Method for calibrating the activation time of at least one vehicle-mounted restraint system

Abstract

The present invention relates to a method and device for calibrating the activation time of at least one vehicle-mounted restraint system. The method obtains (32) a predicted activation time for said at least one restraint system from the output of a activation time prediction model trained on collision scenarios (31). The method adjusts (34, 35) at least one parameter limit value for at least one vehicle-mounted sensor and at least one collision scenario and retrains (36) the activation time prediction model trained on a difference (33) between the predicted activation time and an expected activation time for each collision scenario. Figure for the abstract: Figure 3

Description

Title of the invention: Method for calibrating the triggering time of at least one vehicle-mounted restraint system. Technical field

[0001] The present invention relates to the field of vehicle safety and more particularly to the calibration and prediction of activation times of vehicle-mounted restraint devices. Technological background

[0002] Methods for predicting the deployment time of vehicle restraint systems, such as seat belts or airbags, are known. These methods predict deployment times based on predetermined and fixed rules. These rules are determined by analyzing the results of vehicle collision scenario tests involving obstacles. These results may consist of measuring values ​​of vehicle sensor parameters, such as vehicle speed and / or acceleration, or door pressure, and comparing these measured parameter values ​​with predetermined limit values. These results make it possible, for example, to determine whether a vehicle collision with an obstacle is severe enough to warrant the deployment of the vehicle's restraint systems.

[0003] These prediction methods lack flexibility and cannot easily adapt to new collision scenarios that were not used to determine these rules. Furthermore, the rules of these prediction methods are determined for values ​​from vehicle-mounted sensors, and any new sensor addition requires a new phase of rule determination. In addition, developing these prediction methods is time-consuming because they most often require several manual adjustments of the limit values ​​to ensure correct activation of the vehicle's on-board restraint systems. Summary of the present invention

[0004] One object of the present invention is to solve at least one of the problems of the technological background described above.

[0005] According to a first aspect, the present invention relates to a method for calibrating the triggering time of at least one vehicle-mounted restraint system, said method being implemented by at least one processor and comprising the following steps: - obtaining at least one collision scenario from a memory, each collision scenario comprising a first data point and a second data point, the first data point being representative of limit values ​​of parameter of at least one on-board sensor of the vehicle, each limit value defining a limit to a parameter value of said at least one sensor beyond which said at least one vehicle restraint device is likely to be triggered and, the second data point being representative of an expected triggering time of said at least one vehicle restraint device when the limit values ​​of parameter are at least reached; - obtaining a predicted triggering time of said at least one restraint device at the output of a triggering time prediction model when the first data of each collision scenario is presented as input to the triggering time prediction model; - obtaining a third data point representing a difference between the predicted triggering time for each collision scenario and the second data point for each collision scenario; - adjusting at least one limit value of the parameter of said at least one sensor as a function of the third data obtained; - if said at least one parameter limit value is adjusted, then the first data point of at least one collision scenario is adjusted according to said at least one adjusted parameter limit value; and - if at least one collision scenario is fitted then learning of the trigger time prediction model trained on said at least one fitted collision scenario.

[0006] Such a method uses a vehicle restraint system activation time prediction model to predict the activation time of these restraint systems when values ​​of vehicle onboard sensor parameters exceed limit values. Collision scenarios are obtained. These scenarios are defined by limit values ​​of sensor parameters and restraint system activation times. These collision scenarios are used to estimate the accuracy of the restraint system activation time prediction provided by the activation time prediction model. If the prediction accuracy is insufficient, the method allows the limit values ​​of the sensor parameters to be adjusted, the collision scenarios to be updated according to the adjusted limit values, and the parameters of the activation time prediction model to be updated.

[0007] The method is flexible because it can adapt to any new collision scenario and / or any new sensor used for predicting the triggering time of restraint devices. The method is scalable because it allows the model to be adapted to This prediction method is used when the model is deployed in a vehicle to improve its accuracy when the prediction quality is insufficient. The process is faster than conventional methods for determining limit values.

[0008] The method thus makes it possible to improve the prediction of triggering time by reducing premature or late triggering.

[0009] The method makes it possible to increase road safety because it provides predictions of optimal and appropriate triggering times of restraint devices, thus increasing the protection of vehicle occupants in the event of a collision.

[0010] According to a particular and non-limiting embodiment of the present invention, the trigger time prediction model is a neural network.

[0011] This embodiment is advantageous because the use of a neural network improves the accuracy of calibrating the triggering time of restraint devices, reducing the error to ~0.1 ms, thus exceeding the performance of conventional rule-based prediction methods. Furthermore, the use of a neural network shortens the development time of safety systems compared to traditional prediction methods.

[0012] According to a particular and non-limiting embodiment of the present invention, the neural network is trained during a learning phase using training data comprising first and second data from learning collision scenarios.

[0013] According to a particular and non-limiting embodiment of the present invention, each learning collision scenario comprises a first data point and a second data point, the first data point being representative of parameter limit values ​​of at least one vehicle-mounted sensor, each limit value defining a limit to a parameter value of at least one vehicle-mounted sensor beyond which at least one vehicle restraint device is likely to be triggered and, the second data point being representative of an expected triggering time of said at least one vehicle restraint device when the parameter limit values ​​are at least reached.

[0014] According to a particular and non-limiting embodiment of the present invention, the first and second data of the learning collision scenarios correspond to data obtained from real vehicle collision scenarios with obstacles and / or collision scenarios artificially generated by road safety experts.

[0015] According to a particular and non-limiting embodiment of the present invention, the first data includes a limit value of pressure on a door of the vehicle.

[0016] According to a particular and non-limiting embodiment of the present invention, the first data further includes a deceleration limit value of the vehicle.

[0017] According to a second aspect, the present invention relates to a device for calibrating the triggering time of on-board restraint devices of a vehicle, the device comprising a memory associated with a processor configured for implementing the steps of the process according to the first aspect of the present invention.

[0018] According to a third aspect, the present invention relates to a vehicle, for example of the automobile type, comprising a device as described above according to the second aspect of the present invention.

[0019] According to a fourth aspect, the present invention relates to a computer program which includes instructions adapted for carrying out the steps of the process according to the first aspect of the present invention, in particular when the computer program is executed by at least one processor.

[0020] Such a computer program may use any programming language, and be in the form of source code, object code, or an intermediate code between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0021] According to a fifth aspect, the present invention relates to a computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the process according to the first aspect of the present invention.

[0022] On the one hand, the recording medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, RAM, CD-ROM or a microelectronic circuit-type ROM, or a magnetic recording means or a hard disk drive.

[0023] On the other hand, this recording medium can also be a transmissible medium such as an electrical or optical signal, such a signal being able to be transmitted via an electrical or optical cable, by conventional or radio frequency, by self-directing laser beam, or by other means. The computer program according to the present invention can, in particular, be downloaded from an Internet-type network.

[0024] Alternatively, the recording medium may be an integrated circuit in which the computer program is incorporated, the integrated circuit being adapted to execute or to be used in the execution of the process in question. Brief description of the figures

[0025] Other features and advantages of the present invention will become apparent from the description of the particular and non-limiting embodiments of the present invention below, with reference to the attached Figures 1 to 3, in which:

[0026] [Fig-1] schematically illustrates part of a vehicle's passenger compartment, according to a example of a particular embodiment of the present invention;

[0027] [Fig.2] schematically illustrates a device configured to calibrate a triggering time of on-board restraint devices of a vehicle, according to a particular and non-limiting embodiment of the present invention;

[0028] [Fig.3] schematically illustrates a flowchart of the different steps of a calibration process for the triggering time of at least one on-board restraint device of a vehicle, according to a particular and non-limiting embodiment of the present invention. Description of examples of achievements

[0029] A method and device for calibrating the triggering time of at least one vehicle-mounted restraint device will now be described in the following with joint reference to Figures 1 to 3. The same elements are identified with the same reference signs throughout the following description.

[0030] The terms "first," "second" (or "firsts," "seconds"), etc., are used in this document by arbitrary convention to allow for the identification and distinction of different elements (such as operations, means, etc.) implemented in the embodiments described below. Such elements may be distinct or correspond to a single element, depending on the embodiment.

[0031] Fig. 1 schematically illustrates part of the passenger compartment of a vehicle 10, according to a particular and non-limiting embodiment of the present invention.

[0032] Vehicle 10 corresponds, for example, to a vehicle with an internal combustion engine, with electric motor(s), or even a hybrid vehicle with an internal combustion engine and one or more electric motors. Vehicle 10 thus corresponds, for example, to a land vehicle, for example a car, a truck, a bus.

[0033] According to a particular embodiment, the vehicle 10 may carry one or more on-board ADAS systems, each controlled by one or more computers, for example, a navigation system, a cruise control system, and / or a lane change control system. These computers, together with the IVI computer, form, for example, a multiplexed architecture for providing various services useful for the proper functioning of the vehicle and for assisting the driver and / or passengers in controlling the vehicle 10 by controlling the on-board system(s) in the vehicle 10. The computers communicate and exchange data with each other via one or more buses. computer systems, for example a CAN data bus communication bus (from the English "Controller Area Network" or in French "Réseau de contrôlers"), CAN FD (from the English "Controller Area Network Flexible Data-Rate" or in French "Réseau de contrôlers à débit de données flexible"), FlexRay (according to the ISO 17458 standard), LIN (from the English "Local Interconnect Network" or in French "Réseau interconnecté local") or Ethernet (according to the ISO / IEC 802-3 standard).

[0034] The vehicle 10 further includes at least one restraint device DRk (k=l to K) such as, for example, seat belts (not shown) or airbags 12 which can be housed in the dashboard 11 and / or in the steering wheel 13.

[0035] The vehicle 10 further comprises a DRk restraint system that triggers the DRk restraints after a triggering time that is a function of the times at which parameter values ​​of sensors Cj (j=l to M) on board the vehicle 10 exceed limit values. These parameter values ​​exceed their limit values ​​during collisions of the vehicle 10 with an obstacle.

[0036] According to a particular and non-limiting embodiment of the present invention, one of the sensors Cj may be a pressure sensor 14 located at the level of a door of the vehicle 10 and adapted to provide a pressure value on the door which may be less than, greater than or equal to a pressure limit value during a collision of the vehicle 10 with an obstacle.

[0037] According to a particular and non-limiting embodiment of the present invention, one of the sensors Cj may be an accelerometer configured to obtain a deceleration value (negative acceleration value) of the vehicle that may be less than, greater than, or equal to a vehicle deceleration limit value. Several measurement points may be implemented in the vehicle 10 to obtain this deceleration value.

[0038] The present invention uses a DRk restraint device triggering time prediction model.

[0039] According to a particular and non-limiting embodiment of the present invention, the trigger time prediction model can be a trained neural network such as, for example, a recurrent neural network (RNN) with long short-term memory (LSTM) or a gate recurrent unit (GRU) neural network.

[0040] The neural network is trained during a learning phase. During this learning phase, the parameters of the neural network are determined from a large amount of training data.

[0041] According to a particular and non-limiting embodiment of the present invention, the training data may include first data and second data from learning collision scenarios.

[0042] According to a particular and non-limiting embodiment of the present invention, each learning collision scenario may include a first data point and a second data point. The first data point represents parameter limit values ​​for at least one vehicle-mounted sensor, each limit value defining a threshold beyond which at least one vehicle restraint system is likely to be triggered. The second data point represents an expected triggering time for at least one vehicle restraint system when the parameter limit values ​​are reached.

[0043] According to a particular and non-limiting embodiment of the present invention, the first and second data points of the learning collision scenarios may correspond to data obtained from real-life vehicle collision scenarios with obstacles and / or collision scenarios artificially generated by road safety experts

[0044] For example, frontal, side or rear collision scenarios at different speeds of different vehicles and at different intensities can be collision scenarios used as training data.

[0045] According to one variant, learning collision scenarios can be obtained from other learning collision scenarios using methods called training data augmentation.

[0046] For example, limit values ​​of learning collision scenario parameters can be multiplied by a scalar value, for example 0.85, 1.15, etc.) to take into account calculation errors and sensor uncertainty.

[0047] The neural network learning phase can be divided into a training phase, a testing phase, and a validation phase. The neural network is then usable once it has been trained, tested, and validated.

[0048] The training phase of the neural network can be based on an initial portion (80%) of the training data. The training phase is an iterative phase during which the parameters of the neural network are iteratively adjusted to minimize an error function. More specifically, the neural network provides a trigger time for vehicle restraint devices when training data is presented as input to this neural network. The parameters of the neural network are then iteratively determined to minimize the error function between the predicted trigger times for the training collision scenarios and the expected trigger times (second data point from the first portion of the training collision scenarios) when the Initial data from the first part of the training collision scenarios are presented as input to the neural network. Once trained, the neural network can be tested during the testing phase based on a second part of the training collision scenarios (for example, 10% of the training collision scenarios). A mean squared error is then calculated to measure the quality of the neural network's prediction of the trigger time for the training collision scenarios in the second part of the training collision scenarios. If the mean squared error exceeds a threshold value, for example, a difference of 0.5 ms between the average predicted trigger time of the neural network and the average expected trigger time, then the training phase is iterated by modifying at least one parameter of the neural network.When the mean squared error is less than the threshold value, the neural network is considered to be trained and tested, i.e., it can be used to provide a trigger time of vehicle restraint devices 10 when parameter values ​​of sensors Cj are provided as input to this trained and tested neural network.

[0049] It may happen that several neural networks are trained and tested. In this case, only one of these neural networks is validated to provide a vehicle restraint device activation time 10 when sensor parameter values ​​Cj are provided as input to this trained, tested, and validated neural network. The validation phase can be based on a third of the training collision scenarios (for example, 10% of the training collision scenarios). During the validation phase, a mean squared error can be calculated, and the validated neural network is the one associated with the minimum mean squared error.

[0050] According to a particular and non-limiting embodiment of the present invention, the parameters of the trained (and tested and possibly validated) neural network are stored in a memory of the vehicle 10.

[0051] As will be seen later, a calibration process for the triggering time of the DRk restraint devices (k=l to K) on board the vehicle 10 uses at least one collision scenario Si (i=l to N) which is predetermined, i.e. established by experts and pre-recorded in a memory.

[0052] Each predetermined collision scenario Si comprises a first data point Dli and a second data point D2i. The first data point Dli represents limit values ​​Vp,j (p=l to P) of the parameter of the sensors Cj (j=l to M) onboard the vehicle 10. Each limit value Vp,j defines a limit at a parameter value beyond which at least one restraint device DRk (k=l to K) of the vehicle 10 is likely to be triggered. The second data point D2i represents a time of expected triggering of at least one DRk restraint device of vehicle 10 when the limit values ​​Vp,j of parameters are at least reached.

[0053] A DRk restraint device triggering time calibration process is advantageously implemented by one or more display system processors, for example by one or more processors of one or more computers, for example the IVI system computer.

[0054] The different operations of the calibration process are described below with regard to [Fig.1], according to different examples of implementation of the process.

[0055] In a first operation of the calibration process, at least one collision scenario Si (i=l to N) is obtained from a memory.

[0056] In a second operation of the calibration process, a predicted tripping time of the DRk restraint devices is obtained as output of the trained tripping time prediction model (and tested and possibly validated as explained previously) is obtained when the first data Dli of each collision scenario Si is presented as input to the trained tripping time prediction model.

[0057] In a third operation of the calibration process, a third data point D3i is obtained for each collision scenario Si. The third data point D3i represents a difference between the predicted triggering time for each collision scenario Si and the second data point D2i for each collision scenario Si.

[0058] In a fourth operation of the calibration process, at least one limit value Vp,j of parameter of at least one sensor is adjusted according to the third data obtained.

[0059] For example, if the absolute value of the deviation is less than a threshold value, then no limit value Vp,j of parameter is adjusted.

[0060] For example, if the sign of the deviation is positive then a limit value Vp,j of at least one parameter of a sensor is increased by a predetermined value.

[0061] For example, if the sign of the deviation is negative then a limit value Vp,j of at least one parameter of a sensor is decreased by a predetermined value.

[0062] In a fifth operation of the calibration process, if said at least one limit value Vp,j of parameter is adjusted then the first data Dli of at least one collision scenario Si is adjusted according to said at least one adjusted limit value Vp,j of parameter.

[0063] For example, the limit value of the first data Dli of at least one collision scenario Si is replaced by said adjusted limit value Vp,j of parameter.

[0064] In a sixth operation of the calibration process, if at least one collision scenario is fitted then learning of the trigger time prediction model trained as a function of said at least one fitted collision scenario.

[0065] Figure 2 schematically illustrates a device configured to calibrate a time triggering at least one vehicle-mounted restraint system based on values ​​of vehicle-mounted sensor parameters obtained during vehicle collisions with obstacles, according to a particular and non-limiting embodiment of the present invention. Device 2 corresponds, for example, to a device integrated into the vehicle 10, such as a computer.

[0066] Device 2 is, for example, configured to carry out the operations described opposite [Fig. 1] and / or the steps of the process described opposite [Fig. 3]. Examples of such a device 2 include, but are not limited to, embedded electronic equipment such as a vehicle's on-board computer, an electronic control unit such as an ECU (Electronic Control Unit), a smartphone, a tablet, or a laptop computer. The elements of device 2, individually or in combination, can be integrated into a single integrated circuit, into several integrated circuits, and / or into discrete components. Device 2 can be implemented in the form of electronic circuits or software (or computer) modules, or a combination of electronic circuits and software modules.

[0067] The device 2 comprises one (or more) processor(s) 20 configured to execute instructions for carrying out the steps of the process and / or for executing instructions from the software embedded in the device 2. The processor 20 may include integrated memory, an input / output interface, and various circuits known to those skilled in the art. The device 2 further comprises at least one memory 21, corresponding, for example, to volatile and / or non-volatile memory, and / or includes a memory storage device that may include volatile and / or non-volatile memory, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic disk, or optical disk.

[0068] The computer code of the embedded software(s) including the instructions to be loaded and executed by the processor is for example stored on memory 21.

[0069] According to various particular and non-limiting embodiments, the device 2 is coupled in communication with other similar devices or systems and / or with communication devices, for example a TCU (Telematic Control Unit), for example via a communication bus or through dedicated input / output ports.

[0070] According to a particular and non-limiting embodiment, the device 2 includes a block 22 of interface elements for communicating with external devices, for example a remote server or the "cloud", other nodes of the ad hoc network. The interface elements of block 22 include one or more of the following interfaces: - radio frequency RF interface, for example of the Wi-Fi® type (according to IEEE 802.11), for example in the 2.4 or 5 GHz frequency bands, or of the Bluetooth® type (according to IEEE 802.15.1), in the 2.4 GHz frequency band, or of the Sigfox type using UBN (Ultra Narrow Band) radio technology, or LoRa in the 868 MHz frequency band, LTE (Long-Term Evolution), LTE-Advanced; - USB interface (from the English "Universal Serial Bus" or "Universal Serial Bus" in French); - HDMI interface (from the English "High Definition Multimedia Interface", or "High Definition Multimedia Interface" in French); - LIN interface (from the English "Local Interconnect Network", or in French "Réseau interconnecté local").

[0071] According to another particular and non-limiting embodiment, the device 2 includes a communication interface 23 which allows communication to be established with other devices (such as other computers in the embedded system) via a communication channel 24. The communication interface 23 corresponds, for example, to a transmitter configured to transmit and receive information and / or data via the communication channel 24. The communication interface 23 corresponds, for example, to a wired network of the CAN (Controller Area Network) type, CAN FD (Controller Area Network Flexible Data-Rate), FlexRay (standardized by ISO 17458) or Ethernet (standardized by ISO / IEC 802-3).

[0072] According to a particular and non-limiting embodiment, the device 2 can provide output signals to one or more external devices, such as a display screen 25, touch or not, one or more speakers 26 and / or other peripherals 27 (projection system) via output interfaces 28, 29, 30 respectively. According to a variant, one or more of the external devices is integrated into the device 2.

[0073] Figure 3 schematically illustrates a flowchart of the different steps of a calibration process for the triggering time of at least one on-board restraint device of a vehicle, according to a particular and non-limiting embodiment of the present invention.

[0074] The process is implemented for example by a device embedded in the vehicle 10 or by the device 2 of the [Fig.2].

[0075] In a first step 31, at least one collision scenario is obtained from a memory, each collision scenario comprising a first data point and a second data point, the first data point being representative of parameter limit values ​​of at least one vehicle-mounted sensor, each limit value defining a limit to a parameter value of said at least one sensor beyond which said at least one vehicle restraint device is likely to be triggered and, the second data point being representative of an expected triggering time of said at least one vehicle restraint device when the parameter limit values ​​are at least reached.

[0076] In a second step 32, a predicted triggering time of said at least one restraint device is obtained as output of a triggering time prediction model when the first data of each collision scenario is presented as input to the triggering time prediction model.

[0077] In a third step 33, a third data point is obtained. The third data point represents a difference between the predicted triggering time for each collision scenario and the second data point for each collision scenario.

[0078] In a fourth step 34, at least one limit value of parameter of said at least one sensor is adjusted according to the third data obtained.

[0079] In a fifth step 35, if said at least one parameter limit value is adjusted then the first data of at least one collision scenario is adjusted according to said at least one adjusted parameter limit value.

[0080] In a sixth step 36, if at least one collision scenario is fitted, the trained trigger time prediction model is again trained during a learning phase based on said at least one fitted collision scenario.

[0081] According to one variant, the variants and examples of the operations described in relation to [Fig.1] apply to the steps of the process in [Fig.3].

[0082] Of course, the present invention is not limited to the embodiments described above but extends to a method for calibrating the activation time of vehicle-mounted restraint systems, which would include secondary steps without departing from the scope of the present invention. The same would apply to a device configured for implementing such a method.

[0083] The present invention also relates to a vehicle, for example an automobile or more generally an autonomous land-powered vehicle, comprising device 2 of [Fig.2].

Claims

Demands

1. A method for calibrating the activation time of at least one vehicle-mounted restraint system, said method being implemented by at least one processor and comprising the following steps: - obtaining (31) at least one collision scenario from a memory, each collision scenario comprising a first data point and a second data point, the first data point being representative of parameter limit values ​​of at least one vehicle-mounted sensor, each limit value defining a limit to a parameter value of said at least one sensor beyond which said at least one vehicle restraint system is likely to be activated and, the second data point being representative of an expected activation time of said at least one vehicle restraint system when the parameter limit values ​​are at least reached;- obtaining (32) a predicted tripping time of said at least one restraint device as output of a tripping time prediction model when the first data of each collision scenario is presented as input to the tripping time prediction model; - obtaining (33) a third data representing a difference between the predicted tripping time for each collision scenario and the second data of each collision scenario; - adjusting (34) at least one parameter limit value of said at least one sensor as a function of the third data obtained; - if said at least one parameter limit value is adjusted then adjusting (35) the first data of at least one collision scenario as a function of said at least one adjusted parameter limit value;and - if at least one collision scenario is fitted then learning (36) of the trigger time prediction model trained as a function of said at least one fitted collision scenario.;

2. A method according to claim 1, wherein the trigger time prediction model is a neural network.

3. A method according to claim 2, wherein the neural network is trained during a learning phase using data learning including first data and second data of learning collision scenarios.

4. A method according to claim 3, wherein each learning collision scenario comprises a first data point and a second data point, the first data point being representative of parameter limit values ​​of at least one vehicle-mounted sensor, each limit value defining a limit to a parameter value of at least one vehicle-mounted sensor beyond which at least one vehicle restraint device is likely to be triggered, and the second data point being representative of an expected triggering time of said at least one vehicle restraint device when the parameter limit values ​​are at least reached.

5. A method according to claim 4, wherein the first and second data of the learning collision scenarios correspond to data obtained from real vehicle collision scenarios with obstacles and / or collision scenarios artificially generated by road safety experts.

6. A method according to any one of the preceding claims, wherein the first data item includes a pressure limit value on a vehicle door.

7. A method according to any one of the preceding claims, wherein the first data further includes a vehicle deceleration limit value.

8. Device (2) for calibrating the triggering time of at least one vehicle-mounted restraint device, said device (2) comprising a memory (21) associated with at least one processor (20) configured for carrying out the steps of the method according to any one of claims 1 to 7.

9. Computer program comprising instructions for carrying out the method according to any one of claims 1 to 7, when such instructions are executed by a processor.

10. Vehicle comprising a device according to claim 8.