Tire pressure real-time monitoring method and system based on wireless sensor

Abstract

The invention discloses a tire pressure real-time monitoring method and system based on a wireless sensor, and the method comprises the steps: 101, collecting tire pressure data, and carrying out the filtering processing of the tire pressure data; step 102, carrying out data decoding, verification and formatting on the tire pressure data subjected to the filtering processing; dynamically adjusting the frequency of data acquisition according to the driving state of the vehicle; step 103, performing cross validation and fusion on data acquired by different sensors, and taking the fused data as input of machine learning; and 104, historical tire pressure data are analyzed according to the fused tire pressure data through a machine learning algorithm, a tire pressure model is established, the tire pressure model comprises the relation that the tire pressure changes along with temperature, load and vehicle speed factors, and when the tire pressure data collected in real time deviates from a model prediction value and exceeds a preset threshold value, an alarm is sent to a driver in time. By the adoption of the technical scheme, the accuracy, reliability and real-time performance of tire pressure monitoring are improved.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and in particular relates to a tire pressure real-time monitoring method and system based on a wireless sensor. Background Art

[0002] In the real-time wireless tire pressure monitoring system, the accuracy and reliability of data are crucial. After the tire pressure data collected by the sensor is processed, it needs to be transmitted wirelessly to the receiving device of the vehicle. However, there are many challenges in the wireless transmission process. First, the working environment of the sensor inside the tire is harsh, and factors such as high temperature and vibration may cause data distortion or transmission interruption. Secondly, during the driving process of the vehicle, the distance between the tire and the receiving device is constantly changing, and the signal strength will also fluctuate accordingly, which is easy to cause data packet loss. Furthermore, there is a large amount of electromagnetic interference in the external environment, such as electronic equipment and radio stations of other vehicles, which may affect the wireless transmission of tire pressure data. Therefore, how to ensure the quality of wireless transmission of tire pressure data in a complex and changing environment is a technical problem that needs to be solved urgently. At the same time, the battery life of the sensor is also a limiting factor. Frequent data collection and transmission will accelerate the depletion of the battery, so it is necessary to find a balance between the reliability of data transmission and the energy consumption of the sensor. In short, there are still many problems to be overcome in the wireless transmission technology of the tire pressure monitoring system. Only by solving these problems can the safety of vehicle driving be ensured and drivers can go on the road with peace of mind. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a tire pressure real-time monitoring method and system based on wireless sensors.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] A tire pressure real-time monitoring method based on a wireless sensor, comprising:

[0006] Step 101, collecting tire pressure data and filtering the tire pressure data;

[0007] Step 102: decoding, verifying and formatting the tire pressure data after filtering; dynamically adjusting the frequency of data collection according to the driving state of the vehicle;

[0008] Step 103: cross-validate and fuse the data collected by different sensors, and use the fused data as input for machine learning;

[0009] Step 104: Analyze historical tire pressure data using a machine learning algorithm for the fused tire pressure data to establish a tire pressure model. The tire pressure model includes the relationship between tire pressure changes with temperature, load, and vehicle speed. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is promptly issued to the driver.

[0010] Preferably, in step S102, the driving state of the vehicle is determined by a vehicle speed sensor, and the sampling rate is reduced when the vehicle speed is lower than a preset threshold, and the sampling rate is increased when the vehicle speed is higher than the preset threshold.

[0011] Preferably, step 103 includes:

[0012] Determine the number and location of sensors to be installed in each tire;

[0013] The Kalman filter algorithm is used to fuse the tire pressure data collected by multiple sensors, and the true value of the tire pressure is dynamically estimated by establishing the state equation and observation equation;

[0014] The fused tire pressure data is organized in time series to construct a training data set. An appropriate time window size is selected to extract the statistical features of the data as input features of the machine learning model.

[0015] Preferably, step 104 includes:

[0016] Obtain historical tire pressure data of vehicle tires, including tire pressure value, temperature, load and vehicle speed related factor data, and build a tire pressure dataset;

[0017] Remove abnormal data and normalize the tire pressure data set;

[0018] Train the tire pressure model and establish the relationship mapping between tire pressure and factors such as temperature, load and vehicle speed;

[0019] Input the real-time collected tire pressure data into the trained tire pressure model to obtain the tire pressure value predicted by the model;

[0020] Calculate the deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds the preset threshold;

[0021] If the threshold is exceeded, the tire pressure is considered abnormal and the warning mechanism is triggered.

[0022] The present invention also provides a tire pressure real-time monitoring system based on a wireless sensor, comprising:

[0023] A first processing device, used to collect tire pressure data and filter the tire pressure data;

[0024] The second processing device is used to decode, verify and format the tire pressure data after filtering; dynamically adjust the frequency of data collection according to the driving state of the vehicle;

[0025] A third processing device is used to cross-validate and fuse the data collected by different sensors, and use the fused data as input for machine learning;

[0026] The fourth processing device is used to analyze the historical tire pressure data through a machine learning algorithm to the fused tire pressure data and establish a tire pressure model. The tire pressure model includes the relationship between the tire pressure and the temperature, load, and vehicle speed. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is promptly issued to the driver.

[0027] Preferably, the second processing device is used to determine the driving state of the vehicle through a vehicle speed sensor, and to reduce the sampling rate when the vehicle speed is lower than a preset threshold, and to increase the sampling rate when the vehicle speed is higher than the preset threshold.

[0028] Preferably, the third device comprises:

[0029] a first processing unit for determining the number and location of sensors to be installed in each tire;

[0030] The second processing unit is used to fuse the tire pressure data collected by multiple sensors using a Kalman filter algorithm, and dynamically estimate the real value of the tire pressure by establishing a state equation and an observation equation;

[0031] The third processing unit is used to organize the fused tire pressure data according to time series, construct a training data set, select an appropriate time window size, and extract statistical features of the data as input features of the machine learning model.

[0032] Preferably, the fourth processing device comprises:

[0033] The third processing unit is used to obtain historical tire pressure data of vehicle tires, including tire pressure value, temperature, load and vehicle speed related factor data, and construct a tire pressure data set;

[0034] A fourth processing unit, used for removing abnormal data and normalizing the tire pressure data set;

[0035] The fifth processing unit is used to train the tire pressure model and establish a relationship mapping between tire pressure and factors such as temperature, load and vehicle speed;

[0036] a sixth processing unit, configured to input the tire pressure data collected in real time into the trained tire pressure model to obtain a tire pressure value predicted by the model;

[0037] The seventh processing unit is used to calculate the degree of deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds a preset threshold; if it exceeds the threshold, the tire pressure is considered abnormal and a warning mechanism is triggered.

[0038] In response to problems such as harsh working environment and unstable signal transmission, the present invention adopts a sensor design that is resistant to high temperature and vibration, and combines adaptive power control and frequency hopping spread spectrum communication technology to ensure the reliability of tire pressure data collection and transmission. In order to improve data accuracy, the present invention introduces multi-sensor data fusion technology to eliminate abnormal data through cross-validation. At the same time, the sampling frequency is dynamically adjusted according to the vehicle speed to balance energy consumption and accuracy requirements. The present invention also uses a machine learning algorithm to establish a tire pressure model to analyze the relationship between tire pressure changes with factors such as temperature and load. When the real-time tire pressure data deviates from the predicted value, a warning is issued to the driver in a timely manner. Through the comprehensive application of these innovative technologies, the present invention significantly improves the accuracy, reliability and real-time performance of tire pressure monitoring, effectively ensures driving safety, and optimizes system energy consumption, which has important practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The present invention is a flow chart of a tire pressure real-time monitoring method based on a wireless sensor. DETAILED DESCRIPTION

[0040] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with the accompanying drawings and embodiments. The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant inventions, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0041] Embodiment 1:

[0042] like Figure 1 As shown, an embodiment of the present invention provides a tire pressure real-time monitoring method based on a wireless sensor, comprising:

[0043] Step S101, in view of the problem of the harsh working environment of the sensor, a high temperature and vibration resistant sensor design is adopted to improve the reliability of the sensor. Tire pressure data is collected from the sensor, and the collected tire pressure data is filtered through a software algorithm to remove noise interference introduced by environmental factors and ensure the accuracy of the tire pressure data.

[0044] According to the temperature range and vibration frequency of the sensor's working environment, determine the sensor's high temperature resistance and vibration resistance level, select sensor materials and structural designs that meet the requirements, and improve the reliability of the sensor. Collect tire pressure data through the sensor, convert the collected analog signal into a digital signal, and obtain the original tire pressure data. Preprocess the original tire pressure data to remove obvious outliers and data beyond the normal range to obtain the tire pressure data after preliminary processing. According to the frequency characteristics of noise interference introduced by environmental factors, design a digital filter suitable for tire pressure data, filter the tire pressure data after preliminary processing, remove environmental noise interference, and obtain filtered tire pressure data. Feature extraction is performed on the filtered tire pressure data to extract key feature parameters that reflect the tire pressure change trend and abnormal conditions. A machine learning model is established based on historical tire pressure data, and the model is trained using the extracted feature parameters to obtain a tire pressure prediction model. The tire pressure prediction model is used to predict the current tire pressure data to determine whether the tire pressure is abnormal. If the prediction result is abnormal, an alarm is triggered to prompt the driver to handle it in time to ensure driving safety.

[0045] Specifically, the reliability design of the sensor is the basis for ensuring the normal operation of the tire pressure monitoring system. Taking the internal environment of the tire as an example, the temperature range may fluctuate between -40℃ and 125℃, and the vibration frequency may reach 2000Hz. For this harsh environment, high-temperature resistant ceramic piezoelectric materials can be used to make pressure sensitive elements, and metal packaging structures can be used to improve vibration resistance. At the same time, the sensor structure design can be optimized through thermal stress analysis to reduce the impact of temperature changes on measurement accuracy. Data acquisition and preprocessing are key steps to obtain effective tire pressure information. The analog signal collected by the sensor is usually a 0-5V voltage signal, which can be converted into a digital signal in the range of 0-65535 using a 16-bit ADC. In the preprocessing stage, a reasonable threshold value can be set, such as 30-350kPa, and data outside this range can be marked as abnormal values. In addition, methods such as moving median filtering can be used to remove short-term pulse interference and improve data quality. The impact of environmental noise on tire pressure data cannot be ignored. The frequency of vibration noise generated by tire rolling is generally in the range of 50-500Hz, while the frequency of tire pressure changes is usually less than 1Hz. Based on this feature, a low-pass filter with a cutoff frequency of 5Hz can be designed to effectively filter out high-frequency noise. In specific implementation, a Butterworth filter can be used, which has good frequency response characteristics and can keep the signal amplitude unchanged in the passband while rapidly attenuating in the stopband. Feature extraction is an important part of data analysis. For tire pressure data, the features that can be extracted include statistical features such as mean, variance, peak, and spectrum features after Fourier transform. For example, the mean tire pressure reflects the tire inflation state, and the variance can indicate the degree of tire pressure fluctuation. In addition, the tire pressure change rate can be calculated to detect abnormal conditions such as rapid air leakage. These features together constitute a multidimensional feature vector that describes the tire pressure state. The establishment of a machine learning model provides a powerful tool for tire pressure prediction. Algorithms such as support vector machines (SVM) can be used to build a prediction model. First, collect historical tire pressure data containing normal and abnormal conditions, extract features, and divide them into training sets and test sets. Then, use the training set data to train the SVM model, and optimize the model parameters through cross-validation. Finally, use the test set to evaluate the model performance to ensure that the model has good generalization ability. In practical applications, the model prediction results can be used to determine whether the tire pressure is abnormal. For example, the warning threshold is set to ±20% of the standard tire pressure, and an alarm is triggered when the predicted tire pressure exceeds this range. The alarm method may include the lighting of the dashboard indicator light, voice prompts, etc. to attract the driver's attention. At the same time, the system can also predict possible problems such as slow air leakage and abnormal temperature increase based on the trend of tire pressure changes, and issue early warnings. The design of the entire tire pressure monitoring system reflects the comprehensive application of multidisciplinary knowledge. From sensor design to data processing, to the construction of machine learning models, each link is closely connected to jointly ensure the reliability and accuracy of the system.For example, the anti-interference design of the sensor provides high-quality raw data for subsequent data processing, while data preprocessing and filtering create good conditions for feature extraction, which ultimately enables the machine learning model to accurately predict the tire pressure status. In addition, the design of the system also takes into account the actual application scenarios. For example, when driving at high speeds, the system may require faster response speed and higher accuracy. This requires targeted optimization in terms of data acquisition frequency, filter design, and model selection. For example, the data acquisition frequency can be increased, a more complex adaptive filtering algorithm can be used, or a machine learning model with higher computational efficiency can be selected to meet real-time requirements. In general, this systematic tire pressure monitoring solution can not only detect tire pressure abnormalities in a timely manner, but also predict potential problems, greatly improving driving safety. By continuously collecting and analyzing data, the system can also continuously optimize and self-learn to further improve prediction accuracy, providing important support for the development of intelligent driving technology.

[0046] Step S102, preliminary processing is performed on the received tire pressure data, including data decoding, verification and formatting. The frequency of data collection is dynamically adjusted according to the driving state of the vehicle. The driving state of the vehicle is determined by the vehicle speed sensor. When the vehicle speed is lower than the preset threshold, the sampling rate is reduced, and when it is higher than the preset threshold, the sampling rate is increased to balance the contradiction between energy consumption and data accuracy. Through this optimized data collection and transmission strategy, unnecessary energy consumption is reduced.

[0047] The vehicle speed sensor data is received through the CAN bus to obtain the real-time driving speed of the vehicle, and the current driving state of the vehicle is judged according to the preset vehicle speed threshold. If the vehicle speed is lower than the preset low-speed threshold, the collection frequency of the tire pressure data is reduced to the preset low-frequency sampling rate; if the vehicle speed is higher than the preset high-speed threshold, the collection frequency of the tire pressure data is increased to the preset high-frequency sampling rate. According to the correspondence between the vehicle speed and the sampling rate, a vehicle speed-sampling rate mapping table is established, and the mapping table is stored in the non-volatile memory of the controller for subsequent fast query and call. After the tire pressure sensor collects the tire pressure data, the original data is transmitted to the tire pressure data processing unit, and the received tire pressure data is decoded to obtain the original value of the tire pressure. The decoded original tire pressure value is checked for deviation, and the deviation between the tire pressure measurement value and the calibration value is calculated by comparing it with the calibration value. If the deviation exceeds the preset allowable range, it is determined to be abnormal data. The verified tire pressure data is formatted and converted into a unified data format for subsequent storage and analysis. At the same time, the data is graded and marked according to the importance of the data. According to the driving status of the vehicle and the frequency of tire pressure data collection, the transmission strategy of tire pressure data is dynamically adjusted to reduce the frequency and amount of data transmission as much as possible while ensuring the real-time nature of the data, thereby reducing communication energy consumption.

[0048] Specifically, by receiving the vehicle speed sensor data through the controller area network bus, the vehicle's driving speed information can be obtained in real time. For example, when the vehicle is driving on a city road, the vehicle speed sensor may detect that the vehicle speed is 30 kilometers per hour, and this data will be transmitted to the controller of the tire pressure monitoring system through the controller area network bus. The controller determines whether the current vehicle is in a low-speed, medium-speed or high-speed driving state based on a preset vehicle speed threshold, such as a low-speed threshold of 20 kilometers per hour and a high-speed threshold of 60 kilometers per hour. If the current vehicle speed is lower than the low-speed threshold, it means that the vehicle may be in an idling or slow driving state. At this time, the change of tire pressure is relatively slow, and the frequency of collecting tire pressure data can be reduced to save energy. For example, the frequency of collecting tire pressure data can be reduced from once per second to once every 5 seconds. On the contrary, if the vehicle speed is higher than the high-speed threshold, it means that the vehicle is in a high-speed driving state, and the change of tire pressure may be faster. It is necessary to increase the frequency of collecting tire pressure data to ensure the real-time and accuracy of monitoring. For example, the frequency of collecting tire pressure data can be increased from once per second to once every 2 seconds. In order to facilitate the controller to quickly adjust the sampling frequency, a vehicle speed-sampling rate mapping table can be established in advance. For example, when the vehicle speed is 0 to 20 kilometers per hour, the sampling rate is once every 5 seconds; when the vehicle speed is 20 to 60 kilometers per hour, the sampling rate is once per second; when the vehicle speed is above 60 kilometers per hour, the sampling rate is once every 2 seconds. This mapping table can be stored in the non-volatile memory of the controller, such as flash memory, so that it can be quickly read and used after the vehicle is started. After the tire pressure sensor collects the tire pressure data, it is usually transmitted in a specific encoding format. For example, the sensor may encode the tire pressure value into a 16-bit data packet, in which the upper 8 bits represent the integer part and the lower 8 bits represent the decimal part. After the tire pressure data processing unit receives this data packet, it needs to decode it according to the predefined encoding rules to obtain the original value of the tire pressure. For example, the received data packet is 0x03E8, the high 8 bits 0x03 are converted to decimal 3, the low 8 bits 0xE8 are converted to decimal 232, and then the decimal part is divided by 256 to obtain the value of the decimal part, and the final tire pressure value is 3.9 standard atmospheric pressure. In practical applications, due to the error of the sensor itself or the influence of environmental factors, the collected tire pressure value may have a certain deviation. In order to ensure the accuracy of the tire pressure data, it is necessary to perform a deviation check on the decoded tire pressure original value. For example, the tire pressure sensor can be calibrated in a laboratory environment in advance to obtain a standard tire pressure value, such as 3.0 standard atmospheric pressure. In actual measurement, if the deviation between the measured value and the calibrated value exceeds the preset allowable range, such as plus or minus 0.1 standard atmospheric pressure, the data is considered abnormal and needs to be specially processed or discarded. In order to facilitate subsequent storage and analysis, the verified tire pressure data needs to be formatted. For example, the tire pressure value can be converted into a unified unit, such as kilopascals, and the data can be arranged in a certain format.At the same time, data can be graded and marked according to their importance. For example, normal tire pressure data can be marked as 0, slightly abnormal tire pressure data can be marked as 1, and seriously abnormal tire pressure data can be marked as 2. The transmission strategy of tire pressure data can be dynamically adjusted according to the driving status of the vehicle and the frequency of tire pressure data collection. For example, when driving at low speeds, since the frequency of tire pressure data collection is low, the amount of data transmitted each time can be appropriately increased to reduce the number of transmissions. When driving at high speeds, since the frequency of tire pressure data collection is high, the amount of data transmitted each time can be reduced to improve the real-time performance of the transmission. In this way, the frequency and amount of data transmission can be reduced as much as possible while ensuring the real-time performance of the data, thereby reducing communication energy consumption and extending the battery life of the sensor.

[0049] After the tire pressure sensor collects the tire pressure data, it transmits the original data to the tire pressure data processing unit, decodes the received tire pressure data, and obtains the original value of the tire pressure.

[0050] Based on the original tire pressure data collected by the tire pressure sensor, the original data is sent to the tire pressure data processing unit through data transmission; after the tire pressure data processing unit receives the transmitted original tire pressure data, it uses a preset decoding algorithm to decode the original data; based on the result of the decoding process, the original value of the tire pressure data is obtained; it is determined whether the decoded original tire pressure value is within a preset normal range, if it exceeds the normal range, an abnormal tire pressure alarm is triggered; based on the original tire pressure value, the Kalman filter algorithm is used to filter the tire pressure data to eliminate noise interference in the tire pressure data; based on the tire pressure data after filtering, the least squares fitting algorithm is used to obtain a change curve of the tire pressure data; based on the tire pressure data change curve, it is determined whether the tire pressure fluctuates abnormally, if abnormal fluctuations occur, an abnormal tire pressure warning signal is triggered to prompt the driver to check the tire condition in time.

[0051] Specifically, the raw data collected by the tire pressure sensor usually contains a variety of information, such as tire pressure value, temperature, sensor ID, etc. These data are encoded in a specific format and need to be decoded to obtain valid information. For example, the raw data of a certain model of sensor may be a string of hexadecimal numbers "A5B3C2D1E0", where the first two digits represent the tire pressure value, the middle two digits represent the temperature, and the latter is the sensor ID. The original tire pressure value obtained after decoding may be 165kPa. The preset normal tire pressure range is usually determined by the vehicle type and tire specifications. Taking an ordinary car as an example, the normal tire pressure range may be 180-220kPa. If the decoded tire pressure value is 165kPa, which is lower than the normal range, the system will trigger an abnormal tire pressure alarm to remind the driver to pay attention. The Kalman filter algorithm is a recursive filter that can effectively eliminate measurement noise and improve data accuracy. In tire pressure monitoring, the original tire pressure data may fluctuate due to factors such as road bumps and temperature changes. Through Kalman filtering, smoother tire pressure data can be obtained. For example, the original data may be [200, 198, 203, 197, 201] kPa, and after filtering, it may become [200, 199.5, 200.5, 199.8, 200.2] kPa, which effectively reduces data fluctuations. The least squares fitting algorithm is used to analyze the changing trend of tire pressure data. By fitting the tire pressure data over a period of time, a curve showing the change of tire pressure over time can be obtained. This helps predict future tire pressure changes and detect potential problems early. For example, if the fitted curve shows a clear downward trend, even if the current tire pressure is still within the normal range, it may indicate that the tire has a slow leak problem. The judgment criteria for abnormal fluctuations in tire pressure can be based on multiple factors, such as the rate of change of tire pressure, the amplitude of fluctuation, etc. Assuming that the tire pressure drops by more than 20 kPa in a short period of time (such as 5 minutes), or the tire pressure values ​​measured for multiple consecutive times fluctuate by more than 10 kPa, the system may determine it as an abnormal fluctuation and trigger an early warning signal. This early warning mechanism can help drivers detect tire problems such as nails and cracks in a timely manner, thereby improving driving safety. The entire tire pressure monitoring process reflects a complete chain from data collection, processing to analysis and early warning. Through decoding, filtering and other processing, the reliability and accuracy of the original data are improved. The fitting algorithm and abnormal judgment logic are used to achieve real-time monitoring and prediction of tire pressure status. This intelligent tire pressure monitoring system can not only detect safety hazards in a timely manner, but also provide a basis for vehicle maintenance through data analysis, effectively extend the service life of tires, and improve driving safety and economy.

[0052] Step S103: To further improve the reliability of tire pressure data, multiple sensors are installed in each tire, and multi-sensor data fusion technology is introduced. The data collected by different sensors are cross-validated and fused to eliminate abnormal data that may appear in a single sensor, thereby improving the robustness of the entire system. The fused data is used as the input for the next step of machine learning.

[0053] According to the internal spatial structure of the tire and the technical parameters of the sensor, the number and position of the sensors installed in each tire are determined to ensure that the sensors can fully collect the pressure data in the tire. For the tire pressure data collected by each sensor, a data preprocessing algorithm is designed to perform denoising and normalization on the data to improve the data quality and prepare for subsequent data fusion. The Kalman filter algorithm is used to fuse the tire pressure data collected by multiple sensors. By establishing the state equation and observation equation, the true value of the tire pressure is dynamically estimated to eliminate the abnormal data of a single sensor. On the basis of data fusion, an abnormal data detection algorithm is introduced to determine whether the fused tire pressure data exceeds the normal range by setting a threshold, so as to detect abnormal conditions in time and improve the robustness of the system. The fused tire pressure data is organized according to the time series, a training data set is constructed, and an appropriate time window size is selected to extract the statistical features of the data, such as mean, variance, kurtosis, etc., as the input features of the machine learning model. The support vector machine algorithm is used to establish a classification model for tire pressure data. By training historical data, the characteristic patterns of normal and abnormal tire pressures are learned to achieve real-time classification and early warning of newly collected tire pressure data. Based on the machine learning model and combined with expert knowledge, an interpretation model for tire pressure data is established. According to the changing trend and key indicators of tire pressure, the health status of the tire is judged to provide decision support for the safe operation of the vehicle.

[0054] Specifically, the internal space structure of the tire and the technical parameters of the sensor are the key factors that determine the number and location of sensor installation. Taking an ordinary car tire as an example, its internal space is in the shape of a ring, with a diameter of about 60 cm and a width of about 20 cm. Considering the size of the sensor (usually no more than 5 cm in length) and the measurement range, 3-4 sensors can be evenly installed on the inner wall of the tire to ensure full coverage of the internal pressure distribution of the tire. Data preprocessing is an important step to improve the quality of tire pressure data. Taking denoising as an example, the moving average filter method can be used to smooth the continuously collected tire pressure data. Assuming that the collected original data sequence is [230,228,235,229,231] kPa, and the window size is 3, the processed data is [231,231,232] kPa, which effectively reduces data fluctuations. Normalization processing can unify data of different dimensions into the [0,1] interval for subsequent processing. The Kalman filter algorithm plays an important role in multi-sensor data fusion. Taking two sensors as an example, assuming that sensors A and B measure the tire pressure at a certain moment as 235kPa and 233kPa respectively, considering their respective measurement errors, the Kalman filter algorithm will estimate the tire pressure closer to the true value, such as 234kPa, based on historical data and current observations. This method not only improves the measurement accuracy, but also effectively eliminates abnormal data from a single sensor. Abnormal data detection is the key to ensuring the robustness of the system. The normal tire pressure range can be set to 180-250kPa. When the fused tire pressure data exceeds this range, the system will issue an alarm. For example, if the tire pressure is detected to drop suddenly to 150kPa, the system will immediately remind the driver to prevent safety hazards caused by abnormal tire pressure. Feature extraction of time series data is the basis for the input of machine learning models. With a 30-second time window, the statistical characteristics of tire pressure data such as mean, variance, and kurtosis can be calculated. For example, under normal driving conditions, the mean tire pressure within 30 seconds may be 235kPa, with a variance of 2 and a kurtosis close to 3; when the tire pressure is abnormal, these characteristic values ​​may change significantly, providing an effective basis for judgment for the machine learning model. The support vector machine (SVM) algorithm performs well in tire pressure data classification. By training a historical data set containing normal and abnormal tire pressure data, the SVM can learn the characteristic boundaries of data of different categories. When new tire pressure data is input, the SVM can quickly determine its category and achieve real-time warning. For example, if it is detected that the tire pressure continues to drop and the drop exceeds the preset threshold, the system will issue a tire pressure abnormality warning in time. The explanation model that combines machine learning and expert knowledge can more comprehensively evaluate the health status of the tire. For example, if the system detects that the tire pressure drops sharply in a short period of time and is accompanied by a rise in temperature, it may be judged that the tire is punctured; if the tire pressure drops slowly and the temperature does not change much, it may be caused by normal gas diffusion. This intelligent diagnostic system can provide drivers with more accurate tire status information, which helps to take corresponding maintenance measures in time and improve driving safety.

[0055] Step S104, applying a machine learning algorithm to the fused tire pressure data, analyzing historical tire pressure data, and establishing a model of normal tire pressure. The model includes the relationship between tire pressure and temperature, load, vehicle speed, and other factors. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is issued to the driver in a timely manner to remind him to check the tire condition and deal with potential safety hazards in a timely manner to ensure the driving safety of the vehicle.

[0056] Obtain the historical tire pressure data of the vehicle tire, including tire pressure value, temperature, load, speed and other related factors data, and construct a tire pressure data set. Preprocess the tire pressure data set, remove abnormal data, and normalize the data to ensure data quality and consistency. Select a suitable machine learning algorithm, such as support vector machine or neural network, train the normal tire pressure model, and establish the relationship mapping between tire pressure and factors such as temperature, load, and speed. Collect the vehicle's tire pressure data in real time through the on-board sensor to obtain the current tire pressure value, temperature, load, speed and other parameters. Input the real-time collected tire pressure data into the trained normal tire pressure model to obtain the normal tire pressure value predicted by the model. Calculate the degree of deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds the preset threshold. If it exceeds the threshold, the tire pressure is considered abnormal and the warning mechanism is triggered. Issue a tire pressure abnormality warning to the driver, prompting him to check the tire condition, deal with potential safety hazards in time, and ensure vehicle driving safety. According to the warning information, the driver can take measures such as reducing the speed and stopping for inspection to troubleshoot the cause of abnormal tire pressure and ensure driving safety.

[0057] Specifically, the core of the tire pressure monitoring system is to build an accurate normal tire pressure model. First, a large amount of historical tire pressure data needs to be collected, including tire pressure values ​​under different temperatures, loads and vehicle speeds. For example, in a high temperature environment in summer, the gas in the tire expands and causes the tire pressure to increase; while in a low temperature in winter, the tire pressure will decrease accordingly. The increase in load will aggravate the deformation of the tire, and the tire pressure will increase accordingly; the friction and heat generated by the tire when driving at high speed will also cause the tire pressure to increase. The data preprocessing stage is crucial, and outliers need to be removed and normalized. Suppose a set of tire pressure data is collected, in which outliers that are obviously deviated from the normal range appear, such as the tire pressure measured at room temperature is 0.5 bar, which is far below the normal level. This data may be caused by sensor failure or human operation error and should be eliminated. Normalization processing can unify data of different dimensions to the same scale, which is convenient for subsequent modeling. Choosing a suitable machine learning algorithm has a significant impact on model performance. Support vector machines are suitable for processing high-dimensional feature spaces and can effectively capture the nonlinear relationship between tire pressure and various influencing factors. Neural networks have strong learning capabilities and can automatically extract deep features. In practical applications, multiple algorithms can be tried and their performance can be compared to select the best solution. Real-time data collection is a key link in tire pressure monitoring. Modern vehicle-mounted sensors can collect tire pressure data at high frequency and high accuracy. For example, a certain sensor can collect tire pressure data 10 times per second with an accuracy of ±0.1 bar. These real-time data will be input into the trained model and compared with the predicted value. The judgment of the degree of deviation requires setting a reasonable threshold. Assuming that the normal tire pressure range is 2.2-2.5 bar, the threshold can be set to ±0.3 bar. When the difference between the actual tire pressure and the predicted value exceeds this range, the system will trigger an alarm. The setting of this threshold requires a balance between sensitivity and false alarm rate. Too low may lead to frequent false alarms, while too high may ignore potential risks. The design of the warning mechanism should take into account the driver's perception and reaction characteristics. A combination of sound and light can be used, such as a red warning light displayed on the dashboard accompanied by a beep. At the same time, the vehicle-mounted system can voice broadcast specific abnormal conditions, such as "The left front tire pressure is too low, please check in time", to help the driver quickly locate the problem. The driver's response after receiving the warning is also very important. If you encounter abnormal tire pressure when driving at high speed, you should first slow down to a safe speed and avoid sudden braking or sharp turns. After finding a safe place to park, you should carefully check whether the tires are obviously damaged or have foreign objects stuck in them. If you cannot determine the cause, it is best to contact professional rescue services to ensure driving safety. This complete tire pressure monitoring system can greatly improve driving safety. It can not only detect abnormal tire pressure in time, but also predict tire life through data analysis, providing a scientific basis for vehicle maintenance. This data-driven intelligent management method represents the development direction of automotive safety technology.

[0058] Embodiment 2:

[0059] The embodiment of the present invention further provides a tire pressure real-time monitoring system based on a wireless sensor, comprising:

[0060] A first processing device, used to collect tire pressure data and filter the tire pressure data;

[0061] The second processing device is used to decode, verify and format the tire pressure data after filtering; dynamically adjust the frequency of data collection according to the driving state of the vehicle;

[0062] A third processing device is used to cross-validate and fuse the data collected by different sensors, and use the fused data as input for machine learning;

[0063] The fourth processing device is used to analyze the historical tire pressure data through a machine learning algorithm to the fused tire pressure data and establish a tire pressure model. The tire pressure model includes the relationship between the tire pressure and the temperature, load, and vehicle speed. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is promptly issued to the driver.

[0064] As an implementation mode of an embodiment of the present invention, the second processing device is used to determine the driving state of the vehicle through a vehicle speed sensor, and reduce the sampling rate when the vehicle speed is lower than a preset threshold, and increase the sampling rate when the vehicle speed is higher than the preset threshold.

[0065] As an implementation manner of the embodiment of the present invention, the third device includes:

[0066] a first processing unit for determining the number and location of sensors to be installed in each tire;

[0067] The second processing unit is used to fuse the tire pressure data collected by multiple sensors using a Kalman filter algorithm, and dynamically estimate the real value of the tire pressure by establishing a state equation and an observation equation;

[0068] The third processing unit is used to organize the fused tire pressure data according to time series, construct a training data set, select an appropriate time window size, and extract statistical features of the data as input features of the machine learning model.

[0069] As an implementation manner of the embodiment of the present invention, the fourth processing device includes:

[0070] The third processing unit is used to obtain historical tire pressure data of vehicle tires, including tire pressure value, temperature, load and vehicle speed related factor data, and construct a tire pressure data set;

[0071] A fourth processing unit, used for removing abnormal data and normalizing the tire pressure data set;

[0072] The fifth processing unit is used to train the tire pressure model and establish a relationship mapping between tire pressure and factors such as temperature, load and vehicle speed;

[0073] a sixth processing unit, configured to input the tire pressure data collected in real time into the trained tire pressure model to obtain a tire pressure value predicted by the model;

[0074] The seventh processing unit is used to calculate the degree of deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds a preset threshold; if it exceeds the threshold, the tire pressure is considered abnormal and a warning mechanism is triggered.

[0075] The above description is merely a preferred embodiment of one or more embodiments of the present specification and is not intended to limit one or more embodiments of the present specification. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present specification shall be included in the scope of protection of one or more embodiments of the present specification.

Claims

1. A tire pressure real-time monitoring method based on wireless sensor, characterized in that: include: Step 101, collecting tire pressure data and filtering the tire pressure data; Step 102: decoding, verifying and formatting the tire pressure data after filtering; dynamically adjusting the frequency of data collection according to the driving state of the vehicle; Step 103: cross-validate and fuse the data collected by different sensors, and use the fused data as input for machine learning; Step 104: Analyze historical tire pressure data using a machine learning algorithm for the fused tire pressure data to establish a tire pressure model. The tire pressure model includes the relationship between tire pressure changes with temperature, load, and vehicle speed. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is promptly issued to the driver.

2. The tire pressure real-time monitoring method based on wireless sensor according to claim 1, characterized in that: In step S102, the vehicle driving state is determined by a vehicle speed sensor. When the vehicle speed is lower than a preset threshold, the sampling rate is reduced, and when the vehicle speed is higher than the preset threshold, the sampling rate is increased.

3. The tire pressure real-time monitoring method based on wireless sensor according to claim 2, characterized in that: Step 103 includes: Determine the number and location of sensors to be installed in each tire; The Kalman filter algorithm is used to fuse the tire pressure data collected by multiple sensors, and the true value of the tire pressure is dynamically estimated by establishing the state equation and observation equation; The fused tire pressure data is organized in time series to construct a training data set. An appropriate time window size is selected to extract the statistical features of the data as input features of the machine learning model.

4. The tire pressure real-time monitoring method based on wireless sensor according to claim 3, characterized in that: Step 104 includes: Obtain historical tire pressure data of vehicle tires, including tire pressure value, temperature, load and vehicle speed related factor data, and build a tire pressure dataset; Remove abnormal data and normalize the tire pressure data set; Train the tire pressure model and establish the relationship mapping between tire pressure and factors such as temperature, load and vehicle speed; Input the real-time collected tire pressure data into the trained tire pressure model to obtain the tire pressure value predicted by the model; Calculate the degree of deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds the preset threshold; If the threshold is exceeded, the tire pressure is considered abnormal and the warning mechanism is triggered.

5. A tire pressure real-time monitoring system based on wireless sensors, characterized in that: include: A first processing device, used to collect tire pressure data and filter the tire pressure data; A second processing device, used for decoding, verifying and formatting the tire pressure data after filtering; Dynamically adjust the frequency of data collection according to the driving status of the vehicle; A third processing device is used to cross-validate and fuse the data collected by different sensors, and use the fused data as input for machine learning; The fourth processing device is used to analyze the historical tire pressure data through a machine learning algorithm to the fused tire pressure data and establish a tire pressure model. The tire pressure model includes the relationship between the tire pressure and the temperature, load, and vehicle speed. When the tire pressure data collected in real time deviates from the model prediction value and exceeds a preset threshold, a warning is promptly issued to the driver.

6. The tire pressure real-time monitoring system based on wireless sensor as claimed in claim 5, characterized in that: The second processing device is used to determine the driving state of the vehicle through a vehicle speed sensor, reduce the sampling rate when the vehicle speed is lower than a preset threshold, and increase the sampling rate when the vehicle speed is higher than the preset threshold.

7. The tire pressure real-time monitoring system based on wireless sensor as claimed in claim 6, characterized in that: The third device comprises: a first processing unit for determining the number and location of sensors to be installed in each tire; The second processing unit is used to fuse the tire pressure data collected by multiple sensors using a Kalman filter algorithm, and dynamically estimate the real value of the tire pressure by establishing a state equation and an observation equation; The third processing unit is used to organize the fused tire pressure data according to time series, construct a training data set, select an appropriate time window size, and extract statistical features of the data as input features of the machine learning model.

8. The tire pressure real-time monitoring system based on wireless sensor as claimed in claim 7, characterized in that: The fourth processing device comprises: The third processing unit is used to obtain historical tire pressure data of vehicle tires, including tire pressure value, temperature, load and vehicle speed related factor data, and construct a tire pressure data set; A fourth processing unit, used for removing abnormal data and normalizing the tire pressure data set; The fifth processing unit is used to train the tire pressure model and establish a relationship mapping between tire pressure and factors such as temperature, load and vehicle speed; a sixth processing unit, configured to input the tire pressure data collected in real time into the trained tire pressure model to obtain a tire pressure value predicted by the model; The seventh processing unit is used to calculate the degree of deviation between the real-time tire pressure value and the model prediction value, and determine whether the deviation exceeds a preset threshold; if it exceeds the threshold, the tire pressure is considered abnormal and a warning mechanism is triggered.

Images