Method for detecting an ingress of water into a housing of a control device of a vehicle, and control device

Abstract

The invention relates to a method for detecting a critical ingress of water into a housing (200) of an electronic control unit (202), in particular a steering device (204) or a braking device, of a vehicle (206), wherein the electronic control unit (202) has a temperature and humidity sensor (208), the method comprising the steps of: providing temperature and humidity measurement time series which can be recorded by the temperature and humidity sensor (208); providing a trained explainable boosting machine, EBM; and detecting a critical ingress of water into the housing (200) on the basis of the provided temperature and humidity measurement time series by means of the trained EBM.

Description

[0001] R. 414975

[0002] - 1 -

[0003] Description

[0004] title

[0005] Method for detecting water ingress into the housing of a vehicle control unit and control unit

[0006] The invention relates to a method for detecting critical water ingress into the housing of an electronic control unit, in particular a steering or braking system, of a vehicle. The invention relates to an electronic control unit. The invention also relates to a computer program with program code and a computer-readable data carrier containing the program code of a computer program.

[0007] State of the art

[0008] Electronic control units (ECUs) for motor vehicles, in particular box-by-wire control units used to control steering operations, are known.

[0009] Control units of this type are subject to particular challenges regarding environmental influences, as they are often installed in locations exposed to external elements such as moisture and water. A typical problem is that water penetrating through a damaged or leaking housing can lead to significant malfunctions or even complete failure of the control unit. These failures often result from common causes, such as short circuits or corrosion of electronic components. In critical systems like a vehicle's steering system, such malfunctions are particularly relevant to safety.

[0010] There is a correlation between the stresses placed on a vehicle and a potential failure of the electronic control unit (ECU). For example, R. 414975

[0011] - 2 - External stresses (temperature, relative humidity, and splash water) act on the vehicle, which are transferred to the control unit as internal stresses (internal temperature, humidity, and water). This leads to stress in the ECU circuit, which is intensified by electric field strengths and ionic contaminants, potentially resulting in localized dew and condensation. This stress can eventually lead to damage in the ECU, particularly through migration on the printed circuit board (PCB). This process can ultimately lead to ECU failure, for example, due to a short circuit or increased leakage currents. In such cases, the ECU should be monitored to warn the driver or trigger a fail-safe protocol. Such monitoring is not currently readily available.The stresses and failures depend on various factors, including the vehicle's environment, the position of the control unit in the vehicle, the design of the ECU, and the layout and coating of the circuit board.

[0012] To prevent such damage, it is already known to install humidity sensors near the electronic control unit that can detect elevated humidity and initiate appropriate measures. However, there is still potential for improvement in this area, particularly regarding the processing and use of the data collected by the humidity sensors.

[0013] It is an object of the invention to provide a method for detecting a critical water ingress into a housing of an electronic control unit, in particular a steering or braking device, of a vehicle and / or an electronic control unit.

[0014] The problem is solved by a method according to the features of claim 1. The problem is solved by an electronic control unit according to the features of claim 7.

[0015] Disclosure of invention R. 414975

[0016] - 3 -

[0017] According to a first aspect, a method for detecting critical water ingress into the housing of an electronic control unit, in particular a steering or braking system, of a vehicle is proposed. The electronic control unit has at least one temperature and humidity sensor. The method comprises the following steps: providing temperature and humidity measurement time series that can be acquired by the temperature and humidity sensor; providing a trained Explainable Boosting Machine (EBM); and detecting critical water ingress into the housing based on the provided temperature and humidity measurement time series using the trained EBM.

[0018] It is understood that the steps according to the invention, as well as further optional steps, do not necessarily have to be carried out in the sequence shown, but can also be carried out in a different sequence. Furthermore, additional intermediate steps may be provided. The individual steps may also comprise one or more sub-steps without thereby departing from the scope of the method according to the invention.

[0019] According to a second aspect, an electronic control unit, in particular of a steering or braking system, of a vehicle is proposed, wherein the electronic control unit comprises a housing, at least one temperature and humidity sensor and an evaluation and computing unit, wherein the at least one temperature and humidity sensor is configured to record temperature and humidity measurement time series, and wherein a trained Explainable Boosting Machine (EBM) for detecting a critical water ingress into the housing based on the provided temperature and humidity measurement time series is executable on the evaluation and computing unit.

[0020] Particularly preferably, the electronic control unit comprises at least one printed circuit board for providing control functions, wherein the at least one temperature and humidity sensor is arranged on the printed circuit board and preferably interconnected with the printed circuit board. In this way, it is possible to process the temperature and humidity measurement time series via the electronics of the printed circuit board and to have them evaluated by the EBM. R. 414975

[0021] - 4 -

[0022] For example, the temperature and humidity sensor can be soldered to the circuit board.

[0023] The control unit can be included in a vehicle with or without an autonomous driving function and / or in a robotics system and / or in an industrial machine with a steering or braking device.

[0024] The at least one temperature and humidity sensor can have multiple sensors. The temperature and humidity sensor can include both a temperature sensor and a humidity sensor. Multiple temperature and humidity sensors can also be arranged on the housing, preferably at several locations on the housing, to record temperature and humidity measurement time series. The temperature and humidity measurement time series are recorded according to the cycle time of the at least one temperature and humidity sensor. Thus, for example, measured values ​​can be recorded per millisecond, per second, or per cycle time.

[0025] The Explainable Boosting Machine (EBM) is a machine learning model designed to be both high-performing and explainable. EBM combines the advantages of boosting algorithms with high interpretability, enabling users to better understand the model's decision-making processes. EBM is based on an ensemble of weak learners, typically decision trees, which are trained iteratively. Each new tree attempts to correct the errors of the previous trees, resulting in a strong and accurate overall model. EBM is an explainable model, thus mitigating the risk of failure. Furthermore, it ensures deterministic or probabilistic and scientifically sound model behavior.

[0026] The temperature and humidity measurement time series can preferably be preprocessed by means of data transformation and / or sampling and / or data cleaning.

[0027] Based on the provided temperature and humidity measurement time series, the EBM can determine the criticality of a specific water ingress and R. 414975

[0028] - 5 - thereby deciding whether water ingress is critical or not. Not every instance of water ingress is critical for a control unit, and not every compromised housing integrity poses a risk to the control unit's functionality. The present method makes it possible to derive a risk threshold based on the control unit's internal environment, indicating the potential for undesirable effects on the control unit's functions due to water ingress.

[0029] The explanations given for the procedure apply accordingly to the control unit. It is understood that linguistic modifications of procedurally formulated characteristics can be reformulated for the device according to common linguistic practice, without such formulations needing to be explicitly listed here.

[0030] The present invention aims in particular to further optimize the integration of temperature and humidity sensors and to process the acquired data in such a way that preventive measures can be taken at an early stage before a system failure of the control unit occurs. Improved analysis of the temperature and humidity measurement time series allows for more efficient prevention of control unit malfunctions caused by water ingress and further increases the reliability of the control units. The present invention aims in particular to use a temperature-humidity sensor to detect critical water ingress early enough to enable reactive or even preventive measures for the control unit.

[0031] The present method is based on feature engineering and data identification / labeling by domain experts, which enabled the training of the EBM (Electronic Principles). The trained EBM can then be executed on a control unit during inference. If necessary, further validation and, if required, adjustment of the EBM can also be performed during inference.

[0032] In another aspect, it is proposed that water ingress into the housing depends on at least one of the provided temperature values. R. 414975

[0033] - 6 - and moisture measurement time series derivable key figure when a predetermined limit criterion is met is detected as "critical".

[0034] The key figure can be used to derive the risk of an internal control unit error caused by water ingress. The model output risk is preferably a risk based on prior labeling or marking by experts.

[0035] In another aspect, it is proposed that the indicator has a relative humidity and / or a temperature difference between a measured temperature and a dew point temperature of the housing and / or a difference of the temperature differences squared, and wherein the predetermined limit criterion is selected depending on the indicator.

[0036] To derive key performance indicators from the raw data, it is preferable to perform a preliminary evaluation. For example, temperature and humidity measurement time series can be recorded in a climate chamber, where the sensor is preferably exposed to dynamic climate changes.

[0037] Using kernel density estimation, an inhomogeneous distribution of the raw data can preferably be analyzed and taken into account for subsequent effect separation in the trained model. This can also be used, for example, to determine a maximum permissible aging drift of the sensor over time. Feature engineering can then be based on this.

[0038] Relative humidity can be used to determine an atypical internal climate within the housing. Changes in the internal climate can also be detected based on this measurement. Temperature difference can be used to detect an atypical, condensation-prone internal climate within the control unit. This allows for the detection of both gradual and rapid changes. The temperature difference is also referred to as the temperature spread. The dew point can be calculated from the temperature and humidity values ​​and predetermined constants, or directly from the sensor's internal calculations. The square of the temperature difference enables the detection of atypical design changes in the internal climate. R. 414975

[0039] - 7 - of the control unit, which are impossible, for example, due to a buffering effect of the intact housing. This makes the detection of very rapid changes possible. It should be noted that other possible predictors, based on absolute humidity, (saturation) vapor pressure, dew point, etc., are linearly dependent on the predictors selected above with respect to their changes.

[0040] As previously described, the EBM model has, in particular, input features such as temperature difference, relative humidity, and the square of the change in temperature difference. Since temperature difference and relative humidity contain partially similar information, the displayed importance of these features depends on their training order. Because temperature difference is the most robust feature, the training order is preferably set to maximize its importance.

[0041] Trained EBM (Electro-Brained Materials) professionals know that small temperature differences are risky. If the temperature difference is small, condensation depends on the chemical contamination on the circuit board. The larger the temperature difference, the lower the risk of condensation.

[0042] The trained EBM (Electro-Borne Medicine) knows that high relative humidity is risky. This makes sense, as a well-designed enclosure exhibits damping characteristics against external climatic conditions (like a low-pass filter). If a relative humidity, for example above 80% RH, is measured, the enclosure's integrity is compromised, meaning that its damping characteristics change drastically, for example due to a hole or water ingress.

[0043] The trained EBM knows that large changes in temperature difference are risky. This is clear because the damping behavior of the housing does not allow increased changes in temperature difference within a sampling time of 10 seconds. If there is a significant squared change in temperature difference, water will directly enter the housing. R. 414975

[0044] - 8 -

[0045] In another aspect, it is proposed that a critical water ingress is an ingress of water into the housing where a malfunction of the electronic control unit occurs or is to be expected with a predetermined probability.

[0046] The predetermined probability can be defined, for example, in a code. This code defines, for instance, a risk function that calculates the risk based on various factors such as temperature, humidity, dew point, dispersion, and dispersion differential. The code preferably iterates through the data acquired by the sensor and calculates the risk value for each row. The risk values ​​are preferably stored in a new column. This process is carried out with the involvement of all experts, drawing on their expertise in expert risk assessment. Any markings that are not automated are preferably made manually, in accordance with the aforementioned automated and adapted function.

[0047] In another aspect, it is proposed that the predetermined limit criterion and / or the predetermined probability and / or a model sensitivity of the EBM can be manipulated based on a change in the housing integrity of the housing.

[0048] Furthermore, there are design features of the housing that are known to correspond to a good housing integrity state. If a change in these design features is detected, the issue risk of the model may preferably be increased.

[0049] In a further aspect, it is proposed that the trained EBM is trained on the basis of labeled training data, wherein the training data includes temperature and humidity measurement time series and labels assigned to the temperature and humidity measurement time series regarding critical or non-critical water ingress. The labels can be automatically generated based on a comparison of the difference between a dew point temperature of the housing and a temperature at the coldest point of the housing, which can be determined by simulation or testing, with a definable threshold value. R. 414975

[0050] - 9 -

[0051] If the dew point temperature of the housing differs from the temperature at the coldest point of the housing by, for example, only 1 Kelvin, the associated data can be labelled as data indicating critical water ingress.

[0052] Preferably, the data is split into training and test datasets. Bayesian optimization is preferably performed, defining an objective function for the Bayesian optimization that optimizes the hyperparameters of the EBM model. The function creates and fits an explainable boosting regressor model with selected hyperparameters and calculates a mean squared error (MSE) on a validation dataset. The Bayesian optimization is performed, for example, using Optuna to find the best hyperparameters. Preferably, the EBM model is fitted with the best hyperparameters. The model is preferably saved. Furthermore, the model can be checked for scientific consistency.

[0053] In another aspect, a computer program is claimed to contain program code capable of executing at least parts of the present method in one of its aspects when the computer program is executed on a computer. In other words, a computer program (product) is claimed to comprise instructions that, when executed by a computer, cause it to execute the method(s) in one of its aspects.

[0054] In another aspect, a computer-readable data carrier containing the program code of a computer program is proposed to execute at least parts of the present method in one of its aspects when the computer program is executed on a computer. In other words, the invention relates to a computer-readable (storage) medium comprising instructions which, when executed by a computer, cause it to execute the method / steps of the method in one of its aspects. R. 414975

[0055] - 10 -

[0056] The described configurations and training programs can be combined in any way desired.

[0057] Further possible embodiments, developments and implementations of the invention also include combinations of features of the invention described previously or subsequently with regard to the exemplary embodiments that are not explicitly mentioned.

[0058] Brief description of the drawings

[0059] The accompanying drawings are intended to provide a further understanding of the embodiments of the invention. They illustrate embodiments and, in the context of describing the invention, serve to explain its principles and concepts.

[0060] Other embodiments and many of the aforementioned advantages become apparent with reference to the drawings. The elements depicted in the drawings are not necessarily shown to scale.

[0061] Fig. 1 shows a schematic flowchart of a

[0062] Exemplary embodiment of the present method.

[0063] Fig. 2 schematically shows an electronic control unit according to a

[0064] Design.

[0065] In the figures of the drawings, identical reference symbols denote identical or functionally equivalent elements, parts or components, unless otherwise stated.

[0066] Fig. 1 shows a schematic flowchart of a method for detecting a critical water ingress into a housing 200 of an electronic control unit 202 of a steering system 204 of a vehicle 206. Alternatively, the electronic control unit 202 can, for example, be associated with a braking system. The electronic control unit 202 is shown schematically in Fig. 2. R. 414975

[0067] - 11 -

[0068] The electronic control unit 202 has at least one temperature and humidity sensor 208.

[0069] The electronic control unit 202 further comprises an evaluation and computing unit 210. The electronic control unit 202 may also include several components not shown in detail, for example, one or more provisioning units. It is understood that the provisioning unit may be designed together with the evaluation and computing unit 210 or may be different from it. Furthermore, the electronic control unit 202, which may be part of a system, may include a storage unit and / or an output unit and / or a display unit and / or an input unit. The electronic control unit 202 also includes a printed circuit board 212 for providing control functions. The temperature and humidity sensor 208 is arranged on the printed circuit board 212 and connected to it.

[0070] The procedure includes at least the following steps:

[0071] In step S1, temperature and humidity measurement time series are provided, which can be recorded by the temperature and humidity sensor 208.

[0072] In step S2, a trained Explainable Boosting Machine (EBM) is deployed.

[0073] In step S3, a critical water ingress into housing 200 is detected by the trained EBM based on the provided temperature and humidity measurement time series. Water ingress into housing 200 can be detected as "critical" depending on at least one parameter derived from the provided temperature and humidity measurement time series, provided a predetermined limit criterion is met. The parameter can be a relative humidity and / or a temperature difference between a measured temperature and a dew point temperature of housing R. 414975

[0074] - 12 - and / or a squared difference in the temperature differences. The predetermined limit criterion can be selected depending on the key figure. A critical water ingress is preferably water ingress into the housing 200 where a malfunction of the electronic control unit 202 occurs or is expected with a predetermined probability. The predetermined limit criterion and / or the predetermined probability and / or a model sensitivity of the EBM can be manipulated based on a change in the housing integrity of the housing 200. The trained EBM is trained here on the basis of labeled training data, wherein the training data includes temperature and humidity measurement time series and labels assigned to the temperature and humidity measurement time series regarding a critical water ingress or a non-critical water ingress.The labels are based on a comparison of the difference between a dew point temperature of the housing 200 and another.

[0075] Temperature at the coldest point of the housing 200, which can be determined by simulation or tested, can be automatically generated with a definable limit value.

Claims

R. 414975 - 13 - Claims 1. Method for detecting a critical water ingress into a housing (200) of an electronic control unit (202), in particular a steering device (204) or a braking device, of a vehicle (206), wherein the electronic control unit (202) has a temperature and humidity sensor (208), the method comprising the steps: Providing (S1) temperature and humidity measurement time series that can be recorded by the temperature and humidity sensor (208); Deployment (S2) of a trained Explainable Boosting Machine, EBM; and Detecting (S3) a critical water ingress into the housing (200) based on the provided temperature and humidity measurement time series by the trained EBM.

2. Method according to claim 1, wherein water ingress into the housing (200) is detected as "critical" depending on at least one characteristic value derivable from the provided temperature and humidity measurement time series when a predetermined limit criterion is met.

3. Method according to claim 1 or 2, wherein the characteristic value comprises a relative humidity and / or a temperature difference between a measured temperature and a dew point temperature of the housing (200) and / or a difference of the temperature differences squared, and wherein the predetermined limit criterion is selected depending on the characteristic value. R. 414975 - 14 - 4. Method according to any one of claims 1 to 3, wherein a critical water ingress is a water ingress into the housing (200) in which a malfunction of the electronic control unit (202) occurs or is to be expected with a predetermined probability.

5. Method according to claim 2 or 4, wherein the predetermined limit criterion and / or the predetermined probability and / or a model sensitivity of the EBM is manipulable based on a change in the housing integrity of the housing (200).

6. Method according to one of the preceding claims, wherein the trained EBM is trained on the basis of labeled training data, wherein the training data comprise temperature and humidity measurement time series and labels assigned to the temperature and humidity measurement time series with respect to a critical water ingress or a non-critical water ingress, wherein the labels can be automatically generated on the basis of a comparison of a difference of a dew point temperature of the housing (200) and a temperature at a coldest point of the housing (200), which can be determined by simulation or tested, with a definable limit value.

7. Electronic control unit (202), in particular of a steering device (204) or a braking device, of a vehicle (206), wherein the electronic control unit (202) comprises a housing (200), a temperature and humidity sensor (208) and an evaluation and computing unit (210), wherein the temperature and humidity sensor (208) is configured to record temperature and humidity measurement time series, wherein a trained Explainable Boosting Machine (EBM) for detecting a critical water ingress into the housing (200) based on the provided temperature and humidity measurement time series can be executed on the evaluation and computing unit (210).

8. Electronic control unit according to claim 7, wherein the electronic control unit (202) comprises a printed circuit board (212) for providing control functions, wherein the temperature and R. 414975 - 15 - The humidity sensor (208) is arranged on the circuit board (212) and is preferably connected to the circuit board (212).

9. Computer program with program code to execute at least parts of a method according to any one of claims 1 to 6 when the computer program is executed on a computer.

10. Computer-readable data carrier containing program code of a computer program for executing at least parts of a method according to any one of claims 1 to 6 when the computer program is executed on a computer.

Images