Real-time diagnosis method and system for vehicle faults based on tbox

Abstract

The application discloses a vehicle fault real-time diagnosis method and system based on TBOX, relates to the technical field of fault diagnosis, and comprises the following steps: TBOX adaptively switches a sampling mode according to a calculation load and a bus communication state to obtain a vehicle-mounted bus link characteristic flow; the reflection wave characteristics of the vehicle-mounted bus link characteristic flow are analyzed, and the communication link impedance state is calculated; signal correction and phase compensation are performed; whole-vehicle high-fidelity logic signals and communication link reliability indexes are generated; the whole-vehicle high-fidelity logic signals are mapped into discrete sequences reflecting signal evolution rules; the distribution complexity characteristics of the discrete sequences are quantified; the running state characteristic vector is generated in combination with the communication link reliability indexes; the running state characteristic vector is matched with a preset fault tree model for reasoning; the fault type is determined; and a vehicle health state diagnosis conclusion is output. The application improves the sensitivity and accuracy of real-time diagnosis of the sub-health state of the vehicle.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and in particular to a method and system for real-time vehicle fault diagnosis based on TBOX. Background Technology

[0002] With the deep integration of in-vehicle network technology and telematics units (TBOX), vehicle remote diagnostics technology has evolved from early passive alarms to predictive maintenance based on big data analytics. Current technologies typically use the TBOX to collect application-layer diagnostic fault codes (DTCs) and data stream parameters from the controller area network (CAN) or high-speed bus in real time, and then utilize cloud-based expert systems or machine learning algorithms to assess the vehicle's health status. This approach focuses on parsing the logic protocol layer, identifying functional faults by monitoring periodic anomalies in messages, control deviations, or sensor value overruns, and is currently the mainstream method in the field of in-vehicle remote diagnostics.

[0003] However, existing technologies have significant limitations in handling intermittent faults caused by degradation of the vehicle's physical links or latent damage to hardware entities. Specifically, traditional protocol-layer-based diagnostic mechanisms often ignore the evolution of electrical characteristics at the vehicle bus physical layer, making it difficult to effectively identify microscopic signal distortions caused by wiring harness wear, connector oxidation, impedance mismatch, or drift in the electrical characteristics of hardware interfaces. When the integrity of the physical layer is compromised but has not yet triggered a check error (such as a CRC error) at the layout protocol layer, the system often cannot perceive the underlying physical risks, leading to missed or false alarms for deep hardware-level faults such as intermittent short circuits in sensors or performance degradation of actuators, making it difficult to achieve high-fidelity condition assessment throughout the entire lifecycle. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a real-time vehicle fault diagnosis method based on TBOX to solve the problem of insufficient accuracy in diagnosing hidden damage to the underlying hardware caused by the difficulty in perceiving the evolution of electrical features at the physical layer.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a real-time vehicle fault diagnosis method based on TBOX, comprising:

[0008] TBOX adaptively switches sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status.

[0009] The reflected wave characteristics of the vehicle bus link are analyzed and the impedance state of the communication link is calculated. Signal correction and phase compensation are performed to generate high-fidelity logic signals for the whole vehicle and reliability indicators of the communication link.

[0010] The high-fidelity logic signals of the whole vehicle are mapped into discrete sequences that reflect the evolution of the signals, and the distribution complexity characteristics of the discrete sequences are quantified. Combined with the reliability index of the communication link, an operating status feature vector is generated.

[0011] The system matches and infers the operating status feature vector with the preset fault tree model to determine the fault type and outputs the vehicle health status diagnosis conclusion.

[0012] Based on the vehicle health status diagnosis results, secure evidence storage and remote synchronization operations are performed, and online calibration is performed according to the diagnostic confidence level to adjust the physical layer topology reference parameters.

[0013] Preferably, the method for obtaining the vehicle bus link feature flow includes:

[0014] Compare the available computing load of TBOX with the preset resource threshold, and the vehicle bus communication load with the preset communication threshold; when the available computing load is greater than the preset resource threshold and the vehicle bus communication load is greater than the preset communication threshold, switch the physical layer signal acquisition mode to the full waveform capture mode; otherwise, switch to the edge feature extraction mode based on the hardware pulse capture circuit.

[0015] In a defined physical layer signal acquisition mode, in response to the start bit edge of a message or a hardware abnormal interrupt pulse, a sampling trigger window is opened to achieve synchronous alignment between the signal edge and the sampling clock.

[0016] Within the sampling trigger window, logical frame data and physical link feature data are extracted synchronously, and the sampling points of multiple communication cycles are recombined based on equivalent sampling to output the vehicle bus link feature stream.

[0017] Preferably, the method for analyzing the reflected wave characteristics of the vehicle bus link characteristic flow includes:

[0018] Identify the instantaneous location in the characteristic flow of the vehicle bus link where the voltage amplitude causes a level polarity reversal, and locate the reference moment of the logic level transition edge;

[0019] Starting from the reference time, based on the signal propagation delay corresponding to the physical topology of the vehicle bus, a reflection wave observation time window covering the potential reflection signal area is defined backward; and the difference between the maximum fluctuation point of voltage amplitude deviating from the steady-state level and the steady-state level is extracted as the peak amplitude of the reflection wave.

[0020] Identify the peak time of the reflected wave corresponding to the point of maximum fluctuation within the observation time window; measure the time difference between the peak time of the reflected wave and the reference time as the reflection time delay.

[0021] Preferably, the method for estimating the impedance state of the communication link includes:

[0022] The reflection coefficient is determined based on the ratio of the peak amplitude of the reflected wave to the standard amplitude of the incident wave, and the positive or negative sign information representing the change in impedance polarity is determined based on the offset direction of the reflected wave relative to the steady-state level.

[0023] The physical properties of impedance mismatch are determined based on the positive and negative sign information. When the positive and negative sign information is positive, a high impedance mismatch point is determined to exist, and when the positive and negative sign information is negative, a low impedance mismatch point is determined to exist.

[0024] By utilizing the reflection time delay and the signal propagation rate of the vehicle bus, the physical location of the impedance mismatch is mapped; by associating the physical properties and location of the impedance mismatch, the communication link impedance state associated with the electrical characteristics of the vehicle hardware interface is output.

[0025] Preferably, the method for performing signal correction and phase compensation includes:

[0026] Based on the impedance state of the communication link, a reflection cancellation compensation sequence is generated and superimposed onto the characteristic flow of the vehicle bus link to compensate for signal distortion caused by impedance mismatch.

[0027] After detecting the bit edge timing offset of the characteristic stream of the vehicle bus link after signal correction, the trigger phase of the sampling clock is adjusted to align with the valid data determination interval, thus completing phase compensation.

[0028] The system performs logic level determination on the characteristic flow of the vehicle bus link after signal correction and phase compensation, outputs a high-fidelity logic signal for the whole vehicle, and generates a communication link reliability index based on the waveform characteristics during the compensation process.

[0029] Preferably, the method for mapping the high-fidelity logic signals of the entire vehicle into discrete sequences reflecting the signal evolution law includes:

[0030] The high-fidelity logic signal of the whole vehicle is sampled and extracted according to the sampling clock frequency to obtain the original bit value sequence; the flip points of adjacent values ​​are captured, and the edge evolution features of the logic state transition position are extracted.

[0031] Based on the pulse width reflected by the edge evolution characteristics, the original bit value sequence is converted into symbolic code elements representing different signal evolution states; these are then arranged according to temporal relationships to generate discrete sequences.

[0032] Preferably, the method for generating the running state feature vector includes:

[0033] The frequency of occurrence of each symbol in a discrete sequence is statistically analyzed to determine the symbol probability distribution, which represents the probability distribution of the signal evolution state. The information entropy value of the discrete sequence is calculated using the symbol probability distribution to quantify the distribution complexity characteristics of the signal distortion disorder.

[0034] The distribution complexity characteristics are normalized and aligned with the communication link reliability indicators to form fused data of both physical link and signal logic states; the fused data is then vectorized and combined according to preset dimensions to generate operational state feature vectors.

[0035] Preferably, the method for outputting the vehicle health status diagnostic conclusion includes:

[0036] The running state feature vector is mapped to the bottom node of the preset fault tree model as the initial evidence for logical reasoning. Based on the comparison between the running state feature vector and the threshold of the preset fault tree model, the logic gate instruction in the preset fault tree model is activated to perform matching reasoning using the initial evidence for logical reasoning.

[0037] Track the activated logical paths in the preset fault tree model to identify the fault modes and risk levels associated with the paths and determine the fault type.

[0038] By associating fault types with the reasoning reliability scores generated during the matching reasoning process, the diagnostic confidence level is determined, and the vehicle health status diagnostic conclusion is output.

[0039] Preferably, the method for performing online calibration based on diagnostic confidence to adjust physical layer topology reference parameters includes:

[0040] The vehicle health status diagnosis results are encrypted and stored locally and pushed synchronously to the cloud; when the diagnostic confidence reaches the preset trigger threshold, the online calibration trigger condition is determined to be met.

[0041] Based on the fault type in the vehicle health status diagnosis conclusion, the trigger phase of the sampling clock is corrected to align with the effective sampling range of the physical layer signal, thereby achieving online calibration of the physical layer topology reference parameters.

[0042] Secondly, the present invention provides a real-time vehicle fault diagnosis system based on TBOX, comprising:

[0043] The capture module is used by TBOX to adaptively switch sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status.

[0044] The correction module is used to analyze the reflected wave characteristics of the vehicle bus link characteristic flow and calculate the impedance state of the communication link, perform signal correction and phase compensation, and generate high-fidelity logic signals and communication link reliability indicators for the whole vehicle.

[0045] The quantization module is used to map the high-fidelity logic signals of the whole vehicle into discrete sequences that reflect the evolution of the signals, quantify the distribution complexity characteristics of the discrete sequences, and generate operating status feature vectors by combining communication link reliability indicators.

[0046] The judgment module is used to match and reason with the operating status feature vector and the preset fault tree model to determine the fault type and output the vehicle health status diagnosis conclusion.

[0047] The evidence storage module is used to perform secure evidence storage and remote synchronization operations based on the vehicle health status diagnosis results, and to perform online calibration to adjust the physical layer topology reference parameters according to the diagnostic confidence level.

[0048] The beneficial effects of this invention are as follows: By reconstructing the link impedance based on the time delay and amplitude characteristics of reflected waves, and superimposing a compensation sequence with opposite phases to cancel signal distortion, and by using symbol dynamics to map the logic bit stream into an information entropy feature vector that reflects the law of signal evolution, the invention achieves accurate localization of microscopic electrical damage in the bus physical layer and quantitative characterization of the degree of signal evolution disorder. This not only eliminates communication errors caused by the degradation of physical link quality, but also deeply mines the hardware load interference characteristics hidden behind the waveform. It can be used to identify hardware interface faults or performance degradation precursors of core entities such as engines and chassis. Without relying on traditional diagnostic codes, it can discover and characterize hidden damage to vehicle hardware through physical layer fingerprints, improving the sensitivity and accuracy of real-time diagnosis of vehicle sub-health status. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of the real-time vehicle fault diagnosis method based on TBOX in this invention;

[0051] Figure 2 This is a flowchart of the output vehicle bus link feature flow in this invention;

[0052] Figure 3 This is a flowchart illustrating the output communication link impedance status in this invention;

[0053] Figure 4 This is a schematic diagram of the vehicle fault real-time diagnosis system based on TBOX in this invention. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0057] Reference Figure 1 , Figure 2 , Figure 3 and Figure 4 This is one embodiment of the present invention, which provides a real-time vehicle fault diagnosis method based on TBOX, including the following steps:

[0058] TBOX adaptively switches sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status.

[0059] The reflected wave characteristics of the vehicle bus link are analyzed and the impedance state of the communication link is calculated. Signal correction and phase compensation are performed to generate high-fidelity logic signals for the whole vehicle and reliability indicators of the communication link.

[0060] The high-fidelity logic signals of the whole vehicle are mapped into discrete sequences that reflect the evolution of the signals, and the distribution complexity characteristics of the discrete sequences are quantified. Combined with the reliability index of the communication link, an operating status feature vector is generated.

[0061] The system matches and infers the operating status feature vector with the preset fault tree model to determine the fault type and outputs the vehicle health status diagnosis conclusion.

[0062] Based on the vehicle health status diagnosis results, secure evidence storage and remote synchronization operations are performed, and online calibration is performed according to the diagnostic confidence level to adjust the physical layer topology reference parameters.

[0063] Methods for obtaining the characteristic flow of the vehicle bus link include:

[0064] Compare the available computing load of TBOX with the preset resource threshold, and the vehicle bus communication load with the preset communication threshold; when the available computing load is greater than the preset resource threshold and the vehicle bus communication load is greater than the preset communication threshold, switch the physical layer signal acquisition mode to the full waveform capture mode; otherwise, switch to the edge feature extraction mode based on the hardware pulse capture circuit.

[0065] It should be noted that the internal registers of the vehicle terminal are read in real time to obtain the available computing load of the remaining processing capacity, and the message bit flow per unit time is counted by the communication controller to determine the vehicle bus communication load.

[0066] In full waveform capture mode, the high-speed analog-to-digital converter inside the vehicle terminal controls the bus differential voltage at a sampling frequency much higher than the bus bit rate, for example, 100 million samples per second, converting the continuous analog voltage signal into a discrete voltage value sequence with high time resolution, and fully recording the full projection data of voltage, including voltage overshoot, ringing fluctuations and the fine shape of signal rising edge.

[0067] In the edge feature extraction mode based on hardware pulse capture circuit, the vehicle terminal only activates the capture unit of the internal timer, such as calling the sixteen-bit capture register with input capture function in the general timer, and identifies the instantaneous switching of the signal level through a voltage comparator with a reference voltage set to 0.5V integrated in the hardware circuit.

[0068] When the bus differential voltage crosses the half dominant level position of 0.5V, the voltage comparator outputs a high-level pulse to trigger the 16-bit capture register to instantly latch the current general-purpose timer count value, for example, latch the current hexadecimal count value 0x4A2C, thereby extracting only the timestamp data representing the moment the signal edge occurs and the time difference between adjacent edges, characterizing the time-domain jitter of the signal with extremely low data throughput;

[0069] It should be noted that the safe level before the CPU utilization rate reaches the congestion threshold is set as a preset resource threshold, with an example value of 30% of the remaining available computing resources. At the same time, based on the theoretical bus bandwidth and the risk value of internal buffer overflow, the maximum traffic limit that triggers high-precision sampling without packet loss is set as a preset communication threshold, with an example value of 65% of the packet bit traffic per unit time.

[0070] Under a defined physical layer signal acquisition mode, in response to the start bit edge of a message or a hardware abnormal interrupt pulse, a sampling trigger window is opened to achieve synchronous alignment between the signal edge and the sampling clock.

[0071] Specifically, in full waveform capture mode or edge feature extraction mode, the hardware detection circuit monitors the bus level status in real time. When it recognizes the falling edge signal of the start bit representing the start of a frame message or receives a hardware abnormal interrupt pulse due to abnormal voltage fluctuation, the sampling controller inside the vehicle terminal immediately starts a sampling trigger window whose duration covers the entire bit period.

[0072] The sampling controller uses the message start bit edge or hardware abnormal interrupt pulse as a reference signal. It extracts the phase difference between the reference signal and the internal sampling clock signal in real time through the feedback control circuit, and adjusts the control voltage linearly according to the phase difference to change the output frequency of the internal voltage-controlled oscillator.

[0073] By continuously fine-tuning the frequency and compensating for the phase, the phase of the internal sampling clock is locked at a specified offset position of the reference signal, so that the start time of the sampling trigger window is aligned with the edge of the message start bit or the occurrence time of the hardware abnormal interrupt pulse.

[0074] Within the sampling trigger window, logical frame data and physical link feature data are extracted synchronously, and the sampling points of multiple communication cycles are recombined based on equivalent sampling to output the vehicle bus link feature stream.

[0075] Specifically, during the operation of the open sampling trigger window, the vehicle terminal uses two independent channels of direct memory access technology to read logic frame data reflecting the logic level state from the protocol processor at the same clock cycle, such as the binary bit stream of message identifier, data length code and cyclic redundancy check sequence; and reads physical link characteristic data reflecting the waveform electrical characteristics from the high-speed analog-to-digital converter or capture unit, such as the electrical parameter sequence of instantaneous differential voltage amplitude, signal rise edge slope distribution and bit time offset.

[0076] For high-speed communication environments where the sampling frequency is lower than the signal change frequency, the sampling controller changes the phase offset of the sampling trigger window relative to the signal start position in multiple consecutive communication cycles. It extracts discrete sample points at different phase points in each communication cycle, rearranges and merges the points according to the sampling time axis to complete phase reorganization, and combines the reorganized waveform samples with logic information to output the vehicle bus link feature stream.

[0077] Traditional vehicle fault diagnosis primarily relies on diagnostic fault codes (DTCs) generated from application layer data or protocol layer. Its core logic lies in monitoring the "result"—that is, the system can only detect a fault when damage to the physical link has led to communication interruption, data errors (such as CRC errors), or significantly abnormal sensor values. This approach suffers from severe lag and cannot accurately locate hidden physical damage such as wear, loose connectors, or oxidation of the vehicle's bus harness, often requiring manual, blind inspection of the entire vehicle's wiring harness during repairs. Furthermore, traditional solutions struggle to distinguish between instantaneous fluctuations caused by electromagnetic interference and actual hardware performance degradation. To achieve predictive maintenance and obtain the specific physical coordinates of the fault before it evolves into a communication system collapse, this solution designs a physical link depth analysis method based on reflected wave characteristics, as follows:

[0078] Methods for analyzing the reflected wave characteristics of the characteristic flow of the vehicle bus link include:

[0079] Identify the instantaneous location in the characteristic flow of the vehicle bus link where the voltage amplitude causes a level polarity reversal, and locate the reference moment of the logic level transition edge.

[0080] It should be noted that the analysis of the reflected wave characteristics of the vehicle bus link characteristic flow involves traversing the electrical parameter sequence in the vehicle bus link characteristic flow, performing continuous subtraction on the instantaneous differential voltage amplitude between adjacent sampling points to obtain the voltage change rate value over time, and retrieving the moment when the change rate value reaches the extreme point to identify the instantaneous position where the voltage amplitude generates a level polarity reversal. The time when the voltage crosses the exemplary 0.5V intersection point is confirmed as the reference moment for locating the logic level transition edge.

[0081] Starting from the reference time, based on the signal propagation delay corresponding to the physical topology of the vehicle bus, a reflection wave observation time window covering the potential reflection signal area is defined backward; and the difference between the maximum fluctuation point of voltage amplitude deviating from the steady-state level and the steady-state level is extracted as the peak amplitude of the reflection wave.

[0082] Specifically, starting from the reference moment of the positioning logic level transition edge, the signal propagation delay corresponding to the vehicle bus physical topology that matches the actual wiring length is retrieved from the non-volatile memory of the vehicle terminal. The signal propagation delay is then widened by multiplying it by a preset expansion coefficient (an example value of 2.2, which is obtained by combining the signal propagation delay with the time requirement for the signal to travel back and forth in the actual wiring length, which is 2, plus the transmission delay fluctuation margin caused by the node connectors in the vehicle bus physical topology, which is 0.2). This is used to define a reflection wave observation time window that covers the potential reflection signal area.

[0083] Within the observation window of the reflected wave, the deviation between the instantaneous differential voltage amplitude and the steady-state level representing the interference-free transmission state is continuously monitored. The maximum fluctuation point of the voltage amplitude deviating from the steady-state level is captured, and the absolute value of the difference between the voltage value at the maximum fluctuation point and the steady-state level is calculated. This absolute value is recorded as the peak amplitude of the reflected wave.

[0084] Identify the peak time of the reflected wave corresponding to the point of maximum fluctuation within the observation time window; measure the time difference between the peak time of the reflected wave and the reference time as the reflection time delay.

[0085] It should be noted that the peak time of the reflected wave corresponding to the maximum fluctuation point is retrieved in real time within the observation time window of the reflected wave; the time axis distance between the peak time of the reflected wave and the reference time of the positioning logic level transition edge is measured, and the time difference between the two is calculated. The time difference is confirmed as the reflection time delay that reflects the characteristics of the signal's round-trip transmission path in the physical link.

[0086] Methods for estimating the impedance state of a communication link include:

[0087] The reflection coefficient is determined based on the ratio of the peak amplitude of the reflected wave to the standard amplitude of the incident wave, and the positive and negative sign information representing the change in impedance polarity is determined based on the offset direction of the reflected wave relative to the steady-state level.

[0088] Specifically, the reflection coefficient is determined by retrieving the standard amplitude of the incident wave preset in the non-volatile memory (for example, a value of 2V) and dividing the recorded peak amplitude of the reflected wave by the standard amplitude of the incident wave.

[0089] Simultaneously, the offset direction of the voltage value at the maximum fluctuation point relative to the steady-state level is detected in real time. If the voltage value at the maximum fluctuation point is higher than the steady-state level, the positive polarity is determined; if the voltage value at the maximum fluctuation point is lower than the steady-state level, the negative polarity is determined. This generates positive and negative sign information representing the change in impedance polarity.

[0090] The physical properties of impedance mismatch are determined based on the positive and negative sign information. When the positive and negative sign information is positive, a high impedance mismatch point is determined to exist, and when the positive and negative sign information is negative, a low impedance mismatch point is determined to exist.

[0091] Specifically, the physical properties of impedance mismatch are determined based on the positive and negative sign information. When the positive sign information is positive, it indicates that a load higher than the characteristic impedance of the transmission line has been encountered in the transmission path. At this time, a high impedance mismatch point is determined to exist, which usually corresponds to a wire harness open circuit or poor connector contact. When the positive and negative sign information is negative, it indicates that a load lower than the characteristic impedance of the transmission line has been encountered in the transmission path. At this time, a low impedance mismatch point is determined to exist, which usually corresponds to a wire harness short circuit or insulation failure.

[0092] By utilizing the reflection time delay and the signal propagation rate of the vehicle bus, the physical location of the impedance mismatch is mapped; by associating the physical properties and location of the impedance mismatch, the communication link impedance state associated with the electrical characteristics of the vehicle hardware interface is output.

[0093] Specifically, by using the reflection time delay and the signal propagation rate of the vehicle bus, such as 0.2 meters per nanosecond, a multiplication operation is performed and divided by 2 to cancel the signal round-trip path, thereby mapping the physical location of the impedance mismatch.

[0094] Logically associate the physical properties of impedance mismatch with the physical location of impedance mismatch. The specific process is as follows: retrieve the vehicle topology layout diagram containing the coordinates of the harness branch nodes and the physical spacing of the connectors, and perform addressing and comparison of the physical location of impedance mismatch on the time-distance mapping axis of the vehicle topology layout diagram.

[0095] If the physical location of the impedance mismatch coincides with the installation location of a specific connector, for example, if the physical location of the impedance mismatch is 3.5 meters away from the vehicle terminal, and the vehicle topology layout diagram indicates that there is a gateway connector with the number CON_03 at that location, then an index mapping is established between that location and the "high impedance mismatch point" or "low impedance mismatch point"; by matching the corresponding pin contact resistance or insulation resistance to ground of the connector and other related vehicle hardware interface electrical characteristics, the final communication link impedance status is output.

[0096] This solution achieves a technological leap from "message logic diagnosis" to "link physical diagnosis" through joint diagnosis based on reflected wave characteristics and impedance status. Its local advantages lie in its ability to quantify and identify minute impedance fluctuations in the communication link and to locate fault points at the centimeter level using reflection time delay. Compared to traditional diagnostics, this solution can not only distinguish the physical nature of high impedance (open circuit, poor contact) and low impedance (short circuit, insulation failure), but also accurately pinpoint the specific connector or pin where the fault occurs without disassembling the wiring harness. This real-time mapping mechanism based on electrical characteristics greatly improves the accuracy of identifying the "sub-healthy" state of vehicles, providing the vehicle terminal with an intuitive view from electrical signals to the underlying hardware physical state, thereby significantly reducing the difficulty of fault diagnosis and subsequent maintenance costs.

[0097] Methods for performing signal correction and phase compensation include:

[0098] Based on the impedance state of the communication link, a reflection cancellation compensation sequence is generated and superimposed onto the characteristic flow of the vehicle bus link to compensate for signal distortion caused by impedance mismatch.

[0099] It should be noted that by extracting the physical properties and physical locations of impedance mismatch recorded in the communication link impedance status, and combining the physical location of the impedance mismatch with the loss constant determined by the amplitude attenuation per unit length of the vehicle bus under standard impedance, the energy loss ratio is obtained as the attenuation coefficient by multiplying the loss constant by the transmission distance corresponding to the physical location of the impedance mismatch and performing a negative exponential function transformation on the product. The initial intensity of the reflected wave is obtained by multiplying the incident wave amplitude by the reflection coefficient of the impedance mismatch degree. Simultaneously, the round-trip time required for the reflected wave to return from the mismatch point to the vehicle terminal is calculated by dividing the physical location of the impedance mismatch by the aforementioned signal propagation rate of the vehicle bus, and this time delay is used as the time delay.

[0100] The sampling controller determines the compensation start point based on the calculated time delay, and multiplies the initial intensity of the reflected wave by the attenuation coefficient to obtain the corrected reflected wave amplitude. It generates a set of discrete voltage correction value sequences that correspond to the original sampling time but have opposite voltage polarities and proportional amplitudes. The discrete voltage correction value sequence is used as the reflection cancellation compensation sequence. The reflection cancellation compensation sequence is superimposed on the original voltage sampling points in the characteristic flow of the vehicle bus link in real time, and the signal distortion caused by impedance mismatch is canceled by the algebraic addition of the physical waveforms.

[0101] After detecting the bit edge timing offset of the characteristic stream of the vehicle bus link after signal correction, the trigger phase of the sampling clock is adjusted to align with the valid data determination interval, thus completing phase compensation.

[0102] Specifically, the timing offset of the bit edges of the characteristic flow of the vehicle bus link after signal correction is detected. The timing interval between the corrected logic level transition edge and the standard bit clock is compared. By recording the first time stamp of the logic level transition edge and the second time stamp of the standard bit clock transition point, and calculating the algebraic difference between the first and second time stamps, the residual value of edge jitter caused by physical link damage is obtained. The trigger phase of the sampling clock is adjusted using a feedback adjustment circuit. The compensation is oriented according to the positive or negative polarity of the residual edge jitter value: if the logic level transition edge lags behind the standard bit clock, it is determined to be a phase delay caused by link capacitance, and the sampling clock trigger edge is controlled to move synchronously with the lag; if the logic level transition edge leads the standard bit clock, it is determined to be a phase advance caused by link inductance, and the sampling clock trigger edge is controlled to move synchronously with the lead, so that the sampling point is accurately aligned with the valid data determination interval, thereby completing the phase compensation.

[0103] The system performs logic level determination on the characteristic flow of the vehicle bus link after signal correction and phase compensation, outputs a high-fidelity logic signal for the whole vehicle, and generates a communication link reliability index based on the waveform characteristics during the compensation process.

[0104] Specifically, the logic level is determined for the characteristic flow of the vehicle bus link after signal correction and phase compensation. The reconstructed voltage amplitude is compared with a preset logic threshold, which is set according to the vehicle bus protocol specification. For example, for the CAN bus, the differential voltage determination threshold is set to a certain value (such as 0.7V) in the range of 0.5V to 0.9V. Sampling points higher than the determination threshold are identified as logic 1, and sampling points lower than the determination threshold are identified as logic 0. The high-fidelity logic signal of the whole vehicle that reflects the true communication intention is output.

[0105] Simultaneously, based on the injection strength of the reflection elimination compensation sequence, the jitter range of the bit edge timing offset, and the signal-to-noise ratio gain before and after signal correction during the compensation process, a communication link reliability index reflecting the physical performance of the current physical link is quantified.

[0106] Methods for mapping high-fidelity logic signals of a vehicle into discrete sequences that reflect the evolution of signals include:

[0107] The high-fidelity logic signal of the whole vehicle is sampled and extracted according to the sampling clock frequency to obtain the original bit value sequence; the flip points of adjacent values ​​are captured, and the edge evolution features of the logic state transition position are extracted.

[0108] It should be noted that by calling the sampling clock frequency generated by the internal clock generator, the high-fidelity logic signal of the whole vehicle after the execution signal correction and phase compensation is latched at equal intervals, and the continuous time signal is converted into the original bit value sequence composed of logic 0 and logic 1.

[0109] Subsequently, the flip points of adjacent values ​​in the original bit value sequence are captured in real time through logical XOR operation, and the rising edge position from logical 0 to logical 1 and the falling edge position from logical 1 to logical 0 are recorded, thereby extracting the edge evolution characteristics of the logical state transition position.

[0110] Based on the pulse width reflected by the edge evolution characteristics, the original bit value sequence is converted into symbolic code elements representing different signal evolution states; these are then arranged according to temporal relationships to generate discrete sequences.

[0111] It should be noted that, based on the pulse width reflected by the edge evolution characteristics, the time span between adjacent flip points is calculated and normalized with the standard bit period. The original bit value sequence is then converted into symbolic symbols representing different signal evolution states. For example, a level continuity state that conforms to the standard bit period is mapped to the symbolic symbol "S" representing stable transmission, and an abnormal disturbance state with a significantly shortened or stretched pulse width is mapped to the symbolic symbol "J" representing edge jitter. Alternatively, bit combinations that conform to specific frame structure characteristics are mapped to start or end symbols. The identified symbolic symbols are arranged according to the chronological relationship of the time axis to generate a discrete sequence that can characterize the fluctuation characteristics of the physical link communication quality of the entire vehicle.

[0112] Traditional vehicle fault diagnosis often relies on monitoring data from a single sensor based on preset thresholds. This approach struggles to handle the complex nonlinear evolution of signals and random noise interference in the vehicle bus environment. When physical links exhibit latent damage or suboptimal health, traditional logical judgments often lack quantification methods for the degree of signal disturbance, failing to provide early warnings and struggling to deeply correlate microscopic distortions at the physical layer with macroscopic logical errors. This results in high false alarm rates and a limited diagnostic dimension. Therefore, this solution introduces a state quantization and feature vector fusion method based on information entropy, as detailed below:

[0113] Methods for generating runtime state feature vectors include:

[0114] The frequency of occurrence of each symbol in a discrete sequence is statistically analyzed to determine the symbol probability distribution, which represents the probability distribution of the signal evolution state. The information entropy value of the discrete sequence is calculated using the symbol probability distribution to quantify the distributional complexity characteristics of the signal distortion disorder.

[0115] Specifically, the frequency of occurrence of each symbol code element in the discrete sequence is counted. By traversing the previously generated discrete sequence, the proportion of the symbol code element "S", the symbol code element "J", and the symbol code element representing the start or end in the total length of the discrete sequence is calculated, thereby determining the symbol probability distribution representing the probability distribution of the signal evolution state.

[0116] The information entropy value of a discrete sequence is calculated using the symbolic probability distribution. Specifically, this involves performing a logarithmic weighted summation based on the probability values ​​of each term in the symbolic probability distribution using the information entropy calculation formula.

[0117] ;

[0118] in, It is the information entropy value of the discrete sequence. It represents the total number of symbolic code types contained in the discrete sequence. It is a traversal index variable for symbol code types. It is the first The probability distribution of the occurrence of a symbolic code element in the entire discrete sequence. It is a logarithmic function with base 2, exemplified by when the symbol code element probability of occurrence When it is 1 / 4, its corresponding self-information is bit;

[0119] The information entropy value calculated by summation The mapping is a specific numerical value that reflects the uncertainty of the signal during its evolution in the time domain, thereby realizing a quantitative characterization of the distribution complexity characteristics of the degree of signal distortion and disorder.

[0120] The distribution complexity characteristics are normalized and aligned with the communication link reliability indicators to form fused data of both physical link and signal logic states; the fused data is then vectorized and combined according to preset dimensions to generate operational state feature vectors.

[0121] Specifically, the distributed complexity features and communication link reliability indicators are normalized and aligned. By retrieving the communication link reliability indicators obtained from the aforementioned quantization, the min-max normalization method is used to map the values ​​of the distributed complexity features and the communication link reliability indicators to the same numerical range. For example, the values ​​are mapped to the range of 0 to 1, forming fused data that covers both the physical performance of the physical link and the evolution of signal logic. The fused data is then vectorized and combined according to a preset dimension. By treating the normalized distributed complexity features, communication link reliability indicators, reflection coefficient, and reflection time delay as independent component operators, they are sequentially mapped to the coordinate axis components of four-dimensional Euclidean space and linearly arranged according to the topological order of {distributed complexity, reliability indicator, reflection coefficient, time delay} to construct an N-dimensional numerical sequence of the real-time state of vehicle communication, generating an operating state feature vector.

[0122] Methods for outputting vehicle health status diagnostic conclusions include:

[0123] The running state feature vector is mapped to the bottom node of the preset fault tree model as the initial evidence for logical reasoning. Based on the comparison between the running state feature vector and the threshold of the preset fault tree model, the logic gate instruction in the preset fault tree model is activated to perform matching reasoning using the initial evidence for logical reasoning.

[0124] Specifically, the distribution complexity features, communication link reliability indicators, reflection coefficients, and reflection time delays in the running state feature vector are respectively filled into the corresponding leaf node input interfaces in the preset fault tree model;

[0125] Based on the comparison results of the values ​​of each component in the running status feature vector with the preset thresholds in the fault tree model, such as the reflection coefficient being greater than the example value of 0.3 or the communication link reliability index being lower than 0.6, the probability of occurrence of the underlying event is determined, and the initial logical reasoning evidence formed therefrom is used to activate the AND gate, OR gate and other logic gate instructions in the preset fault tree model to perform matching reasoning from the underlying evidence to the top-level fault event.

[0126] The pre-setting process of the fault tree model is as follows: First, deep data mining is carried out on the historical maintenance records of the entire vehicle life cycle and the failure cases of the physical layer of the vehicle bus (such as CAN / CAN FD). Using the Failure Mode and Effects Analysis (FMEA) method, "bus communication interruption", "continuous signal distortion" and "critical node offline" are defined as the top-level events of the fault tree. Then, the deductive reasoning method is used to decompose the model downwards level by level to identify the intermediate logical causes that lead to the failure of the top level. For example, signal distortion is refined into secondary events such as waveform reflection caused by impedance mismatch, common-mode interference caused by external electromagnetic coupling and edge hysteresis caused by wiring harness parasitic capacitance.

[0127] Furthermore, the causes are decomposed into underlying basic events that can be quantified in real time by TBOX, including physical layer indicators such as abnormal reflection coefficients, sudden increases in information entropy, and reflection delay shifts, thereby establishing a hierarchical mapping matrix from microscopic electrical parameters to macroscopic system faults. During the logic construction process, based on the failure rate distribution of different hardware interfaces (such as connectors and branch nodes), transfer weights and trigger thresholds are configured for each level of logic gates (such as AND and OR gates), and a topology adaptive calibration mechanism is introduced. The fault tree model is iteratively trained and validated using known typical fault samples to ensure that each activated logic path accurately points to a specific fault mode (such as connector oxidation or harness short circuit). The transfer weights of each level of logic gates on the activated logic path are multiplied together, and a weighted sum is calculated based on the occurrence probability of the underlying basic events, simultaneously calculating the inference reliability score that reflects the reliability of the diagnostic conclusion.

[0128] Track the activated logical paths in the preset fault tree model to identify the fault modes and risk levels associated with the paths and determine the fault type.

[0129] It should be noted that by tracking the activated logical paths in the preset fault tree model and monitoring the connected branches formed during the matching inference process due to the logic gate instructions meeting the triggering conditions, the fault modes and risk levels associated with the paths are identified, and the fault type is determined. For example, when the reflection time delay associated with the high impedance mismatch point and the high distribution complexity feature simultaneously meet the triggering conditions and activate a specific logical path, the fault type is determined to be a serious poor connection of the wiring harness or a loose plug, and the corresponding emergency handling risk level is determined according to the hierarchical structure of the activated path in the preset fault tree model.

[0130] By associating fault types with the reasoning reliability scores generated during the matching reasoning process, the diagnostic confidence level is determined, and the vehicle health status diagnostic conclusion is output.

[0131] It should be noted that during the logical reasoning process of the preset fault tree model, the reasoning reliability score of each logical node is calculated by combining the normalized residual value when the operating state feature vector is input. The final fault type is weighted and correlated with the reasoning reliability score, and the percentage value reflecting the current diagnostic accuracy is calculated as the diagnostic confidence. The example value is 95%. Finally, the information including fault type, physical location and diagnostic confidence is summarized and the vehicle health status diagnosis conclusion is output.

[0132] By constructing operational status feature vectors and combining them with a fault tree model, an upgrade from "simple threshold alarms" to "multi-dimensional evidence reasoning" has been achieved. Its local advantages lie in: accurately quantifying the degree of signal distortion using information entropy and integrating it with electrical indicators of the physical link (such as reflection coefficient), providing highly robust input evidence for fault diagnosis. Through logical path tracing of the fault tree, the system can not only determine complex fault modes but also provide diagnostic confidence levels with reference value, significantly improving the accuracy of identifying soft faults such as harness aging and poor contact in complex electromagnetic environments and enhancing maintenance prediction capabilities.

[0133] Methods for performing online calibration based on diagnostic confidence levels to adjust physical layer topology baseline parameters include:

[0134] The vehicle health status diagnosis results are encrypted and stored locally and simultaneously pushed to the cloud; when the diagnostic confidence level reaches the preset trigger threshold, it is determined that the online calibration trigger condition is met.

[0135] Specifically, the vehicle health status diagnosis results are encrypted and stored locally and pushed synchronously to the cloud. By calling the hardware security encryption engine built into the vehicle terminal, the vehicle health status diagnosis results, which include fault type, physical location and diagnostic confidence, are encapsulated and encrypted, and stored in local non-volatile memory.

[0136] Simultaneously, an encrypted transmission link is established with the cloud server using the vehicle's wireless communication unit to upload the encrypted vehicle health status diagnostic results in real time. When the diagnostic confidence level reaches a preset trigger threshold, the calculated diagnostic confidence level is compared with the preset trigger threshold. For example, when the diagnostic confidence level is greater than 90%, the current fault diagnosis result is deemed to have extremely high credibility, thus confirming that the online calibration trigger condition is met.

[0137] Based on the fault type in the vehicle health status diagnosis conclusion, the trigger phase of the sampling clock is corrected to align with the effective sampling range of the physical layer signal, thereby achieving online calibration of the physical layer topology reference parameters.

[0138] Specifically, for impedance mismatch at a specific physical location determined in the vehicle health status diagnosis, the sampling controller recalculates the bit edge timing offset caused by the fault type, and uses a feedback adjustment circuit to compensate and adjust the internal voltage-controlled oscillator. The effective trigger edge of the sampling clock is finely offset towards the stable level after the signal transition, so as to re-align with the effective sampling range of the physical layer signal and realize the online calibration of the physical layer topology reference parameters.

[0139] This embodiment also provides a TBOX-based real-time vehicle fault diagnosis system, including:

[0140] The capture module is used by TBOX to adaptively switch sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status.

[0141] The correction module is used to analyze the reflected wave characteristics of the vehicle bus link characteristic flow and calculate the impedance state of the communication link, perform signal correction and phase compensation, and generate high-fidelity logic signals and communication link reliability indicators for the whole vehicle.

[0142] The quantization module is used to map the high-fidelity logic signals of the whole vehicle into discrete sequences that reflect the evolution of the signals, quantify the distribution complexity characteristics of the discrete sequences, and generate operating status feature vectors by combining communication link reliability indicators.

[0143] The judgment module is used to match and reason with the operating status feature vector and the preset fault tree model to determine the fault type and output the vehicle health status diagnosis conclusion.

[0144] The evidence storage module is used to perform secure evidence storage and remote synchronization operations based on the vehicle health status diagnosis results, and to perform online calibration to adjust the physical layer topology reference parameters according to the diagnostic confidence level.

[0145] This embodiment also provides a computer device applicable to the real-time vehicle fault diagnosis method based on TBOX, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the real-time vehicle fault diagnosis method based on TBOX as proposed in the above embodiment.

[0146] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0147] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the real-time vehicle fault diagnosis method based on TBOX as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0148] In summary, this invention achieves precise localization of microscopic electrical damage at the bus physical layer and quantitative characterization of signal evolution disorder by: reconstructing the link impedance based on the time delay and amplitude characteristics of reflected waves, superimposing a compensation sequence with opposite phases to cancel signal distortion, and using symbol dynamics to map the logic bit stream into an information entropy feature vector reflecting the signal evolution law. This not only eliminates communication errors caused by physical link quality degradation, but also deeply mines the hardware load interference characteristics hidden behind the waveform. It can be used to identify hardware interface faults or performance degradation precursors of core entities such as engines and chassis. Without relying on traditional diagnostic codes, it can discover and characterize hidden damage to vehicle hardware through physical layer fingerprints, improving the sensitivity and accuracy of real-time diagnosis of vehicle sub-health conditions.

[0149] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A real-time vehicle fault diagnosis method based on TBOX, characterized in that, include: TBOX adaptively switches sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status. The reflected wave characteristics of the vehicle bus link are analyzed and the impedance state of the communication link is calculated. Signal correction and phase compensation are performed to generate high-fidelity logic signals for the whole vehicle and reliability indicators of the communication link. The method for analyzing the reflected wave characteristics of the vehicle bus link characteristic flow includes: Identify the instantaneous location in the characteristic flow of the vehicle bus link where the voltage amplitude causes a level polarity reversal, and locate the reference moment of the logic level transition edge; Starting from the reference time, based on the signal propagation delay corresponding to the physical topology of the vehicle bus, a reflection wave observation time window covering the potential reflection signal area is defined backward; and the difference between the maximum fluctuation point of voltage amplitude deviating from the steady-state level and the steady-state level is extracted as the peak amplitude of the reflection wave. Identify the peak time of the reflected wave corresponding to the point of maximum fluctuation within the observation time window of the reflected wave; measure the time difference between the peak time of the reflected wave and the reference time as the reflection time delay. The method for calculating the impedance state of the communication link includes: The reflection coefficient is determined based on the ratio of the peak amplitude of the reflected wave to the standard amplitude of the incident wave, and the positive or negative sign information representing the change in impedance polarity is determined based on the offset direction of the reflected wave relative to the steady-state level. The physical properties of impedance mismatch are determined based on the positive and negative sign information. When the positive and negative sign information is positive, a high impedance mismatch point is determined to exist, and when the positive and negative sign information is negative, a low impedance mismatch point is determined to exist. By utilizing the reflection time delay and the signal propagation rate of the vehicle bus, the physical location of the impedance mismatch is mapped; by associating the physical properties and physical location of the impedance mismatch, the communication link impedance state associated with the electrical characteristics of the vehicle hardware interface is output. The high-fidelity logic signals of the whole vehicle are mapped into discrete sequences that reflect the evolution of the signals, and the distribution complexity characteristics of the discrete sequences are quantified. Combined with the reliability index of the communication link, an operating status feature vector is generated. The method for mapping the high-fidelity logic signals of the whole vehicle into discrete sequences that reflect the evolution law of the signals includes: The high-fidelity logic signal of the whole vehicle is sampled and extracted according to the sampling clock frequency to obtain the original bit value sequence; the flip points of adjacent values ​​are captured, and the edge evolution features of the logic state transition position are extracted. Based on the pulse width reflected by the edge evolution characteristics, the original bit value sequence is converted into symbolic code elements representing different signal evolution states; these are then arranged according to temporal relationships to generate discrete sequences. Methods for generating runtime state feature vectors include: The frequency of occurrence of each symbol in a discrete sequence is statistically analyzed to determine the symbol probability distribution, which represents the probability distribution of the signal evolution state. The information entropy value of the discrete sequence is calculated using the symbol probability distribution to quantify the distribution complexity characteristics of the signal distortion disorder. The distribution complexity characteristics are normalized and aligned with the communication link reliability indicators to form fused data of both physical link and signal logic states; the fused data is vectorized and combined according to preset dimensions to generate operational state feature vectors. The system matches and infers the operating status feature vector with the preset fault tree model to determine the fault type and outputs the vehicle health status diagnosis conclusion. Based on the vehicle health status diagnosis results, secure evidence storage and remote synchronization operations are performed, and online calibration is performed according to the diagnostic confidence level to adjust the physical layer topology reference parameters.

2. The real-time vehicle fault diagnosis method based on TBOX as described in claim 1, characterized in that, The method for obtaining the feature flow of the vehicle bus link includes: Compare the available computing load of TBOX with the preset resource threshold, and the vehicle bus communication load with the preset communication threshold; when the available computing load is greater than the preset resource threshold and the vehicle bus communication load is greater than the preset communication threshold, switch the physical layer signal acquisition mode to the full waveform capture mode; otherwise, switch to the edge feature extraction mode based on the hardware pulse capture circuit. In a defined physical layer signal acquisition mode, in response to the start bit edge of a message or a hardware abnormal interrupt pulse, a sampling trigger window is opened to achieve synchronous alignment between the signal edge and the sampling clock. Within the sampling trigger window, logical frame data and physical link feature data are extracted synchronously, and the sampling points of multiple communication cycles are recombined based on equivalent sampling to output the vehicle bus link feature stream.

3. The real-time vehicle fault diagnosis method based on TBOX as described in claim 1, characterized in that, The method for performing signal correction and phase compensation includes: Based on the impedance state of the communication link, a reflection cancellation compensation sequence is generated and superimposed onto the characteristic flow of the vehicle bus link to compensate for signal distortion caused by impedance mismatch. After detecting the bit edge timing offset of the characteristic stream of the vehicle bus link after signal correction, the trigger phase of the sampling clock is adjusted to align with the valid data determination interval, thus completing phase compensation. The system performs logic level determination on the characteristic flow of the vehicle bus link after signal correction and phase compensation, outputs a high-fidelity logic signal for the whole vehicle, and generates a communication link reliability index based on the waveform characteristics during the compensation process.

4. The real-time vehicle fault diagnosis method based on TBOX as described in claim 1, characterized in that, The method for outputting vehicle health status diagnostic conclusions includes: The running state feature vector is mapped to the bottom node of the preset fault tree model as the initial evidence for logical reasoning. Based on the comparison between the running state feature vector and the threshold of the preset fault tree model, the logic gate instruction in the preset fault tree model is activated to perform matching reasoning using the initial evidence for logical reasoning. Track the activated logical paths in the preset fault tree model to identify the fault modes and risk levels associated with the paths and determine the fault type. By associating fault types with the reasoning reliability scores generated during the matching reasoning process, the diagnostic confidence level is determined, and the vehicle health status diagnostic conclusion is output.

5. The real-time vehicle fault diagnosis method based on TBOX as described in claim 1, characterized in that, The method for performing online calibration based on diagnostic confidence levels to adjust physical layer topology baseline parameters includes: The vehicle health status diagnosis results are encrypted and stored locally and pushed synchronously to the cloud; when the diagnostic confidence reaches the preset trigger threshold, the online calibration trigger condition is determined to be met. Based on the fault type in the vehicle health status diagnosis conclusion, the trigger phase of the sampling clock is corrected to align with the effective sampling range of the physical layer signal, thereby achieving online calibration of the physical layer topology reference parameters.

6. A vehicle fault real-time diagnosis system based on TBOX, based on the vehicle fault real-time diagnosis method based on TBOX as described in any one of claims 1 to 5, characterized in that, include: The capture module is used by TBOX to adaptively switch sampling modes to obtain the characteristic flow of the vehicle bus link based on the computing load and bus communication status. The correction module is used to analyze the reflected wave characteristics of the vehicle bus link characteristic flow and calculate the impedance state of the communication link, perform signal correction and phase compensation, and generate high-fidelity logic signals and communication link reliability indicators for the whole vehicle. The quantization module is used to map the high-fidelity logic signals of the whole vehicle into discrete sequences that reflect the evolution of the signals, quantify the distribution complexity characteristics of the discrete sequences, and generate operating status feature vectors by combining communication link reliability indicators. The judgment module is used to match and reason with the operating status feature vector and the preset fault tree model to determine the fault type and output the vehicle health status diagnosis conclusion. The evidence storage module is used to perform secure evidence storage and remote synchronization operations based on the vehicle health status diagnosis results, and to perform online calibration to adjust the physical layer topology reference parameters according to the diagnostic confidence level.

Images