METHOD FOR DETECTING AN ANOMALY ON A DEVICE OF INTEREST BELONGING TO A FLEET OF DEVICES
The method constructs a main database and uses statistical analysis to compare device data against a reference database of similar devices, enhancing anomaly detection precision by accounting for descriptive characteristics, thereby improving the accuracy of identifying atypical behavior.
Patent Information
- Application Number
- FR2023004834
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-16
- Publication Date
- 2024-11-22
AI Technical Summary
Existing methods for detecting anomalies in devices within a fleet, such as motor vehicles, are imprecise and rely on empirical threshold comparisons, lacking a comprehensive and precise method to identify atypical behavior.
A method involving the construction of a main database from collected data sets, extraction of a reference database excluding the device of interest, and statistical analysis to compare data from the device of interest against similar devices within the fleet, using descriptive characteristics like model, software version, and operating state to determine deviations beyond predetermined thresholds.
Enables precise detection of anomalies by comparing device behavior against similar devices, reducing false positives and improving accuracy in identifying atypical conditions.
Smart Images

Figure 00000013_0000
Abstract
Description
Title of the invention: METHOD FOR DETECTING AN ANOMALY ON A DEVICE OF INTEREST BELONGING TO A FLEET OF DEVICES TECHNICAL FIELD OF THE INVENTION
[0001] One aspect of the invention relates to a method for detecting an anomaly on a device of interest belonging to a fleet of devices. The invention finds a particularly interesting application in the field of motor vehicles, but can be applied to other devices such as high-pressure cleaning devices or any other device equipped with an electrical system. TECHNOLOGICAL BACKGROUND OF THE INVENTION
[0002] Motor vehicles intended for production are usually tested and validated during fleet runs carried out before series production. Data from a fleet of vehicles is recorded and made available to business teams for validation. To determine whether a component of a motor vehicle is defective or whether behavior is abnormal, the data is generally compared to thresholds. Acceptance limits are set empirically based on experience and business knowledge. Another method consists of calculating diagnostics in computers embedded in vehicles based on thresholds or sequences of events.
[0003] These methods make it possible to detect an operating anomaly in a vehicle by comparing it to previously determined thresholds. These methods can nevertheless be improved. Summary of the invention
[0004] The aim of the invention is to overcome the drawbacks of the prior art by proposing a method for detecting an anomaly on a device of interest belonging to a fleet of devices whose detection is more precise.
[0005] In this context, the invention thus relates, in its broadest acceptance, to a method for detecting an anomaly on a device of interest belonging to a fleet of devices, the method comprising the following steps, executed by a system for detecting an anomaly, • Acquire a new data set, in which each data represents the value of a respective quantity relative to said device of interest; • Update a so-called main database, to add said new data set acquired in association with information relating to at least one descriptive characteristic of said device of interest at the time acquisition of said new data set; • From said main database, extract a so-called reference database, the reference database grouping together data sets associated with devices of said fleet of devices having said at least one descriptive characteristic, excluding data sets associated with the device of interest; • From the main database, extract a so-called analysis database grouping together data sets associated with said device of interest and said at least one descriptive characteristic; • For at least one quantity, compare, by statistical analysis, the corresponding data from the analysis database and the corresponding data from the reference database; (in other words, the data from the analysis database associated with said quantity and the data from the reference database associated with this same quantity) • Detect an anomaly on said device of interest, when, for at least one of said quantities, the corresponding data of the analysis database deviates beyond a predetermined threshold relative to the corresponding data of the reference database.
[0006] By means of the invention, it is possible to determine whether a device belonging to a fleet of vehicles is atypical, based on the behavior of other vehicles having the same descriptive characteristic. In other words, the data of a device of interest are not simply compared to thresholds, but are compared to the data of similar devices in order to determine whether the device of interest has an anomaly. The devices are advantageously motor vehicles.
[0007] In addition to the characteristics which have just been mentioned in the preceding paragraph, the method according to this aspect of the invention may have one or more complementary characteristics among the following, considered individually or according to all technically possible combinations.
[0008] According to a non-limiting aspect of the invention, the method comprises a preliminary step of constructing the main database, this preliminary step consisting of grouping together a plurality of collected data sets, said data sets having been determined by the devices of the fleet of devices, each data item representing the value of a respective quantity relating to a device of said fleet of devices, each data set of a device of the fleet of devices being associated with at least one descriptive characteristic of said device at the time of acquisition of said data set.
[0009] According to a non-limiting aspect of the invention, at least one given, among the data from the new acquired data set, is measured by a sensor embedded on the device of interest, or is calculated, by a computer embedded on the device of interest, from at least two data acquired by two sensors embedded on said device of interest.
[0010] According to a non-limiting aspect of the invention, the quantities associated with the new acquired data set are decorrelated from each other.
[0011] According to a non-limiting aspect of the invention, the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set comprises the model of the device of interest, the reference database grouping together data sets associated only with devices of this same model.
[0012] According to a non-limiting aspect of the invention, the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set further comprises a version of software for controlling functionalities of the device of interest, installed on a computer of the device of interest at said time of acquisition, the reference and analysis databases grouping together data sets associated only with devices of this same model and in which the same version of the software is installed.
[0013] According to a non-limiting aspect of the invention, the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set further comprises an operating state of the device of interest at said acquisition time, the reference and analysis databases grouping together data sets associated only with devices having, at least, the same model and said operating state.
[0014] According to a non-limiting aspect of the invention, the reference and analysis databases group together data sets associated with acquisition times included in a time interval of predetermined duration and prior to the acquisition time of the new data set, and / or data sets associated with a maximum distance traveled by the device of interest up to the acquisition time of the new data set.
[0015] According to a non-limiting aspect of the invention, the method comprises a step of weighting the data sets.
[0016] According to a non-limiting aspect of the invention, the new data set, acquired by the anomaly detection system, comes directly from means of transmitting new data sets that the device of interest comprises.
[0017] According to a non-limiting aspect of the invention, the fleet of devices is formed by a fleet of motor vehicles and the device of interest is formed by a motor vehicle of interest.
[0018] Another aspect of the invention relates to a computer program product comprising instructions which, when the program is executed by the anomaly detection system, cause the latter to implement the steps of the method according to any of the aforementioned aspects.
[0019] The invention and its various applications will be better understood upon reading the following description and examining the accompanying figure. BRIEF DESCRIPTION OF THE FIGURES
[0020] The figure is presented for information purposes only and in no way limits the invention.
[0021] [Fig.l] schematically illustrates a system for detecting an anomaly configured to perform the steps of an anomaly detection method on a device of interest according to a non-limiting aspect of the invention. DETAILED DESCRIPTION
[0022] As illustrated in [Fig.l], the anomaly detection system 1 comprises anomaly detection means 2 arranged to communicate with grouping means 3, and a method 100 implemented within said anomaly detection system 1.
[0023] The means for detecting an anomaly 2 comprise at least one processor and at least one memory. They may be formed by one or more computers and / or one or more servers and / or one or more virtual machines.
[0024] These detection means 2 are configured to execute steps of the anomaly detection method 100 on a device of interest belonging to a fleet of devices.
[0025] Similarly, the grouping means 3 comprise at least one processor and at least one memory. They can be formed by one or more computers and / or one or more servers and / or one or more virtual machines.
[0026] These grouping means 3 are configured to execute a preliminary step of constructing a so-called main database, a step of updating the main database and a step of weighting certain data of the main database.
[0027] In the non-limiting example of the description, the devices are formed by motor vehicles. In variants, the devices may be machine tools, or cleaning devices, etc.
[0028] The method 100 here comprises a preliminary step, executed by the grouping means 3, of constructing 101 a main database, consisting of grouping a plurality of collected data sets where each data item represents the value of a respective quantity of a vehicle in the fleet of vehicles. In a non-limiting manner, the data may relate to the engine of the vehicle, to a consumption of the vehicle or even to a braking system of the vehicle. This preliminary step 101 is implemented implemented only once, and is not repeated thereafter so that the invention is not limited to a method comprising this preliminary step 101.
[0029] Each data item in a data set representing the value of a respective quantity of a vehicle can be measured by a sensor on board the vehicle, or calculated from at least two data items acquired by two sensors on board the vehicle. In the invention, the different data sets all relate to the same set of quantities.
[0030] It should be noted that the quantities of the different data sets are decorrelated from each other. For example, the vehicles do not drive on the same day and / or in different geographical areas and / or under different conditions of use. Preferably, at least some of the quantities of the same vehicle are also decorrelated, in that they do not relate to the same physical parameter of the same element of the vehicle.
[0031] The data sets can be transmitted to the grouping means 3 directly by means of transmitting new data sets that the vehicles contain.
[0032] In a different implementation, the data sets are downloaded during a check of a fleet vehicle in a vehicle tracking center, then transmitted from the tracking center to the grouping means 3.
[0033] Furthermore, it should be noted that each data set of a vehicle in the fleet of vehicles is associated with at least one descriptive characteristic of the vehicle at the time of acquisition of the data set.
[0034] The descriptive characteristic of the vehicle may be formed by the vehicle model, the term "model" being understood in its sense of "category", the vehicles of the same model having predetermined common characteristics. The vehicle model may be designated by a vehicle identification number, better known by the acronym VIN (for Vehicle Identification Number in English).
[0035] Optionally, the at least one descriptive characteristic of the vehicle may also include a version of software for controlling vehicle functionalities, installed on a computer of the vehicle.
[0036] Additionally or alternatively, the at least one descriptive characteristic of the vehicle may further comprise an operating state of the vehicle.
[0037] In a non-limiting manner, this operating state can be formed by a range of engine speeds or a range of temperatures outside the vehicle. Indeed, different operating states can sometimes explain drifts or specific wear of certain components.
[0038] In use after the main database has been constructed, the method 100 comprises a step, executed by the anomaly detection means 2, of acquiring 102 a new data set, in which each data item represents the value of a respective quantity relating to the vehicle of interest. The vehicle of interest belongs to a vehicle fleet, distinct from or similar to the vehicle fleet used in the preliminary step 101 of construction of the main database. The new data set can for example be transmitted by transmission means in the vehicle of interest, directly to the anomaly detection means 2.
[0039] The data sets may be transmitted in real time to the anomaly detection means 2, for example at a regular frequency in terms of mileage and / or duration of use. Alternatively, the data sets may be temporarily stored within the vehicle (for example at a regular frequency in terms of mileage and / or duration of use), before being transmitted in packets to the anomaly detection means 2, during a check of the vehicle in a vehicle tracking center.
[0040] The data sets can therefore pass directly from the vehicle to the anomaly detection means 2, by wireless communication. Alternatively, the data sets can pass first from the vehicle to a central computer, for example by wired communication, then from the central computer to the anomaly detection means 2, by wireless communication. Preferably, the data set of step 102 relates to the same data as the data set of step 101, only the values taken by the latter being possibly modified.
[0041] At least one piece of data, among the data of the new acquired data set, is measured by a sensor embedded on the device of interest, or is calculated, by a computer embedded on the device of interest, from of at least two data acquired by two sensors embedded on the device of interest.
[0042] The method 100 then comprises a step, executed by the grouping means 3, of updating 103 the main database, to add thereto the new acquired data set, in association with information relating to at least one descriptive characteristic of the vehicle of interest at the time of acquisition 102 of the new data set. Advantageously, the descriptive characteristic is transmitted to the means for detecting an anomaly 2 at the same time as the data set acquired in step 102.
[0043] Optionally, the at least one descriptive characteristic of the vehicle of interest at the time of acquisition of the new data set comprises the model of the vehicle of interest, for example the identification number of the vehicle of interest.
[0044] In yet another embodiment, the at least one descriptive characteristic of the vehicle of interest at the time of acquisition of the new data set comprises the model of the vehicle of interest and a version of a software for controlling the features of the vehicle of interest installed on a computer of the vehicle of interest.
[0045] In an optional implementation, the at least one descriptive characteristic of the vehicle of interest at the time of acquisition of the new data set comprises the model of the vehicle of interest and an operating state of the vehicle of interest, and possibly a version of software for controlling functionalities of the vehicle of interest installed on a computer of the vehicle of interest.
[0046] Optionally, the method 100 may comprise a step, executed by the grouping means 3, of weighting 104 the data sets. For example, in order to avoid distorting the results, it is possible to assign a lower weight to the data of a vehicle with high mileage (therefore providing data more frequently than the other vehicles), compared to the other vehicles in the fleet of vehicles.
[0047] From the main database, the method 100 comprises a step, executed by the anomaly detection means 2, of extracting 105 a so-called reference database, the reference database grouping together data sets associated with vehicles of the fleet of vehicles having said at least one descriptive characteristic, and the reference database excluding the data sets associated with the vehicle of interest. The exclusion of the data sets associated with the vehicle of interest is an original characteristic and which prevents the subsequent analysis from being distorted, in the case where the vehicle of interest has an anomaly.
[0048] In one embodiment, and in a non-limiting manner, the reference database may group together data sets associated with acquisition times included in a time interval of predetermined duration, prior to the acquisition time of the new data set. This predetermined duration may for example be one month. Preferably, the time interval considered is immediately prior to the acquisition time of the new data set.
[0049] This time interval makes it possible, by construction, to reduce the number of data to be processed and consequently to reduce the data processing time. In addition, it makes it possible to gradually get rid of less relevant data, relating for example to old versions of at least one data processing and acquisition software installed on a computer of the vehicle. This also makes it possible, in the case where an anomaly appears at a specific moment, to be able to be significantly present in the data set and not to be drowned in the middle of old data not presenting the anomaly.
[0050] Alternatively, the reference database groups together data sets associated with a maximum distance traveled by the device of interest up to the time of acquisition of the new data set. Preferably, the maximum distance considered is immediately prior to the time of acquisition of the new data set. data.
[0051] Thus, the data sets of the reference database can be associated with a vehicle model, possibly with a software version and / or with an operating state, and be limited to a time interval of predetermined duration or a maximum distance traveled.
[0052] Then, from the main database, the method 100 comprises a step, executed by the anomaly detection means 2, of extracting 106 a so-called analysis database grouping together data sets associated with the device of interest and with the at least one descriptive characteristic.
[0053] Thus, the data sets of the analysis database can be associated with a vehicle model, possibly with a software version and / or with an operating state, and be limited to a time interval of predetermined duration or a maximum distance traveled.
[0054] For at least one quantity of the data sets, the method 100 comprises a step, executed by the anomaly detection means 2, of comparing 107, by statistical analysis, the corresponding data of the analysis database and the corresponding data of the reference database. Preferably, each of the quantities is subjected to said statistical analysis. Alternatively, part of the quantities is considered irrelevant, taking into account for example information relating to a current operating state of the vehicle of interest.
[0055] According to an optional implementation, if the number of data relating to a specific quantity in the data sets is not large enough, the statistical analysis on said specific quantity is not taken into account because it is considered statistically invalid.
[0056] In a non-limiting manner, this statistical analysis can be formed by a Quantile-Quantile type analysis (also called QQ plot).
[0057] In a different implementation, the statistical analysis may be formed by a Kolmogorov-Smirnov type analysis.
[0058] In any case, statistical analysis makes it possible to identify abnormal values, without a priori knowledge of the expected values. The reference population, formed by the data from the reference database, is not fixed. On the contrary, it is updated as the vehicle fleet is used, to provide the most relevant anomaly detection possible.
[0059] The method 100 further comprises a step, executed by the anomaly detection means 2, of detecting 108 an anomaly on the vehicle of interest, when, for at least one of the quantities compared by statistical analysis, the corresponding data of the analysis database deviate beyond a predetermined threshold relative to the corresponding data of the reference database. The threshold predetermined threshold is advantageously calculated in a preliminary step. It can be based on a priori knowledge relating to acceptable margins of error, in particular when the quantity follows a known probability distribution such as a Gaussian distribution. Alternatively, the predetermined threshold can be based on feedback.
[0060] Thus, the data of the vehicle of interest is compared to the data of similar vehicles in order to determine whether the vehicle of interest has an anomaly. In other words, the behavior of other vehicles is used to determine whether the vehicle of interest has an anomaly.
[0061] The centralization of the data by the grouping means 3 makes it possible to perform calculations on large-scale data with virtual machines whose size allows the processing of this data on a daily and automated basis. The automation and daily processing of the data make it possible to improve the detection of anomalies and also make it possible to detect anomalies occurring over time. Thus, when this method is implemented on production vehicles, it is possible to detect anomalies as the vehicle is used. When an anomaly is detected, the driver can be notified so that he can bring his vehicle to a monitoring center to have it checked and / or repaired.
[0062] Optionally, the method further comprises a step, not shown, of transmitting an alert message from the means for detecting an anomaly 2 to the vehicle of interest, or to a computer of a control center, when an anomaly is detected on said vehicle of interest.
Claims
Claims
1. Method (100) for detecting an anomaly on a device of interest belonging to a fleet of devices, said method (100) comprising the following steps, executed by a system for detecting an anomaly (D, of • Acquire (102) a new data set, in which each data item represents the value of a respective quantity relating to said device of interest; • Update (103) a so-called main database, to add thereto said new acquired data set in association with information relating to at least one descriptive characteristic of said device of interest at the time of acquisition of said new data set;• From said main database, extract (105) a so-called reference database, said reference database grouping together data sets associated with devices of said fleet of devices having said at least one descriptive characteristic, excluding data sets associated with said device of interest; • From said main database, extract (106) a so-called analysis database grouping together data sets associated with said device of interest and with said at least one descriptive characteristic;• For at least one quantity, compare (107), by statistical analysis, the corresponding data of said analysis database and corresponding data of said reference database. • Detect (108) an anomaly on said device of interest, when, for at least one of said quantities, the corresponding data of said analysis database deviate beyond a predetermined threshold relative to the corresponding data of said reference database.;
2. Method (100) according to the preceding claim, characterized in that it comprises a preliminary step of construction (101) of the main database, this preliminary step consisting of grouping a plurality of collected data sets, said data sets having been determined by the devices of the fleet of devices, each data representing the value of a respective quantity relating to a device of said fleet of devices, each data set of a device of the fleet of devices being associated with at least one descriptive characteristic of said device at the time of acquisition of said data set.
3. Method (100) according to any one of the preceding claims, characterized in that at least one data item, among the data of the new acquired data set, is measured by a sensor embedded on the device of interest, or is calculated, by a computer embedded on said device of interest, from at least two data items acquired by two sensors embedded on said device of interest.
4. Method (100) according to any one of the preceding claims, characterized in that the quantities associated with the new acquired data set are decorrelated from each other.
5. Method (100) according to any one of the preceding claims, characterized in that the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set comprises the model of said device of interest, the reference database grouping together data sets associated only with devices of this same model.
6. Method (100) according to the preceding claim, characterized in that the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set further comprises a version of software for controlling functionalities of the device of interest, installed on a computer of the device of interest at said time of acquisition, the reference and analysis databases grouping together data sets associated only with devices of this same model and in which the same version of the software is installed.
7. Method (100) according to claim 5 or 6, characterized in that the at least one descriptive characteristic of the device of interest at the time of acquisition of the new data set further comprises an operating state of the device of interest at said time of acquisition, the reference and analysis databases grouping together data sets associated only with devices having, at least, the same model and said operating state.
8. Method (100) according to any one of the preceding claims, characterized in that the reference and analysis databases re- group together data sets associated with acquisition times included in a time interval of predetermined duration and prior to the acquisition time of the new data set, and / or data sets associated with a maximum distance traveled by the device of interest up to the acquisition time of the new data set.
9. Method (100) according to any one of the preceding claims, characterized in that it comprises a step of weighting (104) the data sets.
10. Method (100) according to any one of the preceding claims, characterized in that the new data set, acquired by the anomaly detection system (1), comes directly from means for transmitting new data sets that the device of interest comprises.
11. Method (100) according to any one of the preceding claims, characterized in that the fleet of devices is formed by a fleet of motor vehicles and the device of interest is formed by a motor vehicle of interest.
12. Computer program product comprising instructions which, when the program is executed by the anomaly detection system (1), cause the latter to implement the steps of the method (100) according to any one of the preceding claims.
Citation Information
Patent Citations
Methods for the external detection of anomalies in motor vehicles
DE102021204924A1
Monitoring of vehicle health based on historical information
US20030208309A1
Centralized facility and intelligent on-board vehicle platform for collecting, analyzing and distributing information relating to transportation infrastructure and conditions
US7421334B2
