresponding to the historical stable braking process. The sorted median position is a representative position field in the rule table used to offset the influence of occasional abnormal coordinate points. The median offset is bound to the start and end mileages of the window and the boundary version status, and then enters the dynamic judgment of boundary expansion process.

[0051] S306: Calculate the dispersion of each coordinate point in the window coordinate record around the representative offset value. Correct the dispersion by incorporating the brake disc temperature fluctuation amplitude, and map the corrected dispersion to the continuous mileage window to generate the discrete width. The brake disc temperature fluctuation amplitude is derived from the temperature quality state and post-peak decline state corresponding to the temperature drop curve. The correction rules are derived from the brake disc thermal state calibration table and historical stable thermal response records. The discrete width serves as a boundary extension field for dynamic judgment boundaries, and this boundary extension field enters the outer coordinate record filtering process in S4.

[0052] S4: The braking energy attenuation trajectory is input into the dynamic judgment boundary for judgment. Coordinate points that appear continuously outside the dynamic judgment boundary are selected to generate an attenuation level. Based on the attenuation level, the front and rear axle braking force ratio and hydraulic pressure build-up rate are adjusted to generate a braking force distribution command. The attenuation level is a level field formed by classifying the continuous state, outward deviation distance state, and outward deviation duration state of the outer coordinate records. The level field is used to drive the correction of the front and rear axle braking force ratio and the correction of the hydraulic pressure build-up rate. The braking force distribution command is a set of control fields that are sent back to the vehicle brake controller from the cloud, carrying the front and rear axle braking force ratio correction amount, hydraulic pressure build-up rate correction amount, command version, validity status, and confirmation feedback status.

[0053] The dynamic boundary judgment process takes the valid boundary version generated by S3 as input and the coordinate points corresponding to the braking energy attenuation trajectory as the judgment object. If the boundary version fails, the trajectory segment is interrupted, or the quality of the coordinate point source record is insufficient, the command suppression state is output while the existing braking control strategy on the current vehicle side is retained. If the coordinate points are continuously located outside the boundary, the attenuation level generation process is initiated. After the control command is generated, it is sent out through the vehicle-to-cloud communication interface, and the vehicle-side confirmation field is sent back to the cloud and written to the command traceability log.

[0054] S401: Obtain the braking energy decay trajectory and dynamic judgment boundary. Input the coordinate points formed by the temperature rise hysteresis value and braking distance drift value into the dynamic judgment boundary one by one. Compare the positional relationship between the coordinate points and the dynamic judgment boundary, and filter the outer coordinate records. The outer coordinate records carry the coordinate point position status, external offset distance status, boundary version, trajectory segment index, and data quality status. The comparison process first verifies whether the boundary version is consistent with the vehicle identifier of the coordinate point, and then verifies whether the mileage corresponding to the coordinate point falls within the applicable window of the boundary version. Coordinate points that meet the above conditions enter the positional relationship judgment.

[0055] If the coordinate point is located inside the dynamic judgment boundary, a stable coordinate state is generated and subsequent trajectory segments are monitored. If the coordinate point is located outside the dynamic judgment boundary, an outer coordinate state is generated and written into the attenuation level candidate queue. If the same coordinate point simultaneously has a temperature pending confirmation state or a drift calculation pending confirmation state, a coordinate state pending verification is generated, and it does not enter the outer coordinate record filtering. It will only enter the judgment process after the corresponding record has completed quality confirmation.

[0056] S402: Determine the number of occurrences, outward deviation distance, and outward deviation duration of the outer coordinate record during continuous braking. Match these values ​​with preset level thresholds step by step, and determine the degree of braking performance degradation based on the matching results, generating a degradation level. The preset level thresholds are derived from the vehicle braking system calibration table, historical braking degradation records in the cloud, and controller safety boundary rules. The threshold field is stored in the cloud rule configuration area and is invoked based on vehicle type, braking system version, and boundary version status.

[0057] The occurrence count field indicates the continuous occurrence of the outer coordinate record during continuous braking; the offset distance field indicates the offset of the coordinate point relative to the dynamic judgment boundary; and the offset duration field indicates the continuation of the outer coordinate state along the driving mileage. The step-by-step matching process is executed in a fixed order of outer coordinate record continuity, offset distance state, and offset duration state. If there is a conflict between rules, the higher-level state defined by the safety boundary rule is used as the output. The attenuation level is written into the control strategy candidate area and, together with the current vehicle load state, the current brake disc temperature drop rate, and the current brake master cylinder pressure response rate, enters the instruction generation process.

[0058] S403: Calculate the front and rear axle braking force ratio correction and hydraulic pressure build-up rate correction corresponding to the attenuation level. Write these corrections into a control field recognizable by the onboard brake controller to generate a braking force distribution command. The front and rear axle braking force ratio correction is a control field derived from the attenuation level, the vehicle's current load state, and braking stability. The hydraulic pressure build-up rate correction is a control field derived from the attenuation level, the current response speed of the master cylinder pressure, and the current rate of temperature drop in the brake disc. These control fields originate from a cloud-based control strategy table, which is maintained by the vehicle braking system calibration rules and safety boundary rules.

[0059] The braking force distribution command is written into the downlink message of the vehicle-to-cloud communication system. The message includes the vehicle identifier, command version, front and rear axle braking force ratio correction field, hydraulic pressure build-up rate correction field, validity status, and confirmation feedback requirements. If the vehicle returns confirmation feedback, the cloud writes the command version, feedback status, and corresponding attenuation level into the command traceability log. If the vehicle does not return confirmation feedback or the feedback status is inconsistent with the command version, an unconfirmed command status is generated, and the attenuation level and boundary version are re-verified during the next braking event sequence update.

[0060] S404: Obtain the number of coordinate points that appear consecutively outside the dynamic judgment boundary, the offset distance of the coordinate points, and the duration of the offset. Divide the grade intervals according to the joint sorting relationship of increasing coordinate point number, increasing offset distance, and increasing offset duration, and generate the attenuation grade. The coordinate point number field represents the continuous accumulation state of the outer coordinate records, the coordinate point offset distance field represents the offset state of the coordinate points from the boundary, and the offset duration field represents the continuous state of the outer state along the driving mileage. The joint sorting relationship is derived from the grade rule table, which is stored in the cloud control strategy area and updated by the vehicle braking safety boundary.

[0061] S405: Matches the correction direction for the brake fade level and the front-to-rear axle braking force ratio. The correction magnitude is limited by the current response speed of the master cylinder pressure and the current rate of temperature drop of the brake disc, resulting in the brake force ratio correction. The current response speed of the master cylinder pressure is derived from the latest pressure change curve, and the current rate of temperature drop of the brake disc temperature is derived from the latest temperature drop curve. The correction magnitude limitation rules are derived from the brake system calibration table and safety control boundaries. The brake force ratio correction is entered into the control field writing process and is simultaneously checked for consistency with the hydraulic pressure build-up rate correction.

[0062] S406: This function correlates the brake force ratio correction, hydraulic pressure build-up rate correction, and the vehicle's current load status. It performs a consistency check on the corrected front and rear axle brake force ratio and hydraulic pressure build-up rate, and outputs a brake force distribution command. The vehicle's current load status is derived from the load estimation field or suspension status field uploaded by the onboard controller. The consistency check verifies whether the brake force ratio correction and hydraulic pressure build-up rate correction are consistent with the vehicle's current load status, braking stability status, and safety boundary rules. If the check passes, a brake force distribution command is output; if the check fails, a command suppression status and the reason for the failure are output. The command suppression status is written to the cloud traceability log and fed back to the onboard brake controller.

[0063] The braking performance degradation prediction system based on cloud-based big data, implemented in conjunction with the aforementioned method, is deployed in a cloud-based braking monitoring platform and a vehicle-to-cloud communication link. The system includes a cloud sequence module, a trajectory generation module, a boundary construction module, and a command generation module. The cloud sequence module is responsible for data reception, timestamp alignment, braking process segmentation, braking event sequence writing, and generation of pressure change curves, wheel speed change curves, and temperature drop curves in S1. Its input interface receives pedal, pressure, wheel speed, temperature, and deceleration fields uploaded by the onboard brake controller, and its output interface sends curve data, braking process boundaries, and data quality status to the trajectory generation module.

[0064] The trajectory generation module handles the processing of pressure build-up slope field, wheel-end deceleration slope field, and temperature rise peak drop interval field from S2, as well as comparison of adjacent braking processes and generation of braking energy decay trajectory. This module does not modify the original curves from the cloud sequence module; it only reads the valid curve version and braking process boundaries, outputting the braking energy decay trajectory, trajectory segment status, and trajectory quality status. The boundary construction module handles the generation of temperature rise hysteresis records, drift calculation records, coordinate plane mapping, median offset processing, discrete width processing, and dynamic boundary construction from S3. Its input is the braking energy decay trajectory, and its output is the boundary version, boundary quality status, and callable status.

[0065] The instruction generation module handles the outer coordinate record filtering, attenuation level generation, front and rear axle braking force ratio correction, hydraulic pressure build-up rate correction, control field writing, and braking force distribution instruction output in S4. The instruction generation module connects to the vehicle-to-cloud communication interface, which is responsible for downlink instruction transmission, vehicle-side confirmation feedback reception, and instruction version status write-back. The system's log storage area records braking data quality logs, trajectory interruption reasons, boundary version update status, and instruction confirmation status. Access permissions are jointly limited by the vehicle identifier, cloud processing task identifier, and control strategy version. These modules carry the corresponding processing content according to the data flow sequence of the method steps, maintaining consistency in data source, processing status, output results, and feedback path under the same vehicle identifier.

[0066] The above embodiments illustrate preferred embodiments of the present invention. Any equivalent adjustments to the technical solution based on software engineering methods are within the scope of protection, including but not limited to: implementing algorithm logic using different programming languages, refactoring functional modules into services, adjusting data interaction protocols, and optimizing resource scheduling strategies. Any implementation scheme derived from reasonable modifications to the data processing flow, service call chain, or system architecture layer without departing from the core technology of the present invention should be considered within the scope of protection defined by the claims of the present invention.

Claims

1. A method for predicting braking performance degradation based on cloud-based big data, characterized in that, Includes the following steps: The system calculates the brake pedal travel, master cylinder pressure, four-wheel angular velocity, brake disc temperature, and longitudinal deceleration uploaded by the vehicle brake controller. The braking process is segmented according to the pedal lifting and lowering moments. The braking process is written into the braking event sequence built in the cloud, generating pressure change curves, wheel speed change curves, and temperature drop curves. The pressure build-up slope is calculated based on the pressure change curve, the wheel end deceleration slope is calculated based on the wheel speed change curve, and the temperature rise peak drop interval is calculated based on the temperature drop curve. The slope difference and interval difference of adjacent braking processes of the same vehicle are compared to generate a braking energy decay trajectory. The temperature rise lag value and braking distance drift value are calculated based on the braking energy decay trajectory. The temperature rise lag value and braking distance drift value are mapped to the same coordinate plane. A dynamic judgment boundary is constructed according to the median offset and discrete width of the coordinate distribution within the continuous driving mileage. The braking energy attenuation trajectory is input into the dynamic judgment boundary for judgment. Coordinate points that are located outside the dynamic judgment boundary and appear continuously are selected to generate an attenuation level. The braking force ratio of the front and rear axles and the hydraulic pressure build-up rate are adjusted according to the attenuation level to generate a braking force distribution command.

2. The braking performance degradation prediction method based on cloud-based big data according to claim 1, characterized in that, The generation process of the pressure change curve, the wheel speed change curve, and the temperature drop curve is as follows: Acquire brake pedal travel, master cylinder pressure, four-wheel angular velocity, brake disc temperature and longitudinal deceleration. Align all data points coaxially according to the sampling timestamps. Delete data segments corresponding to missing timestamps, reverse pressure jumps and sudden wheel speed increases to create the original brake segment. The system detects the continuous rising edge and continuous falling edge of the brake pedal travel, determines the time corresponding to the continuous rising edge as the pedal lifting moment, determines the time corresponding to the continuous falling edge as the falling moment, segments the original braking process based on the pedal lifting moment and falling moment, and generates the braking process. The braking process is correlated with vehicle identification, mileage, and ambient temperature. The braking event sequence is written according to the occurrence time, and the temporal changes of master cylinder pressure, four-wheel angular velocity, and brake disc temperature are extracted to generate pressure change curves, wheel speed change curves, and temperature drop curves.

3. The braking performance degradation prediction method based on cloud-based big data according to claim 2, characterized in that, The specific process for generating the braking energy decay trajectory is as follows: Obtain the pressure change curve and the wheel speed change curve, calculate the rate of increase of the brake master cylinder pressure relative to the sampling time according to the continuous pressure increase interval during the braking process, and calculate the rate of change of the wheel end deceleration relative to the sampling time according to the continuous decrease interval of the four wheel angular velocities, and establish a slope record. The temperature drop curve is obtained, the sampling time when the brake disc temperature reaches the peak temperature rise and the sampling time when it drops back to the stable temperature range are identified, the time interval between the two sampling times is calculated, and the time interval is bound to the vehicle identification and the time of the braking process to obtain the interval record. By comparing the slope records and interval records corresponding to adjacent braking processes of the same vehicle, the difference in pressure build-up slope, the difference in wheel-end deceleration slope, and the difference in temperature rise peak drop interval are extracted and continuously correlated according to the time of braking process to generate a braking energy decay trajectory.

4. The braking performance degradation prediction method based on cloud-based big data according to claim 3, characterized in that, The process of constructing the dynamic boundary judgment is as follows: The braking energy attenuation trajectory is obtained, and the extension direction and extension amplitude of the temperature rise peak drop interval difference during continuous braking are analyzed. The extension amplitude is matched with the position of the brake disc temperature peak, the temperature rise lag value is calculated, and a temperature rise lag record is established. Obtain the difference in pressure build-up slope and wheel-end deceleration slope in the braking energy decay trajectory, calculate the braking distance change trend according to the correspondence between the initial braking velocity and longitudinal deceleration, compare the braking distance change trend with the historical stable braking distance, calculate the braking distance drift value, and obtain the drift calculation record. Map the temperature rise hysteresis value and braking distance drift value to the same coordinate plane, calculate the median offset and discrete width of the coordinate points relative to the stable coordinate center within the continuous driving mileage, and construct a dynamic judgment boundary by expanding the coordinate judgment range according to the median offset and discrete width.

5. The braking performance degradation prediction method based on cloud-based big data according to claim 4, characterized in that, The specific process for generating the braking force distribution command is as follows: Obtain the braking energy attenuation trajectory and the dynamic judgment boundary, input the coordinate points formed by the temperature rise hysteresis value and the braking distance drift value into the dynamic judgment boundary one by one, compare the positional relationship between the coordinate points and the dynamic judgment boundary, and filter the outer coordinate records. The number of occurrences, deviation distance, and deviation duration of the outer coordinate record during continuous braking are determined. The number of occurrences, deviation distance, and deviation duration are matched with preset level thresholds step by step, and the degree of braking performance degradation is determined according to the matching results to generate a degradation level. Calculate the front and rear axle braking force ratio correction and hydraulic pressure build-up rate correction corresponding to the attenuation level, write the front and rear axle braking force ratio correction and hydraulic pressure build-up rate correction into the control field that the vehicle brake controller can recognize, and generate a braking force distribution command.

6. The braking performance degradation prediction method based on cloud-based big data according to claim 5, characterized in that, The process of determining the timing of the pedal's ascent and descent specifically includes: Calculate the brake pedal travel increment and travel drop within a continuous sampling window, eliminate small fluctuations below the pedal static vibration amplitude, and mark the sampling window where the travel increment continuously exceeds the lifting judgment threshold as a candidate lifting window to generate a lifting candidate record; Compare the pressure growth trend of the brake master cylinder pressure with the longitudinal deceleration response trend within the candidate lifting window, retain the candidate lifting windows where the pressure growth trend and longitudinal deceleration response trend continue in the same direction over time, extract the time corresponding to the start of the window, and obtain the pedal lifting time. The braking process boundary is generated by detecting the duration of the brake pedal travel decrease and the duration of the brake master cylinder pressure relief after the pedal rises. The starting point of the sampling window that simultaneously satisfies the drop judgment threshold is determined as the drop moment.

7. The braking performance degradation prediction method based on cloud-based big data according to claim 6, characterized in that, The comparison process of the slope difference and interval difference between adjacent braking processes of the same vehicle specifically includes: Obtain the pressure build-up slope, wheel end deceleration slope, and temperature rise peak drop interval for vehicles with the same identification, arranged by occurrence time. Filter out process pairs whose operating condition span exceeds the calibration threshold based on the difference in mileage and ambient temperature between adjacent braking processes, and generate adjacent comparison records. Calculate the decrease in built-in pressure slope, the decrease in wheel end deceleration slope, and the extension of the temperature rise peak drop interval in adjacent comparison records. Then, based on the braking initial velocity range, merge the decrease and extension intervals within the same range to obtain the attenuation difference record. The attenuation difference records are analyzed in terms of the direction of change and the number of times they last in continuous driving mileage. Attenuation difference records with the same direction of change and the number of times they last reach the trajectory formation threshold are associated by the time of occurrence to generate trajectory segments.

8. The braking performance degradation prediction method based on cloud-based big data according to claim 7, characterized in that, The calculation process for median offset and discrete width specifically includes: Obtain the temperature rise hysteresis value and braking distance drift value of the same vehicle within a continuous driving mileage, divide the coordinate points into multiple continuous mileage windows according to the direction of mileage increase, delete the continuous mileage windows with fewer coordinate points than the window stability threshold, and generate window coordinate records. Compare the offset distances of each coordinate point in the window coordinate record relative to the reference coordinate center, extract the representative offset value according to the median position of the offset distance sorting, and associate the representative offset value with the start and end mileage of the window to obtain the median offset; The dispersion of each coordinate point in the calculation window coordinate record around the representative offset value is corrected by combining the brake disc temperature fluctuation amplitude, and the corrected dispersion is mapped to the continuous mileage window to generate discrete width.

9. The braking performance degradation prediction method based on cloud-based big data according to claim 8, characterized in that, The process of associating the attenuation level with the braking force distribution command specifically includes: The number of coordinate points that appear consecutively outside the dynamic judgment boundary, the distance of the coordinate points from the outside, and the duration of the outside deviation are obtained. The level intervals are divided according to the joint sorting relationship of increasing number of coordinate points, increasing distance from the outside, and increasing duration of the outside deviation, and the attenuation level is generated. The correction direction of the attenuation level and the braking force ratio of the front and rear axles is matched, and the correction range is limited according to the current response speed of the brake master cylinder pressure and the current drop speed of the brake disc temperature to obtain the braking force ratio correction amount. The system correlates the braking force ratio correction, hydraulic pressure build-up rate correction, and the vehicle's current load status to perform consistency verification on the corrected front and rear axle braking force ratio and hydraulic pressure build-up rate, and then outputs a braking force distribution command.

10. A braking performance degradation prediction system based on cloud-based big data, characterized in that, The system is used to implement the braking performance degradation prediction method based on cloud big data as described in any one of claims 1-9, and the system includes: The cloud sequence module calculates the brake pedal travel, brake master cylinder pressure, four-wheel angular velocity, brake disc temperature and longitudinal deceleration uploaded by the vehicle brake controller. It divides the braking process according to the moment the pedal rises and falls, writes the braking process into the braking event sequence built in the cloud, and generates pressure change curve, wheel speed change curve and temperature drop curve. The trajectory generation module calculates the pressure build-up slope based on the pressure change curve, the wheel-end deceleration slope based on the wheel speed change curve, and the temperature rise peak drop interval based on the temperature drop curve. It compares the slope difference and interval difference of adjacent braking processes of the same vehicle to generate a braking energy decay trajectory. The boundary construction module calculates the temperature rise lag value and braking distance drift value based on the braking energy decay trajectory, maps the temperature rise lag value and braking distance drift value to the same coordinate plane, and constructs a dynamic judgment boundary according to the median offset and discrete width of the coordinate distribution within the continuous driving mileage. The instruction generation module inputs the braking energy attenuation trajectory into the dynamic judgment boundary for judgment, filters out coordinate points that appear continuously outside the dynamic judgment boundary, generates attenuation level, adjusts the front and rear axle braking force ratio and hydraulic pressure build-up rate according to the attenuation level, and generates braking force distribution instruction.

Images