Data processing method and system for tire pressure sensor

Abstract

This application relates to a tire pressure sensor data processing method and system, comprising the following steps: S1. A relay receiver predicts the tire pressure information for each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data using a pre-set tire pressure prediction model; S2. The relay receiver transmits the predicted tire pressure information for each wheel in the next cycle to the tire pressure sensor of each wheel in a time-sharing manner; S3. The tire pressure sensor of each wheel collects tire pressure data in real time and calculates the deviation ΔP between the measured value and the predicted value; S4. A data transmission mode is dynamically selected based on a threshold range of the deviation ΔP to feed data back to the relay receiver; S5. The relay receiver periodically iteratively optimizes the tire pressure prediction model based on the feedback data. This application effectively reduces data transmission volume and improves data transmission response efficiency.

Description

Technical Field

[0001] The present application relates to the field of tire pressure data processing, and in particular to a data processing method and system for a tire pressure sensor. Background Art

[0002] With the rapid development of automotive electronics, tire pressure monitoring systems (TPMS) have become a standard safety feature in modern vehicles. However, existing technologies still have the following prominent problems in practical applications:

[0003] 1. Electromagnetic interference: The traditional 315 / 433MHz transmission frequency band is susceptible to electromagnetic interference from vehicle-mounted equipment such as the ignition system (generating transient interference with a peak value of 200V / m) and the electric power steering (generating 20-100kHz harmonic interference). Actual measurement data shows that at the moment of engine startup, the bit error rate of the traditional system can reach 10 -2 Magnitude.

[0004] 2. Transmission conflict: When relay devices transmit and receive signals in the same frequency band, they can cause co-frequency interference. Experiments have shown that when the transmit power exceeds 10dBm, the receiving sensitivity decreases by over 30%.

[0005] 3. Traditional tire pressure monitoring systems (TPMS) need to periodically report all data, resulting in repeated transmission of more than 80% of normal status data.

[0006] Chinese patent publication number CN115091901B discloses a tire pressure information transmission method, system, and device. The method comprises the following steps: wirelessly monitoring tire pressure signals emitted by any one of multiple tire pressure sensors, and upon receiving the first tire pressure signal, entering a batch processing state; continuing to monitor and receive tire pressure signals emitted by the multiple tire pressure sensors within the batch processing state; upon obtaining tire pressure signals from all of the multiple tire pressure sensors, or upon termination of the batch processing state, repackaging the tire pressure information from the received tire pressure signals into comprehensive tire pressure information; and transmitting the comprehensive tire pressure information via a data connection line to a tire pressure host computer for displaying the comprehensive tire pressure information. This method avoids interference with wireless signals caused by wired communication by using only time-segmented, non-overlapping transmission. However, periodic full reporting results in redundant transmission of normal data, occupying communication resources and causing significant response delays in emergency scenarios.

[0007] With respect to the above-mentioned related technologies, the inventors believe that redundant data transmission in existing tire pressure monitoring systems leads to low data processing feedback efficiency and high abnormal response delay. Summary of the Invention

[0008] In order to solve the above problems, the present application provides a data processing method and system for a tire pressure sensor.

[0009] In a first aspect, the present application provides a data processing method for a tire pressure sensor, which adopts the following technical solution:

[0010] A data processing method for a tire pressure sensor comprises the following steps:

[0011] S1. The relay receiver uses a preset tire pressure prediction model to predict the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The historical tire pressure data is the tire pressure data of the vehicle for the past ten cycles. The tire pressure prediction model is an LSTM neural network model iteratively trained using historical cycle data.

[0012] S2. The relay receiver transmits the predicted tire pressure information of each wheel in the next cycle to the tire pressure sensor of each wheel in a time-sharing manner;

[0013] S3, the tire pressure sensor of each wheel collects tire pressure data in real time and calculates the deviation ΔP between the measured value and the predicted value;

[0014] S4. Dynamically select a data transmission mode according to a threshold range of the deviation ΔP to feed back data to the relay receiver, wherein the data transmission mode includes a heartbeat transmission mode, a polling data transmission mode, and an emergency direct connection transmission mode;

[0015] S5. The relay receiver periodically iteratively optimizes the tire pressure prediction model based on the feedback data.

[0016] Preferably, the dynamically selecting the data transmission mode according to the threshold range of the deviation ΔP to feed back data to the relay receiver specifically includes the following steps:

[0017] Determining whether the deviation ΔP is less than a preset first deviation threshold;

[0018] If it is less than, the data transmission mode selects the heartbeat transmission mode, and only feeds back a 1-byte heartbeat signal to the relay receiver;

[0019] If not, determine whether the deviation ΔP is less than a preset second deviation threshold;

[0020] If it is less than, the data transmission mode selects the polling data transmission mode, and the tire pressure data is fed back to the relay receiver through the time-sharing polling channel;

[0021] If it is not less than, the data transmission mode selects the emergency direct transmission mode, triggering the emergency channel to directly transmit the feedback tire pressure data with the relay receiver and generate an early warning.

[0022] Preferably, the first deviation threshold is 5% of the nominal tire pressure value of the vehicle, and the second deviation threshold is 15% of the nominal tire pressure value of the vehicle.

[0023] Preferably, the time-sharing polling channel uses a dynamic time slot allocation algorithm to dynamically adjust the time slot weight of each tire pressure sensor. The time slot weight is calculated and dynamically adjusted according to the sensor's historical abnormal frequency and current deviation ΔP using a time slot weight calculation formula preset by the dynamic time slot allocation algorithm. The time slot weight calculation formula is specifically:

[0024] ;

[0025] in, is the time slot weight of the i-th tire pressure sensor, is the current tire pressure deviation of the i-th tire pressure sensor, is the maximum tire pressure deviation of all tire pressure sensors, is the historical abnormal frequency of the i-th tire pressure sensor, α is the deviation weight coefficient, and β is the historical abnormal weight coefficient, both of which are preset by the management personnel, and α+β=1.

[0026] Preferably, the emergency channel uses frequency hopping spread spectrum technology, has a transmission delay of less than 50ms, and has the highest priority weight.

[0027] Preferably, the tire pressure prediction model predicts the tire pressure information of each wheel in the next cycle based on the wheel speed, temperature and historical tire pressure data, further comprising: activating a preset temperature compensation module after the prediction is completed, calculating a compensation amount using a preset compensation amount calculation formula to correct the predicted tire pressure information to obtain the tire pressure information of each wheel in the cycle;

[0028] The compensation calculation formula is specifically:

[0029] P 补偿 =k*(T 当前 -T 基准 )+b;

[0030] Among them, P 补偿 is the tire pressure compensation caused by temperature change, k is the tire material temperature coefficient, which is calibrated by the management personnel according to the tire specifications and models, T 当前 is the tire temperature collected in real time during prediction, T 基准 is the standard temperature corresponding to the nominal tire pressure value, which is calibrated by the management personnel according to the tire specifications and models. b is a constant term set by the management personnel during regular vehicle inspections to correct tire pressure deviations caused by non-temperature factors.

[0031] Preferably, the relay receiver periodically iteratively optimizes the tire pressure prediction model according to the feedback data, specifically comprising the following steps:

[0032] The relay receiver verifies the vehicle status during a preset time period to determine whether the vehicle is in a driving state;

[0033] If yes, a delayed task is generated. After determining that the user has turned off the engine and locked the car, the tire pressure prediction model is optimized offline based on the feedback data.

[0034] If not, the tire pressure prediction model is optimized offline iteratively based on the feedback data.

[0035] The relay receiver determines whether the abnormal data trigger conditions are met based on the feedback data. The abnormal data trigger conditions are that a tire pressure deviation ΔP of ≥ 10% exists in a single tire for three consecutive cycles during a single drive, and the ambient temperature change rate is ≥ 2°C / minute.

[0036] If not satisfied, the optimization is completed;

[0037] If the condition is satisfied, the corrected tire material temperature coefficient k is calculated based on the abnormal data using the preset temperature compensation coefficient correction calculation formula. 新 The temperature compensation coefficient correction calculation formula is specifically as follows: ;

[0038] Where 1≥θ≥0.7, θ is the smoothing coefficient set by the management personnel, k 旧 is the current tire material temperature coefficient, n is the total number of abnormal data collection times in 3 cycles, is the measured tire pressure value of the i-th abnormal data, is the tire pressure prediction value of the i-th abnormal data, The tire temperature collected in real time when the tire pressure prediction value of the i-th abnormal data is predicted;

[0039] The corrected tire material temperature coefficient k is introduced into each abnormal data. 新 Producing the corrected deviation ΔP 修正 , compare the corrected deviation ΔP in each abnormal data 修正 Whether the ratio of the deviation ΔP that is less than the original deviation exceeds 50%;

[0040] If it is less than, the corrected tire material temperature coefficient k 新 Replace the current tire material temperature coefficient k 旧 ;

[0041] If it is not less than, a cause diagnosis request is generated and sent to the manufacturer's experts, and artificial rules are injected into the temperature compensation coefficient according to the expert's feedback plan.

[0042] Preferably, the tire pressure prediction model predicts the tire pressure information of each wheel in the next cycle based on the wheel speed, temperature, and historical tire pressure data, further comprising the tire pressure prediction model acquiring historical tire pressure data persistently stored in a preset memory, and if the number of historical cycles of the historical tire pressure data is less than 10 cycles, completing the input sequence using a dynamic filling algorithm, specifically comprising the following steps:

[0043] Read N historical cycle data stored persistently in the memory, each historical cycle contains M tire pressure sampling values, and extract the characteristic parameters of each historical cycle. The characteristic parameters of the historical cycle data include the period sampling mean. , Cycle Trend Slope and cyclical volatility ;

[0044] Assign weights to each historical period through the weighted weight calculation formula and calculate the characteristic parameter prediction value of the period to be filled. The characteristic parameter prediction value includes the period sampling prediction mean , Cycle forecast trend slope and periodic forecast volatility ; The weighted weight calculation formula is:

[0045] ; where j is the historical period number in partial order according to the time axis, j=1, 2, 3...N; is the weight coefficient of the jth historical period;

[0046] ; ;in It is a preset maximum fluctuation threshold, set by management personnel;

[0047] Based on the predicted values ​​of the characteristic parameters of the period to be filled, the initial value is calculated and set for each historical period to be filled. , and generate M sample values ; and the sampling value , where X 标称 is the nominal tire pressure of the vehicle; the calculation formula for the historical cycle initial value to be filled is specifically:

[0048] ;in is the initial value of the lth historical period to be filled, l is the serial number of the historical period to be filled, l=N+1, N+2...10; the sampling value generation formula is:

[0049] ;in is the tth sampling value of the lth historical period that needs to be filled, is a random number in the range of [-1,1], used to simulate the fluctuation of the t-th sampling value;

[0050] The original N historical cycles and the padded 10-N cycles are concatenated in chronological order, and a complete input sequence of 10 cycles × M samples is output to the tire pressure prediction model.

[0051] Preferably, the memory is a vehicle-mounted non-volatile memory, which cyclically overwrites and saves the tire pressure, temperature and timestamp data of the last 10 cycles, performs data timeliness verification on the saved historical tire pressure data, and automatically deletes historical tire pressure data whose timestamp interval exceeds a preset time threshold. The time threshold is automatically updated according to the vehicle location information.

[0052] In a second aspect, the present application provides a data processing system for a tire pressure sensor, which adopts the following technical solution:

[0053] A tire pressure sensor data processing system includes a vehicle computer, a relay receiver connected to the vehicle computer by wire, and multiple tire pressure sensors installed in tires and wirelessly connected to the relay receiver. The relay receiver is equipped with a tire pressure prediction model that predicts the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The historical tire pressure data is the tire pressure data of the vehicle for the past ten cycles. The tire pressure prediction model is an LSTM neural network model obtained by iterative training using historical cycle data.

[0054] A microcontroller is arranged at the tire pressure sensor to calculate the deviation ΔP between the measured value and the predicted value, and dynamically selects the data transmission mode according to the threshold range of the deviation ΔP to feed back data to the relay receiver.

[0055] In summary, this application includes at least one of the following beneficial technical effects:

[0056] 1. Build a "prediction-driven, on-demand feedback" intelligent monitoring system. A tire pressure prediction model is installed in the relay receiver to process sequential data to predict the tire pressure information for the next cycle. This information is then sent to the tire pressure sensor on each wheel in a time-sharing manner. The data transmission mode is then dynamically selected based on the deviation ΔP between the measured and predicted values. This reduces unnecessary data transmission while ensuring data accuracy, improves response speed to abnormal data, and ensures that idle channels are used for data transmission when abnormal data is detected. This also significantly extends sensor lifespan.

[0057] 2. Build a triple-threshold dynamic time-sharing channel, dividing it into heartbeat, polling, and emergency modes based on deviation thresholds. This manages communication resources in a hierarchical manner, optimizes channel utilization, prioritizes resources in emergency scenarios, and ensures real-time transmission of critical data, effectively reducing data transmission volume and improving data transmission response efficiency.

[0058] 3. After the prediction is complete, the pre-set temperature compensation module is activated to compensate for the impact of temperature changes on tire pressure, effectively reducing false triggers caused by temperature fluctuations. In particular, it significantly reduces prediction errors in low-temperature scenarios, achieves precise correction of tire pressure data, improves the robustness of the tire pressure monitoring system, and significantly improves tire pressure prediction accuracy, helping to reduce data transmission volume and improve data transmission response efficiency.

[0059] 4. A two-way optimization mechanism is established. First, the tire pressure prediction model is regularly iterated offline based on feedback data to continuously improve its performance. Second, the temperature compensation coefficient is adjusted in real time based on abnormal data fed back by the sensor end and when the abnormal data trigger conditions are met. A sliding average method is used to balance historical experience with current anomalies, avoid overfitting, and improve the accuracy of the temperature compensation coefficient correction. This can effectively improve the compensation accuracy of the temperature compensation module for complex scenarios such as sudden low-temperature changes and slow tire leaks, thereby achieving accurate predictions. Finally, when the deviation improvement ratio after correction is less than 50%, a manual diagnosis request is triggered to prevent coefficient distortion caused by algorithm misjudgment (such as the nonlinear effect of rubber aging). Industry experience is converted into a rule base, improving the robustness of the model under extreme conditions and further improving the compensation accuracy of the temperature compensation module. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a method flow chart of a data processing method for a tire pressure sensor in an embodiment of the present application;

[0061] Figure 2 This is a flow chart of a method for dynamically selecting a data transmission mode according to a threshold range of a deviation ΔP in an embodiment of the present application;

[0062] Figure 3 is a flow chart of a method for iteratively optimizing a tire pressure prediction model in an embodiment of the present application;

[0063] Figure 4 This is a flow chart of a method for completing an input sequence using a dynamic filling algorithm in an embodiment of the present application;

[0064] Figure 5 This is a system block diagram of a data processing system for a tire pressure sensor in an embodiment of the present application.

[0065] Explanation of the accompanying reference numerals: 1. Vehicle computer; 2. Relay receiver; 21. Tire pressure prediction model; 3. Tire pressure sensor; 31. Microcontroller. DETAILED DESCRIPTION

[0066] The following is combined with Figure 1-Figure 5 This application is described in further detail.

[0067] The embodiment of the present application discloses a data processing method for a tire pressure sensor. Figure 1 , a data processing method for a tire pressure sensor, comprising the following steps:

[0068] S1. Predicting Tire Pressure Information: The relay receiver uses a preset tire pressure prediction model to predict the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The historical tire pressure data refers to the tire pressure data of the vehicle over the past ten cycles. The tire pressure prediction model is an LSTM neural network model that is iteratively trained using historical cycle data. In this embodiment, the cycle is set to 5 minutes.

[0069] S2, time-sharing transmission: The relay receiver transmits the predicted tire pressure information of each wheel in the next cycle to the tire pressure sensor of each wheel in a time-sharing manner;

[0070] S3. Calculating the deviation between the measured value and the predicted value: The tire pressure sensor of each wheel collects tire pressure data in real time, and calculates the deviation ΔP between the measured value and the predicted value;

[0071] S4. Dynamically select a data transmission mode: Dynamically select a data transmission mode according to a threshold range of the deviation ΔP to feed back data to the relay receiver, wherein the data transmission mode includes a heartbeat transmission mode, a polling data transmission mode, and an emergency direct connection transmission mode;

[0072] S5. Iterative Optimization: The relay receiver periodically iteratively optimizes the tire pressure prediction model based on feedback data. This step builds a "prediction-driven, on-demand feedback" intelligent monitoring system. The tire pressure prediction model in the relay receiver processes sequential data to predict the tire pressure information for the next cycle, which is then transmitted to each wheel's tire pressure sensor in a time-sharing manner. The data transmission mode is then dynamically selected based on the deviation ΔP between the measured and predicted values. This reduces unnecessary data transmission while ensuring data accuracy, improves response to abnormal data, and ensures immediate access to idle channels for data transmission when abnormal data is detected. This also significantly increases sensor lifespan.

[0073] Reference Figure 2 The method of dynamically selecting a data transmission mode according to a threshold range of the deviation ΔP to feed back data to the relay receiver specifically includes the following steps:

[0074] A1. Determine whether the deviation ΔP is less than a preset first deviation threshold;

[0075] A2. Select heartbeat transmission mode: If less than, the data transmission mode is heartbeat transmission mode, which only feeds a 1-byte heartbeat signal to the relay receiver. The heartbeat signal is a 1-byte (8-bit) checksum, without actual tire pressure data, and contains only the sensor ID and status flag. It is used to maintain the communication connection and status synchronization between the sensor and the relay receiver, preventing the relay from misjudging the sensor as offline due to long periods of silence. The heartbeat interval is used to calibrate the clock reference of the sensor and the relay to ensure the timing alignment of time-sharing transmission.

[0076] A3. If not, determine whether the deviation ΔP is less than a preset second deviation threshold;

[0077] A4. Select polling data transmission mode: If it is less than, the data transmission mode selects polling data transmission mode, and transmits tire pressure data to the relay receiver through the time-sharing polling channel;

[0078] A5. Select Emergency Direct Transmission Mode: If it is not less than, the data transmission mode is selected as Emergency Direct Transmission Mode, triggering the emergency channel to directly transmit tire pressure data to the relay receiver and generate an alert. Through the above steps, a triple-threshold dynamic time-sharing channel is established. Based on the deviation threshold, heartbeat, polling, and emergency modes are divided into hierarchical management of communication resources, optimizing channel utilization. Emergency scenarios prioritize resources and ensure real-time transmission of critical data, effectively reducing data transmission volume and improving data transmission response.

[0079] The first deviation threshold is 5% of the vehicle's nominal tire pressure, and the second deviation threshold is 15% of the vehicle's nominal tire pressure. Using nominal values ​​as a benchmark, the risk of false positives and missed positives is balanced, significantly reducing communication traffic during normal operating conditions and effectively improving anomaly detection accuracy.

[0080] The time-sharing polling channel uses a dynamic time slot allocation algorithm to dynamically adjust the time slot weight of each tire pressure sensor. The time slot weight is calculated and dynamically adjusted based on the sensor's historical abnormal frequency and current deviation ΔP using a time slot weight calculation formula preset by the dynamic time slot allocation algorithm. The time slot weight calculation formula is specifically:

[0081] ;

[0082] in, is the time slot weight of the i-th tire pressure sensor, is the current tire pressure deviation of the i-th tire pressure sensor, is the maximum tire pressure deviation of all tire pressure sensors, is the historical abnormal frequency of the i-th tire pressure sensor, α is the deviation weight coefficient, and β is the historical abnormal weight coefficient, all preset by management, with α + β = 1. A dynamic time slot allocation algorithm dynamically adjusts the time slot weight based on the sensor's historical abnormal frequency and current deviation ΔP. This ensures that more severe abnormal data is transmitted promptly, effectively reducing channel collision rates. This significantly improves sensor response speed for high-risk tires (aging and frequently experiencing problems), thereby ensuring driver safety.

[0083] The emergency channel utilizes frequency hopping spread spectrum (FHSS) technology, with a transmission delay of less than 50ms and the highest priority. This technology offers strong anti-interference capabilities and minimal transmission delay, enabling rapid data transmission and early warning when tire pressure is abnormal.

[0084] The tire pressure prediction model predicts the tire pressure information of each wheel in the next cycle based on the wheel speed, temperature and historical tire pressure data, further comprising: activating a preset temperature compensation module after the prediction is completed, and calculating a compensation amount using a preset compensation amount calculation formula to correct the predicted tire pressure information to obtain the tire pressure information of each wheel in the cycle;

[0085] The compensation calculation formula is specifically:

[0086] P 补偿 =k*(T 当前 -T 基准 )+b;

[0087] Among them, P 补偿 is the tire pressure compensation caused by temperature change, k is the tire material temperature coefficient, which is calibrated by the management personnel according to the tire specifications and models, T 当前 is the tire temperature collected in real time during prediction, T 基准 The standard temperature corresponding to the nominal tire pressure value is calibrated by management based on tire specifications and models. b is a constant term set by management during regular vehicle inspections to correct for tire pressure deviations caused by factors other than temperature. After the prediction is complete, the pre-set temperature compensation module is activated to compensate for the impact of temperature changes on tire pressure, effectively reducing false triggering caused by temperature fluctuations. This significantly reduces prediction errors in low-temperature scenarios, enabling precise correction of tire pressure data, enhancing the robustness of the tire pressure monitoring system, and significantly improving tire pressure prediction accuracy, helping to reduce data transmission volume and improve data transmission response efficiency.

[0088] Reference Figure 3 The relay receiver periodically iteratively optimizes the tire pressure prediction model according to the feedback data, specifically including the following steps:

[0089] B1. The relay receiver verifies the vehicle status during a preset time period to determine whether the vehicle is in a driving state;

[0090] B2. Delayed offline iterative optimization: If in, a delayed task is generated. After determining that the user has turned off the engine and locked the vehicle, the tire pressure prediction model is regularly iterated and optimized offline based on the feedback data.

[0091] B3. Direct offline iterative optimization: If not, the tire pressure prediction model is directly optimized offline iteratively based on the feedback data;

[0092] B4. The relay receiver determines whether the abnormal data trigger condition is met based on the feedback data. The abnormal data trigger condition is that a tire pressure deviation ΔP of ≥ 10% exists in a single tire for three consecutive cycles during a single drive, and the ambient temperature change rate is ≥ 2°C / minute.

[0093] B5. If not satisfied, then complete the optimization;

[0094] B6. Tire material temperature coefficient correction: If the condition is met, the corrected tire material temperature coefficient k is calculated based on the abnormal data using the preset temperature compensation coefficient correction calculation formula. 新 The temperature compensation coefficient correction calculation formula is specifically as follows:

[0095] ;

[0096] Where 1≥θ≥0.7, θ is the smoothing coefficient set by the management personnel, k 旧 is the current tire material temperature coefficient, n is the total number of abnormal data collection times in 3 cycles, is the measured tire pressure value of the i-th abnormal data, is the tire pressure prediction value of the i-th abnormal data, The tire temperature collected in real time when the tire pressure prediction value of the i-th abnormal data is predicted;

[0097] B7. Incorporate the corrected tire material temperature coefficient k into each abnormal data. 新 Producing the corrected deviation ΔP 修正 , compare the corrected deviation ΔP in each abnormal data 修正 Whether the ratio of the deviation ΔP that is less than the original deviation exceeds 50%;

[0098] B8, tire material temperature coefficient update: if it is less than, then the corrected tire material temperature coefficient k 新 Replace the current tire material temperature coefficient k 旧 ;

[0099] B9. Experts inject manual rules: If it is not less than , a cause diagnosis request is generated and sent to the manufacturer's experts. Manual rules are then injected into the temperature compensation coefficient based on the expert's feedback plan. Through the above steps, a two-way optimization mechanism is established. First, the tire pressure prediction model is regularly iterated and optimized offline based on feedback data to continuously improve the model performance of the tire pressure prediction model. Secondly, the temperature compensation coefficient is adjusted in real time based on abnormal data fed back by the sensor end and when the abnormal data trigger conditions are met. The sliding average method balances historical experience with current anomalies to avoid overfitting and improve the accuracy of temperature compensation coefficient correction. For complex scenarios such as sudden low temperature changes and slow tire leaks, it can effectively improve the compensation accuracy of the temperature compensation module and thus achieve accurate predictions. Finally, when the deviation improvement ratio after correction is less than 50%, a manual diagnosis request is triggered to prevent coefficient distortion caused by algorithm misjudgment (such as the nonlinear effect of rubber aging). Industry experience is converted into a rule base, improving the robustness of the model under extreme working conditions and further improving the compensation accuracy of the temperature compensation module.

[0100] Reference Figure 4 The tire pressure prediction model predicts the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The tire pressure prediction model also obtains historical tire pressure data persistently stored in a preset memory. If the number of historical tire pressure data cycles is less than 10 cycles, the input sequence is completed using a dynamic filling algorithm. Specifically, the steps include:

[0101] C1. Extract characteristic parameters of each historical period: read N historical period data stored persistently in the memory, each historical period contains M tire pressure sampling values, and extract characteristic parameters of each historical period. The characteristic parameters of the historical period data include the period sampling mean. , Cycle Trend Slope and cyclical volatility ;

[0102] C2. Calculate the predicted value of the characteristic parameters of the period to be filled: assign weights to each historical period through the weighted weight calculation formula, and calculate the predicted value of the characteristic parameters of the period to be filled. The predicted value of the characteristic parameters includes the period sampling prediction mean , Cycle forecast trend slope and periodic forecast volatility ; The weighted weight calculation formula is:

[0103] ; where j is the historical period number in partial order according to the time axis, j=1, 2, 3...N; is the weight coefficient of the jth historical period;

[0104] ; ; ;in It is a preset maximum fluctuation threshold, set by management personnel;

[0105] C3. Generate filling cycle data: Based on the predicted values ​​of the characteristic parameters of the cycle to be filled, calculate and set the initial value for each historical cycle to be filled , and generate M sample values ; and the sampling value , where X 标称 is the nominal tire pressure of the vehicle; the calculation formula for the historical cycle initial value to be filled is specifically: ;in is the initial value of the lth historical period to be filled, l is the serial number of the historical period to be filled, l=N+1, N+2...10; the sampling value generation formula is:

[0106] ;in is the tth sampling value of the lth historical period that needs to be filled, is a random number in the range of [-1,1], used to simulate the fluctuation of the t-th sampling value;

[0107] C4. Construct a complete input sequence: Concatenate the original N historical cycles with the 10-N cycles generated by padding in chronological order, and output a complete input sequence of 10 cycles × M samples to the tire pressure prediction model. Through the above steps, when the number of historical cycles of the vehicle's historical tire pressure data is less than 10 cycles, a dynamic padding algorithm is used to extract the periodic characteristics (mean, slope, volatility) of the historical data, combine them with weighted weights to generate supplementary data, and inject random noise during the data generation process. , simulating real tire pressure fluctuations, greatly reducing the error between filled data and measured data, greatly improving the authenticity of filled data, solving the core pain points of tire pressure monitoring systems in scenarios with insufficient data, and achieving dual breakthroughs in "zero-wait prediction" and "high-precision completion", providing reliable technical support for real scenarios such as short-distance driving and cross-day driving.

[0108] The memory is an onboard non-volatile memory that cyclically overwrites and stores tire pressure, temperature, and timestamp data for the last 10 cycles. The system also verifies the timeliness of the stored historical tire pressure data and automatically deletes any tire pressure data with a timestamp interval exceeding a preset time threshold, which is automatically updated based on the vehicle's location. Using a pre-set time threshold comparison table, the data retention time threshold is dynamically adjusted based on local road conditions, ensuring the validity of historical tire pressure data and improving tire pressure data prediction accuracy.

[0109] The embodiment of the present application also discloses a data processing system for a tire pressure sensor. Figure 5 A tire pressure sensor data processing system includes a vehicle computer, a relay receiver connected to the vehicle computer by wire, and multiple tire pressure sensors installed in tires and wirelessly connected to the relay receiver; the relay receiver is equipped with a tire pressure prediction model, which predicts the tire pressure information of each wheel in the next cycle based on wheel speed, temperature and historical tire pressure data; the historical tire pressure data is the tire pressure data of the vehicle for the past ten cycles; the tire pressure prediction model is an LSTM neural network model obtained by iterative training through historical cycle data.

[0110] A microcontroller located at the tire pressure sensor calculates the deviation ΔP between the measured and predicted values. Based on the threshold range of the ΔP deviation, it dynamically selects a data transmission mode to feed data back to the relay receiver. The microcontroller also includes a built-in lightweight residual calculation module, occupying less than 2KB of storage space. This module supports real-time deviation calculation, reducing hardware costs and computational burden.

[0111] Build an intelligent monitoring system with "prediction-driven, on-demand feedback". Set up a tire pressure prediction model in the relay receiver to process serial data to predict the tire pressure information of the next cycle, and send it to the tire pressure sensor of each wheel in a time-sharing manner. Then, dynamically select the data transmission mode based on the deviation ΔP between the measured value and the predicted value. While ensuring data accuracy, it can reduce unnecessary data transmission and improve the response speed of abnormal data. When abnormal data exists, there is an idle channel to transmit data as soon as possible. At the same time, it can also greatly extend the service life of the sensor.

[0112] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the invention. Obviously, the embodiments described are only some embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in this field can still combine, add, delete or make other adjustments to the features in the various embodiments of the present invention according to the circumstances without conflict, without making creative work, so as to obtain different other technical solutions that do not deviate from the concept of the present invention in essence, and these technical solutions also fall within the scope of protection of the present invention.

Claims

1. A data processing method for a tire pressure sensor, characterized in that: The following steps are involved: S1. The relay receiver uses a preset tire pressure prediction model to predict the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The historical tire pressure data is the tire pressure data of the vehicle for the past ten cycles. The tire pressure prediction model is an LSTM neural network model iteratively trained using historical cycle data. S2. The relay receiver transmits the predicted tire pressure information of each wheel in the next cycle to the tire pressure sensor of each wheel in a time-sharing manner; S3, the tire pressure sensor of each wheel collects tire pressure data in real time and calculates the deviation ΔP between the measured value and the predicted value; S4. Dynamically select a data transmission mode according to a threshold range of the deviation ΔP to feed back data to the relay receiver, wherein the data transmission mode includes a heartbeat transmission mode, a polling data transmission mode, and an emergency direct connection transmission mode; S5. The relay receiver periodically iteratively optimizes the tire pressure prediction model based on the feedback data; The method of dynamically selecting a data transmission mode according to the threshold range of the deviation ΔP to feed back data to the relay receiver specifically includes the following steps: Determining whether the deviation ΔP is less than a preset first deviation threshold; If it is less than, the data transmission mode selects the heartbeat transmission mode, and only feeds back a 1-byte heartbeat signal to the relay receiver; If not, determine whether the deviation ΔP is less than a preset second deviation threshold; If it is less than, the data transmission mode selects the polling data transmission mode, and the tire pressure data is fed back to the relay receiver through the time-sharing polling channel; If it is not less than, the data transmission mode selects the emergency direct transmission mode, triggering the emergency channel to directly transmit the feedback tire pressure data to the relay receiver and generate an early warning; The time-sharing polling channel uses a dynamic time slot allocation algorithm to dynamically adjust the time slot weight of each tire pressure sensor. The time slot weight is calculated and dynamically adjusted based on the sensor's historical abnormal frequency and current deviation ΔP using a time slot weight calculation formula preset by the dynamic time slot allocation algorithm. The time slot weight calculation formula is specifically: ； in, is the time slot weight of the i-th tire pressure sensor, is the current tire pressure deviation of the i-th tire pressure sensor, is the maximum tire pressure deviation of all tire pressure sensors, is the historical abnormal frequency of the i-th tire pressure sensor, α is the deviation weight coefficient, and β is the historical abnormal weight coefficient, both of which are preset by the management personnel, and α+β=1.

2. The tire pressure sensor data processing method according to claim 1, characterized in that: The first deviation threshold is 5% of the nominal tire pressure value of the vehicle, and the second deviation threshold is 15% of the nominal tire pressure value of the vehicle.

3. The tire pressure sensor data processing method according to claim 1, characterized in that: The emergency channel adopts frequency hopping spread spectrum technology, has a transmission delay of less than 50ms, and has the highest priority weight.

4. The tire pressure sensor data processing method according to claim 1, characterized in that: The tire pressure prediction model predicts tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data, further comprising: activating a preset temperature compensation module after the prediction is completed, calculating a compensation amount using a preset compensation amount calculation formula to correct the predicted tire pressure information to obtain tire pressure information of each wheel in the cycle; The compensation calculation formula is specifically: P 补偿 =k*(T 当前 -T 基准 )+b; Among them, P 补偿 is the tire pressure compensation caused by temperature change, k is the tire material temperature coefficient, which is calibrated by the management personnel according to the tire specifications and models, T 当前 is the tire temperature collected in real time during prediction, T 基准 is the standard temperature corresponding to the nominal tire pressure value, which is calibrated by the management personnel according to the tire specifications and models. b is a constant term set by the management personnel during regular vehicle inspections to correct tire pressure deviations caused by non-temperature factors.

5. The data processing method of a tire pressure sensor according to claim 4, characterized in that: The relay receiver periodically iteratively optimizes the tire pressure prediction model according to the feedback data, specifically comprising the following steps: The relay receiver verifies the vehicle status during a preset time period to determine whether the vehicle is in a driving state; If yes, a delayed task is generated. After determining that the user has turned off the engine and locked the car, the tire pressure prediction model is optimized offline based on the feedback data. If not, the tire pressure prediction model is optimized offline iteratively based on the feedback data. The relay receiver determines whether the abnormal data trigger conditions are met based on the feedback data. The abnormal data trigger conditions are that a tire pressure deviation ΔP of ≥ 10% exists in a single tire for three consecutive cycles during a single drive, and the ambient temperature change rate is ≥ 2°C / minute. If not satisfied, the optimization is completed; If the condition is satisfied, the corrected tire material temperature coefficient k is calculated based on the abnormal data using the preset temperature compensation coefficient correction calculation formula. 新 The temperature compensation coefficient correction calculation formula is specifically as follows: ； Where 1≥θ≥0.7, θ is the smoothing coefficient set by the management personnel, k 旧 is the current tire material temperature coefficient, n is the total number of abnormal data collection times in 3 cycles, is the measured tire pressure value of the i-th abnormal data, is the tire pressure prediction value of the i-th abnormal data, The tire temperature collected in real time when the tire pressure prediction value of the i-th abnormal data is predicted; The corrected tire material temperature coefficient k is introduced into each abnormal data. 新 Producing the corrected deviation ΔP 修正 , compare the corrected deviation ΔP in each abnormal data 修正 Whether the ratio of the deviation ΔP that is less than the original deviation exceeds 50%; If it is less than, the corrected tire material temperature coefficient k 新 Replace the current tire material temperature coefficient k 旧 ; If it is not less than, a cause diagnosis request is generated and sent to the manufacturer's experts, and artificial rules are injected into the temperature compensation coefficient according to the expert's feedback plan.

6. The tire pressure sensor data processing method according to claim 1, characterized in that: The tire pressure prediction model predicts the tire pressure information of each wheel in the next cycle based on wheel speed, temperature, and historical tire pressure data. The tire pressure prediction model also obtains historical tire pressure data persistently stored in a preset memory. If the number of historical tire pressure data cycles is less than 10 cycles, the input sequence is completed using a dynamic filling algorithm. Specifically, the steps include: Read N historical cycle data stored persistently in the memory, each historical cycle contains M tire pressure sampling values, and extract the characteristic parameters of each historical cycle. The characteristic parameters of the historical cycle data include the period sampling mean. , Cycle Trend Slope and cyclical volatility ; Assign weights to each historical period through the weighted weight calculation formula and calculate the characteristic parameter prediction value of the period to be filled. The characteristic parameter prediction value includes the period sampling prediction mean , Cycle forecast trend slope and periodic forecast volatility ; The weighted weight calculation formula is: ; where j is the historical period number in partial order according to the time axis, j=1, 2, 3...N; is the weight coefficient of the jth historical period; ; ; ;in It is a preset maximum fluctuation threshold, set by management personnel; Based on the predicted values ​​of the characteristic parameters of the period to be filled, the initial value is calculated and set for each historical period to be filled. , and generate M sample values ; and the sampling value , where X 标称 is the nominal tire pressure of the vehicle; the calculation formula for the historical cycle initial value to be filled is specifically: ;in is the initial value of the lth historical period to be filled, l is the serial number of the historical period to be filled, l=N+1, N+2...10; the sampling value generation formula is: ;in is the tth sampling value of the lth historical period that needs to be filled, is a random number in the range of [-1,1], used to simulate the fluctuation of the t-th sampling value; The original N historical cycles and the padded 10-N cycles are concatenated in chronological order, and a complete input sequence of 10 cycles × M samples is output to the tire pressure prediction model.

7. The tire pressure sensor data processing method according to claim 6, characterized in that: The memory is an on-board non-volatile memory that cyclically overwrites and stores tire pressure, temperature, and timestamp data for the last 10 cycles. It also performs a data timeliness check on the stored historical tire pressure data and automatically deletes historical tire pressure data whose timestamp interval exceeds a preset time threshold. The time threshold is automatically updated based on the vehicle's location information.

8. A data processing system for a tire pressure sensor, applied to a data processing method for a tire pressure sensor according to any one of claims 1 to 7, characterized in that: The invention comprises a vehicle computer (1), a relay receiver (2) connected to the vehicle computer (1) by wire, and a plurality of tire pressure sensors (3) respectively installed in tires and wirelessly connected to the relay receiver (2); the relay receiver (2) is provided with a tire pressure prediction model (21) for predicting tire pressure information of each wheel in the next cycle based on wheel speed, temperature and historical tire pressure data; the historical tire pressure data is tire pressure data of the vehicle in the past ten cycles; the tire pressure prediction model (21) is an LSTM neural network model obtained by iterative training through historical cycle data; A microcontroller (31) is arranged at the tire pressure sensor (3) for calculating a deviation ΔP between a measured value and a predicted value, and dynamically selecting a data transmission mode according to a threshold range of the deviation ΔP to feed back data to the relay receiver (2).

Images