Method of diagnosis of a motor vehicle sensor

Abstract

The invention relates to a method of diagnosis of a motor vehicle (10) sensor (16, 17) suitable for detecting road infrastructures (101, 102, 103, 104), said motor vehicle comprising a means of communication (18) adapted for communicating with a remote server (50) and a computer (14) connected to the sensor and to the means of communication. According to the invention, the method of diagnosis comprises steps in the course of which: a) the computer identifies an infrastructure and assigns to it an effective score, which relates to the visibility of this infrastructure, b) the remote server acquires a reference score which is assigned to said infrastructure and which relates to the visibility of this infrastructure, and c) the effective score and the reference score are compared so as to deduce therefrom a state of operation of the sensor.

Description

TECHNICAL FIELD TO WHICH THE INVENTION RELATES

[0001] The present invention relates generally to driving aids for motor vehicles.

[0002] It relates more specifically to a diagnostic process for a sensor in a motor vehicle.

[0003] It applies to motor vehicles equipped with a sensor adapted to detect road infrastructure, a means of communication adapted to communicate with a remote server and a computer connected to the sensor and the means of communication. TECHNOLOGICAL BACKGROUND

[0004] To make driving a motor vehicle easier and safer, we want to provide the driver with information (for example, the maximum speed allowed on the road) and warn them of problems (for example, when the vehicle deviates from its trajectory).

[0005] To develop this information and these reports, it is known to use sensors adapted to detect and interpret road infrastructure (signs, position of continuous and broken lines, ...).

[0006] The reliability of this information and these reports depends largely on the quality of road infrastructure detection. Unfortunately, this detection is affected by two problems.

[0007] The first problem is that infrastructure deteriorates over time, for example due to weather conditions, sun exposure, and the number of cars driving on it. It is therefore necessary to monitor its condition in order to replace it before it becomes unusable.

[0008] Currently, this monitoring work is carried out by individuals, who are employed to drive on roads and to complete a database in which each infrastructure is rated, according to its readability and therefore its need or not to be repaired.

[0009] The second problem is that the sensors sometimes malfunction.

[0010] While it is relatively easy to detect a complete sensor failure, it remains more difficult to detect a problem affecting the sensor without causing it to shut down completely. Such a problem, however, risks compromising the quality of road infrastructure detection.

[0011] For example, movement of the sensor relative to its mount can affect the quality of road infrastructure detection, even if it is not easily detectable. Dust particles stuck to the sensor lens can also affect detection quality.

[0012] Document US2013033603 discloses the monitoring of traffic sign status. A traffic sign is detected using image recognition from a vehicle camera. The image is sent to a server if the recognition result differs from recorded information regarding the expected presence or absence of traffic signs. A person can then visually verify the image.

[0013] Document US2014071281 discloses the diagnosis of a traffic camera using current and reference image analysis.

[0014] In document US2010099353, a sensor on a first vehicle detects an obstacle at time t and compares this detection with that of the same obstacle detected at time t+1 by a second vehicle. If the difference is large, the detection by the second vehicle is incorrect. The detection made by the second vehicle is recorded in the first vehicle for comparison. SUBJECT OF THE INVENTION

[0015] In order to remedy the aforementioned drawback of the prior art, the present invention proposes a statistical method for diagnosing the proper functioning of sensors.

[0016] More specifically, the invention proposes a method for diagnosing a sensor of a motor vehicle, which includes steps defined in claim 1.

[0017] Thus, the invention takes advantage of the fact that the remote server has a database in which ratings assigned to infrastructure are stored, based on their visibility. The invention then proposes to compare this rating with a rating that the computer itself will have calculated based on the difficulties it encountered in interpreting the road infrastructure.

[0018] So, if the two scores are very close, we can deduce that the sensor is functioning correctly.

[0019] Conversely, if these two notes are very different (over a predetermined period of time or for a number of occurrences identified as significant), it can be deduced that the sensor has a significant malfunction.

[0020] Finally, if these two notes are slightly different, and this slight difference is noted for all the infrastructures encountered, we can deduce that the sensor has a slight malfunction.

[0021] We can then monitor the evolution of this difference between the two scores in order to control the drift in reliability of the sensor.

[0022] Other advantageous and non-limiting features of the diagnostic method according to the invention are as follows: In step b), the remote server transmits the reference score to the computer and, in step c), the operating state of the sensor is determined by the computer; at the end of step a), the computer sends a request to the remote server containing an identifier of the identified infrastructure and / or the geographical coordinates of the motor vehicle and, in step b), the remote server acquires the reference score associated with said infrastructure, taking into account said identifier and / or said geographical coordinates; after step a), there is a step of transmitting the actual score to the remote server and a step of calculating by the remote server a new reference score based on said actual score; said calculation step is implemented only if the difference between the actual score and the reference score is less than a predetermined threshold;Steps b) and c) are implemented for only part of the infrastructure identified by the computer; steps b) and c) are implemented at regular intervals; prior to step b), a step is planned to determine an indicator relating to the meteorological and / or glare conditions of the sensor, and steps b) and c) are implemented only when said indicator relates to satisfactory meteorological and / or glare conditions; prior to step b), a step is planned to determine the time of day, and steps b) and c) are implemented only when the time is within a determined interval;Prior to step a), a step is planned to determine an indicator relating to meteorological conditions and / or sensor glare conditions and / or the time of day, and the reference score acquired by the remote server is a function of the value of said indicator. DETAILED DESCRIPTION OF A PROJECT EXAMPLE

[0023] The description that follows, with regard to the attached drawing, given as a non-limiting example, will make it clear what the invention consists of and how it can be carried out.

[0024] On the attached drawing, the figure 1 is a schematic perspective view of a motor vehicle traveling on a road and a remote server on that road.

[0025] As this shows figure 1 , the motor vehicle 10 is here a car with four wheels 11. Alternatively, it could be a motor vehicle with two or three wheels, or more wheels.

[0026] Typically, this motor vehicle 10 includes a chassis which supports in particular a powertrain 12 (namely an engine and means of transmitting torque from the engine to the drive wheels), bodywork elements and interior elements.

[0027] The motor vehicle 10 also includes an electronic control unit (or ECU, from the English "Electronic Control Unit"), referred to here as computer 14.

[0028] This calculator 14 includes a processor and a storage unit, for example a non-volatile rewritable memory or a hard disk.

[0029] The memory unit stores, in particular, computer programs including instructions whose execution by the processor allows the computer to implement the process described below.

[0030] For the implementation of this process, the computer 14 is connected to various pieces of equipment in the motor vehicle 10.

[0031] Among these equipment, the motor vehicle 10 includes at least one sensor 16, 17 and means of communication 18. It also includes here a means of geolocation 15.

[0032] As depicted on the figure 1 The motor vehicle 10 is equipped with several sensors adapted to acquire information relating to road infrastructure.

[0033] The motor vehicle 10 thus includes a camera 16 which is located at the front of the vehicle and is oriented forwards, so that it can acquire images of the infrastructure of Route 100.

[0034] The motor vehicle 10 also includes two LIDAR sensors (acronym for the English expression "light detection and ranging"), which here take the form of two laser remote detectors 17. These two laser remote detectors 17 are located at the front of the vehicle and are oriented in oblique directions, so that they can determine the shape of the infrastructure of the road 100, on either side of the motor vehicle.

[0035] Since such a laser remote detector is well known to those skilled in the art, it will not be described in detail here. It will simply be stated that it is a remote measurement system whose operation is based on the emission of a beam of light by a transmitter and the analysis of the properties of the light beam reflected back from the obstacle to its transmitter.

[0036] Therefore, a LiDAR sensor is able to detect a foreign object on the road, such as a tire left on the road or a branch fallen onto the road. It is also able to detect snow on the road.

[0037] It could of course be expected that the vehicle will have more LIDAR sensors, for example located on the sides and rear of the vehicle.

[0038] Alternatively, the sensors could be different. These could include SONAR or RADAR sensors. They could be positioned differently on the vehicle, for example, to acquire images of the road from the rear view of the vehicle.

[0039] The geolocation means 15 is intended to determine the position of the vehicle and / or that of the infrastructure targeted by the sensors 16, 17.

[0040] If the road were equipped with geolocation modules distributed along its length, one could consider that the means of geolocation is formed by an antenna adapted to communicate with these geolocation modules.

[0041] Here, we will consider that the geolocation means 15 is formed by a GPS antenna, allowing the geographical coordinates of the motor vehicle 10 to be determined.

[0042] As will be explained later in this presentation, the geolocation means may also potentially use the signals emitted by sensors 16, 17 to determine more precisely the position of the motor vehicle 10 on the road 100 (which lane it is in, how far from each infrastructure it is located...).

[0043] Finally, the means of communication 18 is designed to communicate with a remote server 50, via a relay antenna 51. It is more specifically designed here to connect to a mobile telephone network which includes in particular the said relay antenna 51 and a gateway for connection to a public network (for example the Internet network).

[0044] The remote server 50 is then also connected to the public network so that the computer 14 of the motor vehicle 10 and the remote server 50 can communicate and exchange data via the mobile phone network.

[0045] This remote server 50 stores here a database register comprising a plurality of records each associated with a road infrastructure of the automotive network.

[0046] Each record then stores an identifier for this infrastructure, as well as the geographic coordinates of this infrastructure and at least one note relating to the state of this infrastructure.

[0047] The form of this identifier will be detailed later in this presentation.

[0048] The rating could have been recorded in the database by an operator responsible for monitoring the condition of the infrastructure. Such an operator would then be employed to drive on the roads, to observe the condition of the infrastructure and to assign it a rating (which they would then record in the corresponding entry in the database register).

[0049] However, in the embodiment described here, each record will contain not one but several notes relating to the state of this infrastructure. These notes will have been previously transmitted to the remote server by vehicles automatically monitoring the state of the infrastructure. The communication protocol for these notes will be described in detail later in this presentation.

[0050] In this embodiment, the remote server 50 is then able to calculate a reference score NO relating to the state of each infrastructure, for example by determining the average of the scores stored in the record.

[0051] Motor vehicle 10 is shown on the figure 1 such as driving on a route 100 comprising different infrastructures 101, 102, 103, 104.

[0052] Here, for illustrative purposes, this Route 100 has two traffic lanes 105 separated from each other by a solid line 101. The lateral edges of this Route 100 are formed by the shoulder 104. Broken lines 102 mark the position of these shoulders 104. A traffic sign 103 is also shown on the edge of Route 100.

[0053] The invention relates to a method implemented by the computer 14 of the motor vehicle 10 and by the remote server 50 to diagnose the operating state of each sensor 16, 17 of the motor vehicle 10.

[0054] The invention proposes to verify that the sensor detects road infrastructure in the same way as other vehicles traveling on the road 100. In other words, the invention proposes to verify that the score assigned by the vehicle's computer 14 to each infrastructure (depending on whether it considers this infrastructure to be in good condition or not) corresponds substantially to the reference score NO stored in the remote server 50.

[0055] More specifically, according to a particularly advantageous feature of the invention, the diagnostic process comprises three main steps, including: a first step a) during which the computer 14 identifies an infrastructure 101, 102, 103, 104 and assigns it an effective note N1, which is relative to the visibility of this infrastructure 101, 102, 103, 104, a second step b) during which the remote server 50 searches in its database for the record which corresponds to said infrastructure 101, 102, 103, 104 and then determines the reference note NO assigned to this infrastructure 101, 102, 103, 104, and a third step c) during which the effective note N1 and the reference note NO are compared to deduce an operating state of the sensor 16, 17.

[0056] More specifically, during the first stage, camera 16 acquires an image of road 100 on which each of the infrastructures appears, namely the signpost 103, the continuous line 101 and the broken lines 102.

[0057] Laser remote detectors 17 allow the shape and position of the shoulders 104 to be determined.

[0058] Alternatively, the sensors equipping the motor vehicle 10 could acquire more information (including the presence of a pothole in the road), but for the sake of clarity in this presentation, only this information will be considered here.

[0059] Computer 14 then uses the signals it receives from these sensors 16, 17 to determine an effective score N1 relating to the condition of each infrastructure 101, 102, 103, 104 of route 100.

[0060] At this stage, it should be noted that each note will be assigned to a particular infrastructure, as seen by a particular sensor. In other words, if several sensors detected the same infrastructure, the state of that infrastructure would be noted multiple times in order to determine how that infrastructure is seen by each sensor considered separately.

[0061] More specifically, the computer 14 uses the image acquired by the camera 16 and the shapes seen by the laser remote detectors 17 in the following way.

[0062] Using a classic image analysis, he identifies the continuous lines 101 and discontinuous lines 102 and the road sign 103 in the acquired image.

[0063] It also identifies, in the signals received from the laser remote detectors 17, by a classic signal analysis, the shoulders 104.

[0064] An identifier is assigned to each type of infrastructure to facilitate its identification. This identifier will preferably be chosen according to the type of infrastructure. Thus, one could plan to assign the identifier #101 to all solid lines, the identifier #102 to all broken lines, the identifier #103 to all traffic signs bearing a "danger" symbol, and the identifier #104 to all shoulders.

[0065] Calculator 14 will then assign an effective N1 rating to each of the identified infrastructures.

[0066] This effective score N1 can be expressed as a probability level that the infrastructure has been correctly identified, or in any other conceivable form. Here, the effective score N1 will be determined as follows.

[0067] Calculator 14 determines the variations in width of the continuous line 101. If the width of this continuous line 101 varies, meaning that the continuous line 101 is likely degraded, it assigns a reduced effective score N1 (e.g., equal to 1) to the readability of the continuous line 101. Otherwise, it assigns a high effective score N1 (e.g., equal to 2 or 3).

[0068] Calculator 14 then determines the variations in width and length of each stroke of the dashed lines 102. If the width or length of these strokes varies, meaning that the corresponding dashed line 102 is likely degraded, it assigns a reduced effective score N1 (e.g., equal to 1) to the legibility of the dashed line 102. Otherwise, it assigns a high effective score N1 (e.g., equal to 2 or 3).

[0069] Using image recognition software stored in its memory unit, which stores the various symbols that may appear on road signs, the computer 14 determines the symbol displayed on the road sign 13. If it cannot do so (meaning that the symbol is partially erased or hidden by vegetation), it assigns a reduced effective score N1 (for example, equal to 0) to the legibility of the road sign 13. Otherwise, and depending on the degree of certainty of the symbol recognition, it assigns a higher effective score N1 (for example, equal to 1, 2, or 3).

[0070] The calculator 14 finally determines the variations in distances between the dashed lines 102 and the shoulders 104 and identifies any irregularities in these shoulders 104. If these distances vary and / or if the shoulders 104 are irregular, indicating that the shoulders 104 are likely in poor condition, it assigns the shoulders 104 a reduced effective score N1 (for example, equal to 0 or 1). Otherwise, it assigns a higher effective score N1 (for example, equal to 2 or 3).

[0071] During the second step, the computer 14 sends a request to the remote server 50 so that the latter transmits the reference note N0 associated with each infrastructure 101, 102, 103, 104.

[0072] This query contains the identifier of each of the identified infrastructures 101, 102, 103, 104 and the geographical coordinates of the motor vehicle 10 recorded by the geolocation means 15. It may also contain other data, including the direction of travel of the vehicle on the road (obtained from the positions of the vehicle successively recorded by the geolocation means 15) or the traffic lane 105 on which the vehicle is traveling, here the left lane (obtained from the image acquired by the camera 16).

[0073] This data allows the remote server 50 to identify the infrastructures seen by sensors 16, 17 and therefore to find in its database register the records corresponding to these infrastructures.

[0074] Then, the remote server 50 determines the reference notes N0 associated with these infrastructures 101, 102, 103, 104, here by averaging the notes stored in each record found.

[0075] Then, it sends these reference notes N0 back to the computer 14 of the motor vehicle 10.

[0076] During the third step, the calculator 14 determines, for each infrastructure, the difference ΔN between the reference note N0 and the effective note N1 (in absolute value).

[0077] It could be predicted that if the difference ΔN between these two scores exceeds a predetermined threshold (for example, equal to 2), the computer 14 would deduce a malfunction of the corresponding sensor. Indeed, in this case, it would mean that the sensor was unable to detect the infrastructure in question in the same way as other vehicles (those that transmitted data to the remote server, which allowed the calculation of the reference score NO).

[0078] However, here, before deducing such a defect, the calculator 14 will rather repeat the aforementioned steps for different infrastructures.

[0079] If, for each infrastructure detected by the sensor in question, the difference ΔN between the reference rating N0 and the actual rating N1 exceeds the predetermined threshold, the computer deduces a sensor malfunction. It then stores an error code in its memory unit, which will allow a technician to identify this fault.

[0080] Otherwise (i.e., if the sensor detects infrastructure in the same way as other vehicles), the computer deduces that the sensor is working correctly.

[0081] It is also possible to analyze this difference ΔN more precisely. The computer can store successively calculated ΔN differences for a sensor in its memory unit and observe how this difference changes over time. If it observes an increasing trend in this difference ΔN, it can deduce a slight malfunction of the sensor. It can also anticipate when the sensor will be considered faulty, thus predicting when it will need to be replaced.

[0082] In the embodiment considered here, it is planned to transmit the actual grade N1 to the remote server 50. This transmission step can be done during the second step, when the computer transmits a request to the remote server 50.

[0083] Therefore, the remote server 50 can store this effective N1 note in the record associated with the infrastructure in question, in order to complete its database.

[0084] By completing its database, the remote server 50 will be able to obtain a large quantity of notes assigned to each infrastructure, which will allow the value of the reference note N0 to be refined.

[0085] It will be observed that the greater the number of notes transmitted to the remote server 50, the closer to reality will be the reference note NO, even if some of the vehicle sensors were faulty and even if weather conditions sometimes distorted the data collected by the sensors.

[0086] Alternatively, the remote server 50 could be configured to record this actual score N1 in its database only if the difference ΔN between the reference score N0 and the actual score N1 is less than the predetermined threshold. In this way, if the sensor has a defect, the actual score N1 calculated using that sensor will not be recorded in the database and therefore will not distort the calculation of the reference score N0.

[0087] Furthermore, it can be foreseen that the remote server 50 will only record this effective note N1 if the weather conditions are sufficiently good or if the sensor 16, 17 is not dazzled by the sun or by any light source or if it is still daytime.

[0088] To achieve this, the computer can determine the value of a weather condition indicator (1 if sunny, 2 if cloudy, 3 if snowy, etc.) and the value of a glare indicator (1 if dazzled, 0 otherwise), and transmit these values ​​to the remote server 50, so that the server only records the actual rating N1 if these values ​​are satisfactory (for example, if it is not raining or snowing, if there is no fog, and if the sensor is not dazzled). It can also be stipulated that the actual rating N1 is only recorded if it is still daylight at the vehicle's location (taking into account the time and sunrise / sunset times at the vehicle's location).

[0089] Alternatively, the remote server 50 could always record the actual score N1 in its database, regardless of weather and glare conditions or the time of day, but it could associate this actual score N1 with the weather and glare conditions encountered and the time of day. More specifically, the remote server 50 could store the actual score in a sub-record corresponding to the weather conditions encountered, the degree of glare from the sensor, and whether it is day or night. In this variant, the request transmitted by the control unit 14 to the remote server 50 would include the aforementioned indicators. In this way, the reference score N0 returned by the remote server 50 to the control unit 14 would be equal to the average of the scores stored in the sub-record corresponding to the weather and / or lighting (day or night) and / or glare conditions encountered by the vehicle.

[0090] Finally, it should be noted that the second and third steps of the aforementioned diagnostic process can be implemented for each detected infrastructure or at each time step.

[0091] Alternatively, to avoid this process consuming a large part of the processor's computing power, the second and third steps can be performed less frequently.

[0092] For example, they can be carried out at regular intervals (for example, once a day, or after each time the vehicle is started).

[0093] For example, they may only be carried out if the weather and / or glare conditions of the sensor 16, 17 are very satisfactory (in sunny weather, when the sensor is not glareed).

Claims

1. Method for diagnosing a sensor (16, 17) of a motor vehicle (10) designed to detect road infrastructures (101, 102, 103, 104), said motor vehicle (10) comprising a means of communication (18) designed to communicate with a remote server (50) and a computer (14) connected to the sensor (16, 17) and to the means of communication (18), characterized in that it comprises steps during which: a) the computer (14) identifies one of the infrastructures (101, 102, 103, 104) and assigns it an effective score (N1), which relates to the degree of visibility of this infrastructure (101, 102, 103, 104), the visibility being obtained by the sensor b) the remote server (50) acquires a reference score (N0) which is assigned to said infrastructure (101, 102, 103, 104) and which relates to the degree of visibility of this infrastructure (101, 102, 103, 104), the visibility being obtained from records of scores previously communicated to the remote server, and c) the effective score (N1) and the reference score (N0) are compared so as to deduce therefrom a state of operation of the sensor (16, 17).

2. Diagnostic method according to the preceding claim, in which, in step b), the remote server (50) transmits the reference score (N0) to the computer (14) and, in step c), the state of operation of the sensor (16, 17) is determined by the computer (14).

3. Diagnostic method according to the preceding claim, in which, at the end of step a), the computer (14) sends a request to the remote server (50) containing an identifier of the infrastructure (101, 102, 103, 104) identified and / or the geographical coordinates of the motor vehicle (10) and, in step b), the remote server (50) acquires the reference score (N0) associated with said infrastructure (101, 102, 103, 104), taking into account said identifier and / or said geographical coordinates.

4. Diagnostic method according to one of the preceding claims, in which, after step a), a step of transmitting the effective score (N1) to the remote server (50) and a step of the remote server (50) calculating a new reference score (N0) as a function of said effective score (N1) are provided.

5. Diagnostic method according to the preceding claim, in which said calculation step is implemented only if the difference between the effective score (N1) and the reference score (N0) is below a predetermined threshold.

6. Diagnostic method according to one of the preceding claims, in which the steps b) and c) are implemented for only some of the infrastructures identified by the computer (14).

7. Diagnostic method according to Claim 6, in which steps b) and c) are implemented at regular intervals.

8. Diagnostic method according to Claim 6, in which, prior to step b), a step of determining an indicator relating to the weather and / or glare conditions of the sensor (16, 17) is provided, and in which steps b) and c) are implemented only when said indicator relates to satisfactory weather and / or glare conditions.

9. Diagnostic method according to Claim 6, in which, prior to step b), a step of determining the time of day is provided, and steps b) and c) are implemented only when the time is contained in a given interval.

10. Diagnostic method according to one of Claims 1 to 7, in which, prior to step a), a step of determining an indicator relating to the weather conditions and / or to the glare conditions of the sensor (16, 17) and / or to the time of day is provided, and in which the reference score (N0) acquired by the remote server (50) is a function of the value of said indicator.

Images