Intelligent tire pressure monitoring system

Abstract

The invention provides an intelligent tire pressure monitoring system, which adopts a collaborative architecture of a distributed sensor group and an intelligent central processing unit, breaks through the limitation that a traditional TPMS (tire pressure monitoring system) only monitors static tire pressure through fusion detection of three parameters of pressure, temperature and vibration, and realizes three-dimensional evaluation of a tire health state; an innovative dynamic working mode switching mechanism automatically adjusts a detection strategy according to a vehicle state, and the energy consumption of a parking mode is reduced by more than 80% while the monitoring precision is ensured; based on a tire state evaluation model, through real-time analysis of vibration spectrum characteristics, pavement type self-identification and alarm threshold dynamic adjustment functions are realized, and a false alarm rate is reduced. Response delay is compressed to 100 milliseconds by the aid of the red light mode quick awakening technology, and monitoring continuity under urban working conditions is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle safety monitoring, and in particular to an intelligent tire pressure monitoring system. Background Art

[0002] With the rapid development of the automotive industry and increasing awareness of driving safety, tire pressure monitoring systems (TPMS), a crucial component of vehicle safety systems, have become standard equipment in modern vehicles. By monitoring parameters such as tire pressure and temperature in real time, TPMS provides drivers with timely tire status information, effectively preventing accidents caused by abnormal tire pressure.

[0003] Currently, tire pressure monitoring systems on the market are primarily categorized as direct and indirect. Direct tire pressure monitoring systems use sensors installed inside the tire to directly measure tire pressure and temperature and transmit the data to an onboard receiver. Indirect tire pressure monitoring systems use the vehicle's ABS system to indirectly calculate tire pressure changes. When some sensors fail, existing systems often lack effective backup mechanisms, resulting in a complete loss of monitoring functionality and compromising driving safety. Summary of the Invention

[0004] In view of this, the present invention provides an intelligent tire pressure monitoring system to solve the technical defects in the prior art.

[0005] Specifically, the present invention provides an intelligent tire pressure monitoring system, comprising: It includes a distributed sensor group and a central processing unit installed in the tire; The sensor group includes a pressure sensor, a temperature sensor, and a three-axis accelerometer to collect real-time tire pressure, temperature, and vibration spectrum; The central processing unit switches working modes according to the vehicle status, including periodic low-power detection in parking mode, dynamic high-frequency detection in driving mode, and short-term sleep detection in red light mode; The system calculates the current tire pressure health through a trained tire condition assessment model, which is trained by fusing collected data from a sensor group. The tire condition assessment model dynamically adjusts the alarm threshold based on road surface characteristics.

[0006] In some embodiments, the vehicle state includes a parking mode, specifically including: When the vehicle speed signal obtained by the central processing unit through the CAN bus is zero, the control sensor group enters an intermittent working state of being activated once every hour; If the pressure drop rate is detected to exceed the preset threshold, the detection frequency will be automatically increased; The pressure sensor collects the original analog signal and performs AD conversion to generate a digital pressure value; The temperature sensor collects ambient temperature data and performs linear compensation to generate a compensated temperature value; All detection data are transmitted to the central processing unit through the LoRa wireless communication module.

[0007] In some embodiments, the vehicle status includes a driving mode, specifically including: The triaxial accelerometer monitors the tire ground contact deformation characteristics in real time, and the raw accelerometer signal is processed by Kalman filtering to generate denoised vibration data; Combined with the yaw rate data from the vehicle's ESP system, the actual tire load distribution is calculated using the least squares method; The central processing unit dynamically adjusts the sampling frequency of each tire based on the load distribution results, and the sampling frequency adjustment range is 10-50Hz; The vibration sensor collects the time domain signal and performs FFT transformation to generate the frequency domain energy spectrum.

[0008] In some embodiments, the road surface feature recognition process specifically includes: The original signal collected by the vibration sensor is decomposed by wavelet packets, and the energy value of the 30-50Hz frequency band is extracted as the characteristics of ice and snow roads, and the energy value of the 80-120Hz frequency band is extracted as the characteristics of gravel roads. The energy values ​​of each frequency band are compared with the benchmark model established by historical data, and the sliding average algorithm is used to generate the estimated road adhesion coefficient; The central processing unit corrects the standard tire pressure threshold proportionally based on the estimated road adhesion coefficient.

[0009] In some embodiments, the red light mode specifically includes: When the central processing unit identifies that the vehicle is within the stop line of the traffic light through GPS positioning information, it controls the sensor group to enter a deep sleep state; At the same time, the backup capacitor power supply circuit is started to maintain the basic clock operation; the accelerometer original signal is processed by waveform matching to generate a characteristic vibration waveform; When a characteristic vibration waveform is detected, all sensors are woken up and self-tested within 100 milliseconds.

[0010] In some embodiments, an energy management subsystem is further included, specifically including: In driving mode, a micro-generator driven by tire rotation charges the sensor battery; The generator output voltage is rectified and stabilized to generate charging current; Switching to the main power supply to the central processing unit in parking mode; The power management chip monitors the supercapacitor voltage in real time and automatically shuts down the communication function of non-core sensors when the voltage is lower than the threshold.

[0011] In some embodiments, the formula for calculating tire pressure health includes:

[0012] in: For tire pressure health, The tire pressure value obtained for the i-th sampling comes from the real-time data of the pressure sensor; The average tire pressure value during the statistical period; The tire temperature value obtained for the jth sampling is derived from the real-time data of the temperature sensor; Preset value for optimal working temperature; is the spectrum entropy value of the kth vibration sensor, which comes from the wavelet analysis result of the vibration signal; n, m, and p are the sampling times of pressure, temperature, and vibration data, respectively.

[0013] In some embodiments, the formula for calculating the spectrum entropy value includes:

[0014] in: is the spectrum entropy value, is the normalized energy spectral density at frequency f, which comes from the FFT transform result of the vibration sensor; is the wavelet packet energy of the bth frequency band, which comes from the three-level wavelet packet decomposition; is the maximum energy value of the entire frequency band; b is a natural number, indicating different frequency band division intervals, including: 20-50Hz, 50-80Hz, 80-120Hz, 120-160Hz and 160-200Hz.

[0015] In some embodiments, a cloud-based collaborative diagnosis module is also included, specifically including: The central processing unit uploads the key parameters output by the tire condition assessment model to the cloud server via the onboard T-BOX; After receiving the data, the cloud performs feature extraction and processing to generate a fault feature vector; The cloud server combines meteorological data and historical fault records to generate preventive maintenance recommendations using a decision tree algorithm; When the detected abnormal tire pressure pattern matches the typical fault cases stored in the cloud and exceeds the threshold, visual maintenance instructions are pushed to the vehicle display screen.

[0016] In some embodiments, a failure protection mechanism is also included, specifically including: When communication with any sensor group is interrupted, the central processing unit reconstructs the missing data based on the data from the remaining normal sensors and the vehicle dynamics model; The Kalman filter algorithm is used to process the vibration data of adjacent tires to generate correlation coefficients; When reconstructing missing data, the vibration correlation coefficient of adjacent tires is used as the weight factor. The correlation coefficient is calculated by the LSTM network trained with historical driving data.

[0017] At least one embodiment of the present invention utilizes a collaborative architecture combining a distributed sensor array with an intelligent central processing unit. By integrating pressure, temperature, and vibration, the system overcomes the limitations of traditional TPMS systems, which only monitor static tire pressure, and achieves comprehensive tire health assessment. An innovative dynamic operating mode switching mechanism automatically adjusts detection strategies based on vehicle status, reducing parking mode energy consumption by over 80% while maintaining monitoring accuracy. A tire condition assessment model based on machine learning analyzes vibration spectrum characteristics in real time, enabling the first self-identification of road surface types and dynamic adjustment of alarm thresholds, reducing false alarm rates by over 60%. The system's unique red light mode rapid wake-up technology reduces response latency to under 100 milliseconds, ensuring continuous monitoring in urban conditions. Compared to traditional solutions, this system achieves breakthrough improvements in three key metrics: monitoring dimensionality, energy efficiency, and warning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a structural block diagram of an intelligent tire pressure monitoring system provided by the present invention. DETAILED DESCRIPTION

[0019] The following description sets forth many specific details to facilitate a thorough understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0020] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms of "a" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items. The modifications of "one" and "a plurality" mentioned in this disclosure are illustrative and not restrictive, and those skilled in the art should understand that unless the context clearly indicates otherwise, it should be understood as "one or more".

[0021] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0022] See also Figure 1 , Figure 1 The following is a structural block diagram of an intelligent tire pressure monitoring system provided according to some embodiments of this specification, wherein the intelligent tire pressure monitoring system includes: An intelligent tire pressure monitoring system includes a distributed sensor group installed within the tire and a central processing unit (CPU). The sensor group includes a pressure sensor, a temperature sensor, and a three-axis accelerometer, which are used to collect real-time tire pressure, temperature, and vibration spectrum. The CPU switches operating modes based on the vehicle's status, including periodic low-power detection in parking mode, dynamic high-frequency detection in driving mode, and short-term sleep detection in red light mode. The system calculates the current tire pressure health using a trained tire condition assessment model, which is trained by fusing collected data from the sensor group. The tire condition assessment model dynamically adjusts the alarm threshold based on road surface characteristics.

[0023] The sensor group of an intelligent tire pressure monitoring system adopts a distributed architecture, with a set of independent sensor units installed in each tire. Each sensor unit includes a piezoresistive pressure sensor, a PT100 temperature sensor, and a three-axis MEMS accelerometer. The pressure sensor adopts a silicon-based piezoresistive design, with a measurement range of 0-800kPa, an accuracy of ±1kPa, and a sampling frequency that can be dynamically adjusted between 1-50Hz. The temperature sensor has a measurement range of -40℃ to 120℃ and an accuracy of ±0.5℃. The three-axis accelerometer has a measurement range of ±16g and a sampling frequency of up to 200Hz, which is used to capture the vibration characteristics generated when the tire contacts the road.

[0024] The sensor cluster integrates signal conditioning circuitry, an analog-to-digital converter, and a low-power microcontroller. The microcontroller utilizes the ARM Cortex-M0+ architecture, operates at 48MHz, and features 256KB of flash memory and 32KB of RAM. It is responsible for data acquisition, preliminary processing, and wireless transmission. The sensor cluster exchanges data with the central processing unit via a LoRa wireless communication module, operating at 433MHz. The communication range is up to 50 meters, sufficient to cover the distance from the tire to the central processing unit for various vehicle models.

[0025] The central processing unit (CPU), mounted below the vehicle's dashboard, utilizes a dual-core processor architecture with a main frequency of 1.2GHz, 4GB of RAM, and 32GB of flash memory, running a real-time operating system. The CPU connects to the vehicle's other electronic control units via the CAN bus, acquiring information such as vehicle speed, steering angle, and brake pressure to determine vehicle status and switch operating modes. The CPU is also equipped with a GPS module to obtain vehicle location information with an accuracy of 3 meters and an update frequency of 1Hz.

[0026] An intelligent tire pressure monitoring system automatically switches between three operating modes depending on the vehicle's status: parking mode, driving mode, and red light mode. In parking mode, if the vehicle speed signal received by the central processing unit via the CAN bus is zero and persists for more than three minutes, the system controls the sensor group to enter an intermittent operating state, activating once every hour. During each activation, the pressure sensor collects the raw analog signal, performs A / D conversion to generate a digital pressure value, and the temperature sensor collects ambient temperature data for linear compensation to generate a compensated temperature value. If the pressure drop rate detected exceeds a preset threshold (typically set at 5 kPa / hour), the detection frequency is automatically increased, for example, to once every second, to promptly detect potential air leaks. All detection data is transmitted to the central processing unit via the LoRa wireless communication module, with a transmission power of 10 dBm and a data packet size of 64 bytes.

[0027] In driving mode, when the vehicle speed exceeds 5 km / h, the triaxial accelerometer monitors tire contact deformation characteristics in real time, with the sampling frequency increased to 50 Hz. The raw accelerometer signal is processed using a Kalman filter to generate denoised vibration data. Filter parameters Q = 0.01 and R = 0.1 are used to eliminate the influence of random road noise. The system combines yaw rate data from the vehicle's ESP system and uses the least squares method to calculate the actual tire load distribution. The central processing unit dynamically adjusts the sampling frequency of each tire based on the load distribution results. The sampling frequency adjustment range is 10-50 Hz, with higher sampling frequencies for heavier tires. The vibration sensor collects the time-domain signal and performs an FFT transform to generate a frequency-domain energy spectrum. The FFT has 1024 points and uses a Hanning window function.

[0028] The system also implements road feature recognition, performing wavelet packet decomposition of the raw signals collected by the vibration sensor using the db4 wavelet basis function with five decomposition layers. The system extracts energy values ​​in the 30-50Hz frequency band as characteristics of icy and snowy roads, and energy values ​​in the 80-120Hz frequency band as characteristics of gravel roads. The energy values ​​in each frequency band are compared with a baseline model established using historical data, and a sliding average algorithm is used to generate an estimated road adhesion coefficient using a 10-second sliding window. The central processing unit proportionally adjusts the standard tire pressure threshold based on the estimated road adhesion coefficient. For example, on icy and snowy roads, the tire pressure warning threshold is reduced to 80% of the standard value to provide better grip.

[0029] In red light mode, when the central processing unit identifies the vehicle within the stop line of a traffic light using GPS positioning information (compared to a preset traffic light geolocation database with an accuracy of 5 meters), the sensor group enters a deep sleep state, at which point power consumption drops to 0.5mW. Simultaneously, a backup capacitor power supply circuit with a capacitance of 1F and a voltage of 5V is activated to maintain base clock operation. The raw accelerometer signal undergoes waveform matching to generate a characteristic vibration waveform. The matching algorithm uses a dynamic time warping algorithm with a similarity threshold set to 0.85. When the characteristic vibration waveform is detected, all sensors are awakened and self-tested within 100 milliseconds, ensuring that the system can immediately resume normal operation when the vehicle starts.

[0030] An intelligent tire pressure monitoring system also includes an energy management subsystem. In driving mode, a micro-generator driven by tire rotation charges the sensor battery. The micro-generator has a diameter of 15mm and a height of 8mm. It is mounted on the inner sidewall of the tire and uses the centrifugal force generated by the rotation of the tire to drive the generator rotor. The generator outputs a voltage of 0.8-3.2V AC, which is rectified and stabilized to generate a 5V DC charging current with a maximum charging current of 50mA. In parking mode, the system switches to the main power supply of the central processing unit and obtains power through the vehicle's CAN bus. The power management chip monitors the supercapacitor voltage in real time. When the voltage falls below the 3.2V threshold, it automatically shuts down the communication functions of non-core sensors, retaining only the basic monitoring functions of the pressure sensor to extend battery life.

[0031] The system's tire condition assessment model utilizes a multi-layer perceptron neural network architecture. The input layer contains 16 neurons, corresponding to parameters such as pressure, temperature, and vibration spectrum characteristics. The hidden layer consists of two layers, each with 32 neurons, using Reluctant Unit (ReLU) activation function. The output layer consists of a single neuron, representing the tire pressure health score (0-100). The model is trained using extensive historical data, including tire pressure variations under various road conditions, load conditions, and temperature environments. Training utilizes the Adam optimizer with a learning rate of 0.001, a batch size of 64, and 200 training epochs.

[0032] An intelligent tire pressure monitoring system also includes a cloud-based collaborative diagnosis module. The central processing unit uploads the key parameters output by the tire condition assessment model to the cloud server through the vehicle-mounted T-BOX. The upload frequency is once an hour, and the data packet size is approximately 2KB. After receiving the data, the cloud performs feature extraction processing to generate a fault feature vector. The feature dimension is 24 dimensions, including pressure fluctuation patterns, temperature change trends, vibration spectrum characteristics, etc. The cloud server uses a decision tree algorithm based on meteorological data and historical fault records to generate preventive maintenance recommendations. The decision tree depth is 5 layers, and the feature importance threshold is 0.05. When the detected tire pressure abnormality pattern matches the typical fault case stored in the cloud and the match degree exceeds the 0.8 threshold, a visual maintenance guide is pushed to the vehicle display screen, including the fault type, severity, recommended measures and information about nearby repair points.

[0033] In some embodiments, the formula for calculating tire pressure health includes:

[0034] in: For tire pressure health, The tire pressure value obtained by the i-th sampling is derived from the real-time data of the pressure sensor; The average tire pressure value during the statistical period; The tire temperature value obtained by the j-th sampling is derived from the real-time data of the temperature sensor; Preset value for optimal working temperature; is the spectrum entropy value of the kth vibration sensor, which comes from the wavelet analysis result of the vibration spectrum; n, m, and p are the sampling times of pressure, temperature, and vibration data, respectively.

[0035] Tire pressure health can refer to a quantitative indicator that comprehensively assesses tire condition, used to characterize the degree of deviation between the tire's current operating state and ideal operating conditions. The tire pressure value obtained at the i-th sampling time can refer to the instantaneous pressure data collected by the pressure sensor at a specific time point, reflecting the real-time fluctuations in the tire's internal pressure. The real-time data from the pressure sensor can refer to the pressure measurement signal continuously output by the embedded sensor device, providing a temporal characteristic of the tire's internal pressure. The average tire pressure value within a statistical period can refer to the arithmetic mean of multiple sampling data within a set time window, used to eliminate the impact of instantaneous fluctuations on the evaluation results. The tire temperature value obtained at the j-th sampling time can refer to the tire's internal temperature reading collected by the temperature sensor at a preset frequency, used to monitor changes in the tire's thermodynamic state during operation. The preset optimal operating temperature value can refer to the optimal operating temperature range threshold set based on the tire's material properties, used to determine whether the current temperature is within a safe range. The spectral entropy value can refer to the frequency domain features of the vibration signal extracted through wavelet transform, which can characterize the energy distribution complexity of the tire's ground contact vibration. The wavelet analysis results of the vibration signal can refer to the time-frequency domain features obtained after multi-scale decomposition of the accelerometer signal, which are used to identify vibration modes under different road excitations. n can refer to the pressure data sampling frequency setting, which controls the sample size of the pressure data within the statistical window. m can refer to the temperature data sampling frequency setting, which adjusts the time resolution of the temperature assessment. p can refer to the vibration data sampling frequency setting, which determines the sample size for spectral analysis.

[0036] By integrating and processing multi-source sensor data, a three-dimensional tire condition assessment system has been constructed. This system not only monitors basic air pressure parameters in real time, but also senses the internal temperature distribution and external vibration characteristics of the tire. An innovative health calculation model effectively integrates three key indicators: pressure fluctuation, temperature deviation, and vibration spectrum, ensuring that the assessment results reflect both static parameters and dynamic characteristics. A dynamic threshold adjustment mechanism adapts to different operating conditions and significantly improves early warning accuracy. This overall solution represents a technological leap from single air pressure monitoring to comprehensive tire condition diagnosis, providing more comprehensive safety assurance for intelligent tire systems.

[0037] In some embodiments, the formula for calculating the spectrum entropy value includes:

[0038] in: is the spectrum entropy value, is the normalized energy spectral density at frequency f, which is derived from the FFT transformation result of the vibration sensor; is the wavelet packet energy of the bth frequency band, which comes from the three-level wavelet packet decomposition; is the maximum energy value of the entire frequency band; b is a natural number, representing different frequency band division intervals, and the frequency band division intervals include: 20-50Hz, 50-80Hz, 80-120Hz, 120-160Hz and 160-200Hz.

[0039] The normalized energy spectral density at frequency f can refer to the ratio of the energy of each frequency component to the total energy after the FFT transform, which can reflect the relative strength of the vibration energy in different frequency bands. The FFT transform result of a vibration sensor can refer to the frequency domain representation obtained after the accelerometer signal is subjected to the Fast Fourier Transform, which provides information on the frequency component distribution of the vibration signal. Wavelet packet energy can refer to the signal energy value of a specific sub-band after three-level wavelet packet decomposition, which is used to identify the vibration mode characteristics of different frequency bands. Three-level wavelet packet decomposition can refer to a time-frequency analysis method that decomposes the signal layer by layer into eight sub-bands, which can extract multi-scale vibration characteristics. The maximum energy value of the entire frequency band can refer to the peak value of the energy amplitude in all frequency bands, which is used to normalize the energy value of each frequency band. The frequency band division interval can refer to the manually set frequency range division standard, which can make the spectrum analysis more targeted and interpretable.

[0040] Through a combined time-frequency domain analysis approach, a multi-dimensional extraction and quantitative characterization of tire vibration characteristics is achieved. This innovative frequency band partitioning strategy balances engineering practicality with theoretical robustness, enabling the system to capture typical vibration patterns within specific frequency bands while also assessing overall vibration complexity through entropy metrics. This multi-scale analysis framework effectively integrates the global frequency domain perspective of the FFT with the local time-frequency characteristics of wavelet packets, providing a richer dimension of condition monitoring for intelligent tire systems.

[0041] The system also incorporates a failsafe mechanism. If communication with any sensor group is interrupted, the central processing unit reconstructs the missing data based on data from the remaining functioning sensors and the vehicle dynamics model. This reconstruction algorithm, based on correlation analysis between the vehicle dynamics model and historical data, achieves accuracy exceeding 85% of the original data. The system uses a Kalman filter to process the vibration data of adjacent tires to generate correlation coefficients. The filter parameters, Q matrix, are a diagonal matrix of [0.01, 0.01, 0.01], and R matrix, a diagonal matrix of [0.1, 0.1, 0.1]. When reconstructing missing data, the vibration correlation coefficients of adjacent tires are weighted. The correlation coefficients are calculated using an LSTM network trained on historical driving data. The LSTM network consists of 64 hidden units, a time step of 10, and an input feature dimension of 12. Training utilizes a stochastic gradient descent optimizer with a learning rate of 0.005.

[0042] The beneficial effects of one of the embodiments of this specification include at least the following: the system utilizes a collaborative architecture of distributed sensors and an intelligent central processing unit. By integrating pressure, temperature, and vibration parameters, the system overcomes the limitations of traditional TPMS, which only monitors static tire pressure, and achieves a comprehensive assessment of tire health. An innovative dynamic operating mode switching mechanism automatically adjusts the detection strategy based on the vehicle's status, reducing parking mode energy consumption by over 80% while maintaining monitoring accuracy. A tire condition assessment model built based on machine learning analyzes vibration spectrum characteristics in real time, enabling the first self-identification of road surface types and dynamic adjustment of alarm thresholds, reducing false alarm rates by over 60%. The system's unique red light mode rapid wake-up technology reduces response latency to under 100 milliseconds, ensuring continuous monitoring in urban conditions. Compared to traditional solutions, this system achieves breakthrough improvements in three key metrics: monitoring dimension, energy efficiency, and warning accuracy.

[0043] The preferred embodiments disclosed above are intended only to help illustrate this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of the present invention. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent tire pressure monitoring system, characterized in that: It includes a distributed sensor group and a central processing unit installed in the tire; The sensor group includes a pressure sensor, a temperature sensor and a three-axis accelerometer, which are used to collect real-time pressure values, temperature values ​​and vibration spectrum of the tire; The central processing unit switches working modes according to the vehicle status, including periodic low-power detection in parking mode, dynamic high-frequency detection in driving mode, and short-term sleep detection in red light mode; The system calculates the current tire pressure health through a trained tire condition assessment model, wherein the tire condition assessment model is obtained by fusion training based on the collected data of the sensor group, and the tire condition assessment model dynamically adjusts the alarm threshold according to the road surface characteristics.

2. The system according to claim 1, wherein: The vehicle status includes parking mode, including: When the vehicle speed signal obtained by the central processing unit through the CAN bus is zero, controlling the sensor group to enter an intermittent working state of being activated once every hour; If the pressure drop rate is detected to exceed the preset threshold, the detection frequency will be automatically increased; The pressure sensor collects the original analog signal and performs AD conversion to generate a digital pressure value; The temperature sensor collects ambient temperature data and performs linear compensation to generate a compensated temperature value; All detection data are transmitted to the central processing unit via the LoRa wireless communication module.

3. The system according to claim 1, wherein: The vehicle status includes the driving mode, including: The triaxial accelerometer monitors the tire ground contact deformation characteristics in real time, and the original accelerometer signal is processed by Kalman filtering to generate denoised vibration data; Combined with the yaw rate data from the vehicle's ESP system, the actual tire load distribution is calculated using the least squares method; The central processing unit dynamically adjusts the sampling frequency of each tire according to the load distribution results, and the sampling frequency adjustment range is 10-50Hz; The vibration sensor collects the time domain signal and performs FFT transformation to generate the frequency domain energy spectrum.

4. The system according to claim 3, characterized in that The road feature recognition process specifically includes: Performing wavelet packet decomposition on the original signal collected by the vibration sensor, extracting the energy value of the 30-50 Hz frequency band as the ice and snow road feature, and the energy value of the 80-120 Hz frequency band as the gravel road feature; The energy values ​​of each frequency band are compared with the benchmark model established by historical data, and the sliding average algorithm is used to generate the estimated road adhesion coefficient; The central processing unit proportionally modifies the standard tire pressure threshold according to the road adhesion coefficient estimate.

5. The system according to claim 1, wherein: The red light mode specifically includes: When the central processing unit identifies that the vehicle is within the stop line of a traffic light through GPS positioning information, the central processing unit controls the sensor group to enter a deep sleep state; At the same time, the backup capacitor power supply circuit is started to maintain the basic clock operation; the accelerometer original signal is processed by waveform matching to generate a characteristic vibration waveform; When a characteristic vibration waveform is detected, all sensors are woken up and self-tested within 100 milliseconds.

6. The system according to claim 1, wherein: It also includes an energy management subsystem, specifically including: In the driving mode, the sensor battery is charged by a micro-generator driven by tire rotation; The generator output voltage is rectified and stabilized to generate charging current; Switching to a main power supply to the central processing unit in the parking mode; The power management chip monitors the supercapacitor voltage in real time and automatically shuts down the communication function of non-core sensors when the voltage is lower than the threshold.

7. The system according to claim 1, wherein: The formula for calculating tire pressure health includes: in: For tire pressure health, The tire pressure value obtained by the i-th sampling is derived from the real-time data of the pressure sensor; The average tire pressure value during the statistical period; The tire temperature value obtained by the j-th sampling is derived from the real-time data of the temperature sensor; Preset value for optimal working temperature; is the spectrum entropy value of the kth vibration sensor, which comes from the wavelet analysis result of the vibration spectrum; n, m, and p are the sampling times of pressure, temperature, and vibration data, respectively.

8. The system according to claim 7, characterized in that The calculation formula for calculating the spectrum entropy value includes: in: is the spectrum entropy value, is the normalized energy spectral density at frequency f, which is derived from the FFT transformation result of the vibration sensor; is the wavelet packet energy of the bth frequency band, which comes from the three-level wavelet packet decomposition; is the maximum energy value of the entire frequency band; b is a natural number, representing different frequency band division intervals, and the frequency band division intervals include: 20-50Hz, 50-80Hz, 80-120Hz, 120-160Hz and 160-200Hz.

9. The system according to claim 1, wherein: It also includes cloud-based collaborative diagnosis modules, including: The central processing unit uploads the key parameters output by the tire condition assessment model to the cloud server via the vehicle-mounted T-BOX; After receiving the data, the cloud performs feature extraction and processing to generate a fault feature vector; The cloud server generates preventive maintenance recommendations using a decision tree algorithm in combination with meteorological data and historical fault records; When the detected abnormal tire pressure pattern matches the typical fault cases stored in the cloud and exceeds the threshold, visual maintenance instructions are pushed to the vehicle display screen.

10. The system according to claim 1, wherein: It also includes failure protection mechanisms, including: When communication of any of the sensor groups is interrupted, the central processing unit reconstructs the missing data based on data from the remaining normal sensors and a vehicle dynamics model; The Kalman filter algorithm is used to process the vibration data of adjacent tires to generate correlation coefficients; When reconstructing missing data, the vibration correlation coefficient of adjacent tires is used as a weight factor. The correlation coefficient is calculated through an LSTM network trained with historical driving data.

Images