Unmanned vehicle fault detection system

Abstract

The invention belongs to the technical field of intelligent traffic, and particularly relates to an unmanned vehicle fault detection system which comprises a calibration detection module, an information analysis module, an abnormity judgment module, a fault alarm module and a redundancy guarantee module. The calibration detection module collects environmental parameters and calibration parameters and sends the environmental parameters and the calibration parameters to the information analysis module, the information analysis module analyzes data and calculates calibration detection coefficients, the abnormity judgment module compares the coefficients with a fault threshold value and then sends a result to the fault alarm module, and the fault alarm module executes vehicle operation and remote notification. The redundancy guarantee module guarantees operation under extreme working conditions. The extreme environment can be monitored in real time, sensor calibration can be triggered, the blank of calibration detection after the extreme environment is filled up, and hidden problems such as parameter drift are solved; redundancy guarantee and fault alarm are linked, the stability of the system is ensured, the running safety of the vehicle in a complex environment is improved, and the probability of safety accidents is reduced.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology, and specifically relates to a fault detection system for unmanned vehicles. Background Technology

[0002] The safe operation of autonomous vehicles relies heavily on the accurate data output of perception sensors such as LiDAR and cameras. The calibration accuracy of these sensors directly determines the reliability of the entire autonomous driving system. Currently, sensor fault detection systems are typically used to detect and warn of obvious faults such as sensor data transmission interruptions and hardware damage during autonomous driving.

[0003] However, certain technical defects were found during the implementation of the above technology. In some extreme environments, such as sandstorms, rainstorms, and extreme low temperatures, the sensor attitude may be deflected and the parameters may drift. However, the existing technology lacks targeted detection of the sensor calibration capability after extreme environments, making it difficult to fully guarantee the operational safety of unmanned vehicles in complex all-weather environments. Therefore, an improved unmanned vehicle fault detection system was designed. Summary of the Invention

[0004] To address the aforementioned shortcomings in the existing technology, this invention provides an unmanned vehicle fault detection system to solve the problems described in the background section.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] It includes a calibration and testing module, an information analysis module, an anomaly detection module, a fault alarm module, and a redundancy protection module. The calibration and testing module collects environmental parameters and calibration parameters and sends them to the information analysis module. The information analysis module analyzes the data collected by the calibration and testing module and calculates the calibration and testing coefficients and sends them to the anomaly detection module. The anomaly detection module determines the fault threshold and sends the judgment result to the fault alarm module. The fault alarm module executes according to the judgment result. The redundancy protection module ensures operation under extreme conditions.

[0007] Furthermore, the calibration and detection module includes an extreme environment threshold database, which compares the actual environmental parameters with those of the vehicle-mounted environmental sensors. When the actual environmental parameters reach the threshold, an extreme environment signal is triggered and the time T0 is recorded. At the same time, the sensor is triggered to quickly calibrate and record the calibration start time T1 and the completion time T2. T1-T0+TX and T2-T1=WC, where TX is the calibration response time and WC is the calibration completion time.

[0008] Furthermore, the calibration parameters include the real-time parameter Cs after calibration and the standard parameter Cb, calculated using the deviation formula. And calculate the parameter deviation PC.

[0009] Furthermore, the information analysis module includes a preset quantization model, which calculates the calibration detection coefficient XZ based on the calibration response time TX, calibration completion time WC, parameter deviation PC, and preset error adjustment factor β, maximum allowable calibration response time TX0, maximum allowable calibration completion time WC0, and weighting factors d1, d2, and d3 in the calibration and detection module.

[0010] Furthermore, the preset quantization model includes the formula for calculating the calibration detection coefficient XZ as follows: .

[0011] Furthermore, the anomaly detection module makes a judgment by setting a preset calibration fault threshold XZy and comparing XZy with the calibration detection coefficient XZ. If XZy ≤ XZ, the sensor calibration is determined to be abnormal and marked as faulty; if XZy > XZ, the sensor calibration is determined to be normal.

[0012] Furthermore, the fault alarm module includes vehicle alarm execution and remote notification. Vehicle alarm execution includes downgrading driving, restricting starting, and safe parking. Remote notification includes notifying a human to take over remotely.

[0013] Furthermore, the redundancy protection module includes a dual power supply unit and a manual calibration trigger unit. The dual power supply unit includes an on-board main power supply and a backup power supply, which automatically switches to the backup power supply when the main power supply fails. The manual calibration trigger unit can be manually calibrated via a remote control terminal.

[0014] Compared with the prior art, the present invention has the following advantages:

[0015] 1. By adopting a calibration and detection module to monitor extreme environmental parameters such as sandstorms, rainstorms, and extreme low temperatures in real time, and by actively triggering the sensor calibration process and evaluating the calibration capability, the aim is to fill the gap in the existing technology for sensor calibration and detection after extreme environments, solve implicit calibration problems such as parameter drift and attitude deviation, and improve the operational safety of unmanned vehicles in complex all-weather environments.

[0016] 2. By employing redundant protection modules to ensure the stable operation of the detection system, and in conjunction with fault alarm modules to link vehicles to perform degraded driving, restrict start-up, safely park, and trigger remote notifications, the system is guaranteed not to fail under extreme operating conditions, improving system reliability and risk control efficiency, shortening risk response time, and reducing the probability of safety accidents caused by calibration anomalies. Attached Figure Description

[0017] Figure 1 This is an overall flowchart of an unmanned vehicle fault detection system according to the present invention;

[0018] Figure 2This is a flowchart illustrating the calibration and testing module of an unmanned vehicle fault detection system according to the present invention.

[0019] Figure 3 This is a flowchart of the information analysis and anomaly judgment module of the fault detection system for unmanned vehicles according to the present invention; Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solution of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0021] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual images. They should not be construed as limiting the scope of this application. To better illustrate the embodiments of the present invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0022] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present application. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0023] In the description of this invention, unless otherwise explicitly specified and limited, the term "connection" or similar designation indicating a connection between components should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0024] Example 1:

[0025] like Figure 1 As shown in Figure 3, this invention discloses a fault detection system for unmanned vehicles. Specifically, it includes a calibration and detection module that collects extreme environmental parameters and calibration parameters, and sends the collected parameters to an information analysis module in real time.

[0026] The calibration and detection module has a built-in environmental threshold database, where the preset standards are: when the dust concentration is ≥50mg / m³, it is a sandstorm environment; when the rainfall is ≥20mm / h, it is a rainstorm environment; and when the ambient temperature is ≤-30℃, it is an extreme low temperature environment. The real-time environmental parameters are collected in real time by an on-board environmental sensor. The on-board environmental sensor is preferably a BME680 sensor. When the calibration and detection module detects that the environmental parameters have reached the extreme environmental threshold, it triggers an extreme environmental signal and records the trigger time T0.

[0027] The sensor fast calibration process is automatically triggered when the extreme environment signal is triggered. The sensor fast calibration process is a publicly available technical solution commonly used in the preferred Velodyne VLP-16 lidar. The calibration start time T1 and calibration completion time T2 are recorded simultaneously, and the calibration response time TX=T1-T0 and the calibration completion time WC=T2-T1 are calculated.

[0028] The calibration parameters include the real-time parameter Cs after calibration, such as the lidar ranging value, which is compared with the preset standard parameter Cb using the deviation formula. And calculate the parameter deviation PC, where:

[0029] After calibration, the real-time parameter Cs is the core parameter measured value collected in real time by the sensor's built-in detection unit, such as the ranging module of the lidar, after the sensor completes the rapid calibration process. The core parameter measured value is specifically the measured result of the sensor's core performance parameters, such as the real-time ranging value of the lidar, and the measured values ​​of the imaging clarity parameter and frame rate parameter of the camera sensor.

[0030] The standard parameter Cb is a standard parameter value determined after benchmark calibration of the sensor under standard laboratory conditions, and is stored in the system database in advance as a comparison benchmark.

[0031] The information analysis module receives parameters sent by the calibration and detection module, including dust concentration, rainfall, ambient temperature, calibration parameters, calibration response time TX, calibration completion time WC, parameter deviation PC, calculates the calibration and detection coefficient XZ based on the preset quantization model, and sends the calculated XZ to the anomaly judgment module.

[0032] The specific calculation model is as follows: The error adjustment factor β = 0.98 ensures that the calculation accuracy of XZ meets the requirements for fault diagnosis, while avoiding excessive consumption of computing resources due to the factor value being too close to 1.

[0033] The maximum allowable calibration response time is preset to TX0=100ms. To consider the real-time safety of autonomous vehicles, 100ms is set as the minimum real-time requirement threshold for fault warning in autonomous driving systems. This allows sufficient time for subsequent fault alarm modules to trigger alarms and execute degradation strategies.

[0034] The preset maximum allowable calibration completion time is WC0 = 500ms;

[0035] The weighting factors are d1=2.6, d2=2.4, and d3=4.0. The parameter deviation PC directly reflects the sensor calibration accuracy and is a core prerequisite for determining the reliability of sensor data. Data reliability is directly related to the accuracy of perception decisions; any deviation can directly lead to safety accidents, thus having the most direct impact on safety risks and the highest priority, and is assigned the maximum weight d3=4.0. Calibration response time TX reflects the timeliness of calibration triggering; excessive delays will lead to delayed fault warnings and missed risk control opportunities, thus having a lower priority, and is assigned the weight d1=2.6. Calibration completion time WC only reflects calibration efficiency; as long as the calibration accuracy meets the standard and the response is timely, the direct impact on safety risks is minimal, and is assigned the minimum weight d2=2.4. The specific numerical differences of d1, d2, and d3 are distinguished by numerical gradients to differentiate the degree of influence, ensuring that the core parameter PC dominates the calculation of the calibration detection coefficient XZ while avoiding the excessive weakening of the influence of secondary parameters such as calibration response time TX and calibration completion time WC. In selecting the weighting factor values, if the difference in the selected weighting factor values ​​is too large, it will over-amplify the influence of a certain parameter and severely weaken the role of a certain part. At the same time, the actual difference in the priority of the calibration response time TX and the calibration completion time WC is small, so the difference in the weighting factor values ​​should be small.

[0036] The anomaly detection module includes a preset calibration capability fault threshold XZy, which is set to 1.5. The safety threshold setting for autonomous vehicles needs to balance "avoiding missed alarms" and "reducing false alarms". The threshold of 1.5 can ensure that calibration deviations that affect driving safety, such as response delay ≥ 50%, are addressed. Combined with the preset β=0.98, d1=2.6, and TX0=100ms, when TX≥150ms, i.e., a delay of 50%, which is a delay that affects safety, the model calculation yields (TX / TX0)×d1=(150 / 100)×2.6=3.9. The contribution of this single parameter alone can make XZ=0.98×3.9≈3.82≥1.5, triggering the fault detection.

[0037] If the parameter deviation PC is ≥5%, combined with d3=4.0, when PC=5% and accompanied by a slight response delay TX=120ms, with a delay of 20%, we can calculate PC×d3=5%×4.0=0.2, (TX / TX0)×d1=(120 / 100)×2.6=3.12, and comprehensively calculate XZ=0.98×(3.12+0.2)≈3.26≥1.5, which can accurately identify the fault.

[0038] The nominal normal calibration accuracy of the lidar is, for example, ≤2% parameter deviation after calibration for the Velodyne VLP-16 lidar. This means that after calibration under standard operating conditions, the upper limit of normal fluctuation in core parameters such as ranging values ​​falls within an acceptable range. The XZy=1.5 threshold corresponds to the abnormal scenario of "overall deviation of calibration parameters ≥50%". Here, "overall deviation" is not the deviation of a single parameter (such as PC), but rather a comprehensive deviation combining calibration response time TX, calibration completion time WC, and parameter deviation PC, i.e., determined through a quantification model. The calculation shows that when XZ reaches 1.5, it means that the overall deviation of the three calibration parameters has reached 50%, far exceeding the sensor's nominal 2% normal calibration deviation limit. At this point, it can be clearly determined that the calibration process has not met the preset accuracy requirements, which is a fault-level calibration failure. Conversely, when XZ < 1.5, even if individual parameters have slight fluctuations, the overall deviation does not exceed 50%, which is still within the range of normal calibration fluctuations. This can accurately delineate the boundary between normal calibration and calibration failure, avoid misjudging normal fluctuations as faults, and ensure that fault-level calibration failures are not missed.

[0039] Simultaneously considering sensor performance degradation under extreme environments, such as low temperatures leading to decreased calibration accuracy, a threshold of 1.5 provides a 25% safety redundancy compared to critical values ​​like 1.2, avoiding threshold misjudgment due to environmental interference and improving system stability.

[0040] The information analysis module compares XZ and XZy to generate a fault command, which is then simultaneously sent to the fault alarm module and the vehicle's main control system. Judgment logic: If XZ ≥ 1.5, the sensor calibration capability is deemed abnormal and marked as faulty; if XZ < 1.5, the calibration capability is deemed normal.

[0041] Fault Alarm Module: Its core function is to receive fault commands and trigger corresponding alarm signals, such as light alarms and remote notifications, to ensure timely response from maintenance personnel. At the same time, it outputs fault signals to the vehicle's main control system to trigger the vehicle's self-handling process. The vehicle's self-handling strategies include, but are not limited to: degraded driving, which limits the driving speed to within a safe threshold; restricted start, which prohibits the vehicle from starting when it is not in motion; and safe parking, which automatically pulls over to the side of the road in areas with safe conditions.

[0042] The redundancy protection module ensures stable operation of the testing system under extreme conditions, such as low temperatures, sandstorms, heavy rain, or fluctuations in the vehicle's electrical system, preventing calibration failures due to system malfunctions. Specifically, it includes a dual power supply unit and a manual calibration trigger unit.

[0043] The dual power supply unit includes an onboard main power supply and a backup power supply. The onboard main power supply is a conventional power source, directly connected to the vehicle's existing power supply system, meeting the daily power requirements of the testing system. The backup power supply is an independent energy storage power source, pre-charged and in standby mode. The dual power supply unit has a built-in voltage monitoring module that monitors the main power supply voltage and stability in real time. When a voltage drop or interruption is detected in the main power supply, a power switching mechanism is automatically triggered to ensure uninterrupted power supply to the testing system.

[0044] In extreme environments, the sensor may experience hardware malfunctions, such as jamming of the calibration trigger component, which may prevent or interrupt the automatic calibration process. By establishing communication with the detection system through a remote control terminal, such as the operation and maintenance backend or a dedicated handheld terminal, the operation and maintenance personnel can send a manual calibration trigger command after receiving an alarm signal. The detection system will then skip the automatic calibration trigger step and directly start the sensor calibration process, while simultaneously recording the calibration parameters to ensure that the calibration and testing can proceed normally.

[0045] This embodiment uses lidar as the detection object, and its specific working principle is as follows:

[0046] The system initializes, and each module completes self-tests to confirm that the calibration and detection module, information analysis module, anomaly judgment module, fault alarm module, redundancy protection module, and the LiDAR and vehicle main control system communicate normally. The preset values ​​for each parameter are: XZy=1.5, β=0.98, weighting factors d1=2.6, d2=2.4, d3=4.0, TX0=100ms, WC0=500ms, and the dust environment threshold is dust concentration ≥50mg / m³.

[0047] The calibration and detection module collects environmental data in real time through the vehicle-mounted environmental sensors. It detects a dust concentration of 60 mg / m³, reaching the extreme environmental threshold for sandstorms, triggering an extreme environmental signal and recording the trigger time T0. Subsequently, it automatically triggers the rapid calibration process for the lidar, recording the calibration start time T1 and calibration completion time T2. The calibration response time TX = T1 - T0 = 80 ms and the calibration completion time WC = T2 - T1 = 400 ms are calculated. The lidar's built-in ranging module collects a real-time ranging value of 101 m after calibration, while the preset standard ranging value is 100 m. The parameter deviation PC = 1% is calculated using the formula PC = |101 - 100| / 100 × 100%. The values ​​TX = 80 ms, WC = 400 ms, and PC = 1% are then sent to the information analysis module.

[0048] The information analysis module calculates the calibration detection coefficient XZ based on the quantization model. Substituting the parameters, we get: XZ=0.98×[(80 / 100)×2.6+(400 / 500)×2.4+1%×4.0]=0.98×[2.08+1.92+0.04]=3.96; The calculated XZ=3.96 is sent to the anomaly judgment module.

[0049] The anomaly detection module compares XZ=3.96 with the preset fault threshold XZy=1.5. Since 3.96≥1.5, it determines that the lidar calibration capability is abnormal, generates a calibration capability fault command, and sends it to the fault alarm module and the vehicle main control system simultaneously.

[0050] The fault alarm module receives fault commands, triggers light alarms, and sends remote notifications to maintenance personnel; the on-board main control system executes a degradation strategy, reducing the vehicle speed to 20km / h to reduce safety risks.

[0051] The redundancy protection module monitors the power supply status of the detection system in real time to confirm that the vehicle's main power supply is normal. If the calibration capability is abnormal and continues, maintenance personnel can trigger the manual calibration trigger unit through the remote control terminal to start the manual calibration process of the lidar.

[0052] The above are merely embodiments of the present invention. The circuits, electronic components, and modules involved are all prior art, fully achievable by those skilled in the art, and require no further explanation. The content protected by this application does not involve improvements to the software and methods. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art who possess all the common technical knowledge in the field prior to the application date or priority date, are aware of all prior art in that field, and have the conventional experimental methods applicable before that date, can improve and implement this solution based on the guidance provided in this application and their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the application.

Claims

1. A fault detection system for unmanned vehicles, characterized in that: It includes a calibration and detection module, an information analysis module, an anomaly detection module, a fault alarm module, and a redundancy protection module. The calibration and detection module collects environmental parameters and calibration parameters and sends them to the information analysis module. The information analysis module analyzes the data collected by the calibration and detection module and calculates the calibration and detection coefficients and sends them to the anomaly detection module. The anomaly detection module determines the fault threshold and sends the determination result to the fault alarm module. The fault alarm module executes according to the determination result. The redundancy protection module ensures operation under extreme conditions.

2. The unmanned vehicle fault detection system as described in claim 1, characterized in that: The calibration and detection module includes an extreme environment threshold database, which compares actual environmental parameters with those of the vehicle-mounted environmental sensors. When the actual environmental parameters reach the threshold, an extreme environment signal is triggered and the time T0 is recorded. At the same time, the sensor is triggered to quickly calibrate and record the calibration start time T1 and the completion time T2. T1-T0+TX and T2-T1=WC are the calibration response time and the calibration completion time, respectively.

3. The unmanned vehicle fault detection system as described in claim 2, characterized in that: The calibration parameters include the real-time parameter Cs after calibration and the standard parameter Cb, which are determined by the deviation formula. And calculate the parameter deviation PC.

4. The unmanned vehicle fault detection system as described in claim 1, characterized in that: The information analysis module includes a preset quantization model, which calculates the calibration detection coefficient XZ based on the calibration response time TX, calibration completion time WC, parameter deviation PC, and preset error adjustment factor β, maximum allowable calibration response time TX0, maximum allowable calibration completion time WC0, and weighting factors d1, d2, and d3 in the calibration detection module.

5. The unmanned vehicle fault detection system as described in claim 4, characterized in that: The preset quantization model includes the following formula for calculating the calibration detection coefficient XZ: .

6. The unmanned vehicle fault detection system as described in claim 1, characterized in that: The anomaly detection module makes a judgment by setting a preset calibration fault threshold XZy and comparing XZy with the calibration detection coefficient XZ. If XZy ≤ XZ, the sensor calibration is determined to be abnormal and marked as faulty; if XZy > XZ, the sensor calibration is determined to be normal.

7. The unmanned vehicle fault detection system as described in claim 1, characterized in that: The fault alarm module includes vehicle alarm execution and remote notification. The vehicle alarm execution includes downgrading driving, restricting starting, and safe parking. The remote notification includes notifying a human to take over remotely.

8. The unmanned vehicle fault detection system as described in claim 1, characterized in that: The redundancy protection module includes a dual power supply unit and a manual calibration trigger unit. The dual power supply unit includes an on-board main power supply and a backup power supply. When the main power supply fails, it automatically switches to the backup power supply. The manual calibration trigger unit can be manually calibrated via a remote control terminal.

Images