Method for configuring at least one vehicle onboard system that contributes to creating an ambiance inside the vehicle's passenger compartment

Abstract

The present invention relates to a method and device for configuring at least one embedded system in a vehicle that contributes to creating an ambiance inside the vehicle's passenger compartment. The method adjusts (45-47) the internal parameters of a pre-trained neural network according to user preferences during a journey in the vehicle in order to improve configuration predictions for that user during subsequent journeys. The adjustment of the neural network's internal parameters is performed during a reinforcement learning phase. (See Figure 4 for an abstract.)

Description

Title of the invention: Method for configuring at least one embedded system of a vehicle contributing to creating an atmosphere inside a vehicle's passenger compartment technical field

[0001] The present invention falls within the field of artificial intelligence and more specifically within the reinforcement learning of a neural network used to configure embedded systems of a vehicle contributing to creating an atmosphere inside a vehicle's passenger compartment. Technological background

[0002] The field of intelligent vehicles has seen considerable progress in recent years, particularly with the integration of artificial intelligence (AI) technologies to improve the comfort, safety, and efficiency of transportation. One of the challenges is the real-time adaptation of parameters of vehicle onboard systems that contribute to creating an atmosphere inside the vehicle's cabin. This involves adjusting the configuration of these onboard systems according to the preferences and needs of the vehicle's users.

[0003] To personalize a vehicle interior, that is to say to adapt (personalize) in real time parameters of vehicle on-board systems which contribute to creating an atmosphere inside the vehicle's interior, it is necessary that this adaptation be carried out on each new journey of the vehicle in order to promote the comfort of the users and in particular the driver (driver's seat tilt, music choices, interior lighting, etc.) and safety (angles of the exterior and exterior mirrors).

[0004] Several methods have been proposed in the prior art for personalizing a vehicle's interior according to the tastes or needs of a vehicle user. These methods can be classified according to whether or not they use machine learning and, if so, whether the machine learning model is trained in the vehicle or pre-trained, i.e., trained before being installed in the vehicle. Some methods (such as, for example, US patent application US20220234593 A1) use pre-trained neural networks to define multimodal interactions with the driver, in order to generate parameters for the vehicle's embedded systems that contribute to creating an ambiance inside the vehicle's interior. Other methods (such as US patent US20230122813, Chinese patent CN115991072, or German patent DE 102022127026) present methods for customizing the passenger compartment of a vehicle based on driver identity using both pre-trained models to classify drivers into groups associated with particular preferences for cabin personalization, and field learning, i.e. on board the vehicle in use, to train an emotion detection model that can provide output data relating to adjustments of the vehicle's onboard systems parameters that contribute to creating an atmosphere inside the vehicle's cabin.

[0005] These methods use pre-trained models to personalize the vehicle interior and assume that the driver is able to provide feedback. However, drivers rarely provide feedback on proactive suggestions from intelligent systems.

[0006] The problem solved by the present invention is to provide a method for configuring at least one embedded system of a vehicle contributing to creating an atmosphere inside a vehicle's passenger compartment that takes into account situations where frequent or detailed feedback from the driver cannot be assumed, but where his satisfaction can nevertheless be estimated. Summary of the present invention

[0007] One object of the present invention is to solve at least one of the problems of the technological background described above.

[0008] Another object of the present invention is to improve the personalization of the passenger compartment of a vehicle during a journey.

[0009] Another object of the present invention is to improve the configuration prediction of at least one vehicle-mounted system contributing to creating an atmosphere inside a vehicle cabin when user feedback is infrequent.

[0010] According to a first aspect, the present invention relates to a method for configuring at least one vehicle embedded system contributing to creating an atmosphere inside a vehicle's passenger compartment, the method comprising the following steps: - obtaining a neural network whose internal parameters are initially determined during a learning phase to provide as output from the neural network an output data representative of a configuration of said at least one embedded system when an input data representative of a characteristic associated with a vehicle user and / or with an atmosphere inside and / or outside the vehicle's passenger compartment is presented as input to the neural network; - If a change condition is met, then the method further comprises a step of adjusting the configuration of at least one of said at least one embedded system according to the output data; - If the internal parameters of the neural network need to be adjusted, then the process further includes the following steps before providing the output data to the neural network: - obtaining (45) a fitting data representative of a configuration fitting of at least one of said at least one embedded system; - calculation (46) of a reward value as a function of the adjustment data; and - adjustment (47) of the internal parameters of the neural network during a reinforcement learning phase using a gradient method taking as a parameter the calculated reward value.

[0011] According to the present invention, the reinforcement learning used by the method is performed entirely at the edge. Therefore, it is not necessary to label the data because supervised learning is not used. The training costs of the neural network are low thanks to computations that can be performed at the edge and the use of a small neural network.

[0012] The present invention makes it possible to provide configuration predictions for embedded systems to personalize the passenger compartment of a vehicle and also makes it possible to train a neural network based on adjustments made to one of these embedded systems during a journey by a vehicle user, so that, during subsequent journeys, the configurations predicted by the neural network are better adapted to the preferences of that user.

[0013] The method is also advantageous because reinforcement learning of the neural network does not require the transfer of information concerning the vehicle user to a remote third party.

[0014] According to a particular and non-limiting embodiment of the present invention, the change condition is met if at least one of the following conditions is met: - a vehicle user modifies a setting of said at least one on-board system; - change in the level of interior and / or exterior brightness of the vehicle's passenger compartment; - change in weather conditions; - change in the interior and / or exterior temperature of the vehicle's passenger compartment.

[0015] According to a particular and non-limiting embodiment of the present invention, the internal parameters of the neural network must be adjusted if at least one of the following conditions is met: - an adjustment of the internal parameters of the neural network is validated by the vehicle user; - a vehicle user modifies a setting of said at least one on-board system; - a new vehicle route is initiated; - the number of journeys made by the user without modification of a configuration of said at least one embedded system is less than a threshold.

[0016] According to a particular and non-limiting embodiment of the present invention, the input data is representative of at least one of the following pieces of information: - information representing an identifier of a driver of the vehicle; - information representing a facial emotion of the driver; - information representative of the time at which the first data point is obtained; - information representative of a level of brightness outside the vehicle's passenger compartment; - information representative of weather conditions; - information representative of the vehicle's outside temperature; - information representative of the temperature inside the vehicle's passenger compartment.

[0017] According to a particular and non-limiting embodiment of the present invention, the output data is representative of at least one of the following information: - information representative of an adjustment of a position and / or inclination of a driver's seat of the vehicle; - information representative of an adjustment of a position and / or inclination of a steering wheel of the vehicle; - information representative of an adjustment of a position and / or an inclination of at least one rearview mirror of the vehicle; - information representative of an adjustment of a vehicle's interior lighting system; - information representative of an adjustment of an audio system.

[0018] According to a particular and non-limiting embodiment of the present invention, the neural network is a multilayer neural network.

[0019] According to a second aspect, the present invention relates to a configuration device for at least one embedded system of a vehicle contributing to creating an atmosphere inside a vehicle cabin, the device comprising a memory associated with a processor configured for the implementation of the steps of the process according to the first aspect of the present invention.

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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.

[0025] 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.

[0026] 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

[0027] 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 4, in which:

[0028] [Fig. 1] schematically illustrates a principle of reinforcement learning of a neural network.

[0029] [Fig.2] schematically illustrates the operation of the method of configuring at least one embedded system of a vehicle contributing to creating an atmosphere inside a vehicle cabin according to the present invention.

[0030] [Fig.3] schematically illustrates a device configured to configure at least one vehicle-mounted system contributing to creating an atmosphere inside a vehicle's passenger compartment, according to a particular and non-limiting embodiment of the present invention.

[0031] [Fig.4] illustrates a flowchart of the different stages of a configuration process for at least one embedded system of a vehicle contributing to creating a ambiance inside a vehicle's passenger compartment, according to a particular and non-limiting embodiment of the present invention. Description of examples of achievements

[0032] A method and a device for configuring at least one vehicle-mounted system contributing to creating an atmosphere inside a vehicle's passenger compartment will now be described in what follows with joint reference to Figures 1 to 4. The same elements are identified with the same reference signs throughout the following description.

[0033] 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.

[0034] The vehicle interior personalization method according to the present invention is based on reinforcement learning.

[0035] Reinforcement learning is an important part of artificial intelligence, which deals with the design of agents that interact with complex environments and learn to do so without explicit human supervision. In reinforcement learning, a neural network is trained to maximize an assumed available reward function, rather than adjusting internal parameters of that neural network so that the output data is as close as possible to ground truth data. Reinforcement learning has been used, in particular, in autonomous driving systems to predict and plan vehicle trajectories, but also to make Large Language Models (LLMs) more efficient in their interactions with users of those models.

[0036] Reinforcement learning (RL) consists of learning by interacting with an environment 110 ([Fig. 1]). An agent 100 learns from the consequences of its actions 101, rather than receiving explicit instruction. The agent 100 selects its actions 101 based on past experiences (exploitation) and new choices (exploration), which essentially amounts to trial-and-error learning. The reinforcement signal that the agent 100 receives is a numerical reward signal 111, which encodes the success of the outcome of an action 101, and a state 112 of the model, and the agent 100 seeks to learn to select actions that maximize the reward accumulated over time.

[0037] The present invention relates to a method and device for configuring at least one embedded system of a vehicle that contributes to creating an atmosphere inside of a vehicle's interior. The process trains a pre-trained neural network that learns to adjust parameters of these embedded systems according to driver (or vehicle user) preferences to personalize the vehicle's interior during the vehicle journey.

[0038] Fig. 2 schematically illustrates the operation of the method for configuring at least one vehicle-mounted system contributing to creating an atmosphere inside a vehicle cabin according to the present invention.

[0039] For each new journey, the pre-trained neural network 200 provides an initial configuration of at least one vehicle-mounted system that contributes to creating an ambiance inside the vehicle's passenger compartment during each new journey. At least one vehicle-mounted system is then configured according to this initial configuration. If it is detected that the internal parameters of the neural network need to be adjusted, for example, when the driver changes the initial settings of one of these vehicle-mounted systems, a reward value 201 (reward signal) is calculated. Based on this reward signal, the neural network undergoes a reinforcement learning phase to improve its future prediction of the configuration of at least one vehicle-mounted system. The method assumes that there are few reward signals (i.e., that the user makes few configuration corrections).Said at least one embedded system is configured from the initial configuration and then by a new configuration 202 as soon as a change condition is detected, such as for example a change in climatic conditions or interior and / or exterior lighting of the passenger compartment. This new configuration 202 is generated by the neural network.

[0040] The method of configuring at least one embedded system of a vehicle according to the present invention can be validated or invalidated by a driver of the vehicle by means of, for example, a human-machine interface which may take the form of a button on the dashboard.

[0041] Vehicle on-board systems that create an atmosphere inside a vehicle's passenger compartment can be systems for adjusting the position and / or tilt of the driver's seat, a ventilation and / or air conditioning system, a passenger compartment lighting system, a rearview mirror adjustment system, etc.

[0042] The neural network is pre-trained, for example, on a set of training input data for different users and different interior and exterior climatic conditions of the passenger compartment. The purpose of pre-training is to adjust the internal parameters of the neural network so that it provides configurations of the embedded systems that better correspond to the wishes of the vehicle's drivers for a given journey.

[0043] Once trained and installed in the vehicle, during an inference phase, the neural network receives input data representative of a characteristic associated with a vehicle user and / or an atmosphere inside and / or outside the passenger compartment.

[0044] This input data can be obtained from on-board sensors of the vehicle. For example, the driver's identity can be obtained from an input system comprising a human-machine interface which can be implemented as a graphical interface displayed on a touchscreen.

[0045] This input data can also be obtained from a driver emotion acquisition system. It can also be obtained from a clock or light sensors located inside and / or outside the vehicle. It can also be obtained from a weather condition acquisition system located on or off the vehicle, or from a temperature acquisition system located inside and / or outside the vehicle.

[0046] According to a particular, non-limiting embodiment of the present invention, the neural network is relatively small so that field learning converges quickly. Reinforcement learning can thus be performed by the vehicle's computer during a journey, thereby avoiding the transfer of the driver's (or other vehicle user's) preferences to an external computer and enabling rapid convergence of reinforcement learning.

[0047] According to a particular and non-limiting embodiment of the present invention, the neural network is a multilayer perceptron network (MLP for MultiLayer Perceptron).

[0048] A perceptron is a linear classifier generally comprising several inputs and a single output and characterized by an activation function, weights (or synaptic coefficients) and a bias (or threshold).

[0049] For example, a perceptron with n inputs (xh ...,xn) and a single output o is defined by n weights (W], and a bias (or threshold) 0:

[0050] o = f(z) 0 otherwise

[0051] The output o then results from the application of the Heaviside function to the postsynaptic potential z given by:

[0052]

[0053]

[0054] z=Li=iWiXi -e with a non-linear activation function H(x) given for example by: " - " x [ 0 sz x < 0 VxgP, € H(x) - . . . It if x > 1

[0055] The internal parameters of the neural network are then these weights and biases for the set of perceptrons.

[0056] The present invention is not limited to this definition of perceptron or to the use of other basic elements forming a layer of the neural network. It is also not limited to the number of perceptrons (or other basic elements) used per layer or to the number of layers. The internal parameters of the neural network are usually weights and biases, regardless of the basic elements of the neural network layers.

[0057] A configuration process for at least one vehicle embedded system contributing to creating an atmosphere inside a vehicle cabin will now be described according to the present invention.

[0058] In a first operation, the neural network is obtained.

[0059] This neural network is pre-trained, that is to say that its internal parameters are initially determined during a learning phase to provide as output of the neural network an output data representative of a configuration of the vehicle's embedded systems that create an atmosphere inside the vehicle's passenger compartment when an input data representative of a characteristic associated with a driver of the vehicle and / or with an atmosphere inside and / or outside the vehicle's passenger compartment is presented as input to the neural network.

[0060] According to a particular and non-limiting embodiment of the present invention, the input data is representative of at least one of the following pieces of information: - information representing an identifier of a driver of the vehicle; - information representing a facial emotion of the driver; - information representing the time at which the first data point is obtained; - information representing the level of brightness outside the vehicle's passenger compartment; - information representative of weather conditions; - information representative of the vehicle's outside temperature; - information representative of the temperature inside the vehicle's passenger compartment.

[0061] The present invention is not limited to one type or number of input data but extends to any type of input data and / or combination of input data that is related to the personalization of a vehicle interior.

[0062] According to a particular and non-limiting embodiment of the present invention, the output data is representative of at least one of the following information: - information representative of an adjustment of a position and / or an inclination of a driver's seat of the vehicle; - information representing the adjustment of a position and / or tilt of a vehicle's steering wheel; - information representative of an adjustment of a position and / or an inclination of at least one rearview mirror of the vehicle; - information representative of an adjustment of a vehicle's interior lighting system; - information representative of an adjustment of an audio system.

[0063] The present invention is not limited to one type or number of output data but extends to any type of output data and / or combination of output data that is related to a configuration of an embedded vehicle system contributing to creating an atmosphere inside a vehicle's passenger compartment.

[0064] In a second operation, if a change condition is met then the configuration of at least one of said at least one embedded system is adjusted according to the output data.

[0065] The second operation allows an automatic change of the configuration of said at least one on-board system when conditions requiring configuration changes occur during a vehicle journey.

[0066] According to a particular and non-limiting embodiment of the present invention, the change condition is met if at least one of the following conditions is met: - a vehicle user modifies a setting of said at least one on-board system; - change in the level of interior and / or exterior brightness of the vehicle's passenger compartment; - change in weather conditions; - change in the interior and / or exterior temperature of the vehicle's passenger compartment.

[0067] For example, following a modification of at least one of said at least one on-board system, for example, a modification of the driver's seat angle, steering wheel angle, and / or mirrors by the user, a new configuration of said at least one on-board system is present at the output of the trained neural network when the new input data, including the new setting of said at least one on-board system, is presented as input to this neural network. The passenger compartment lighting system and / or the audio system can also be adapted during the second operation according to the user's configuration preferences (input data).

[0068] In a third operation, it is detected whether the internal parameters of the neural network need to be adjusted.

[0069] According to a particular and non-limiting embodiment of the present invention, it is detected that the internal parameters of the neural network must be updated when at least one of the following conditions is met: - an adjustment of the internal parameters of the neural network is accepted by the vehicle driver; - a vehicle user modifies a setting of said at least one on-board system; - a new vehicle route is initiated; - the number of journeys made by a user without modification of a configuration of said at least one embedded system is less than a threshold.

[0070] If it is detected that the internal parameters of the neural network need to be adjusted then the process further includes a fourth, fifth and sixth operation before providing the output data.

[0071] In the fourth operation, an adjustment data point is obtained. The adjustment data point is representative of a configuration adjustment of at least one of said at least one component of the vehicle.

[0072] In the fifth operation, a reward value is calculated based on the adjustment data.

[0073] The calculation of the reward is based on user adjustments, which means that the user does not need to provide explicit feedback.

[0074] The adjustments made by the user to the settings of said at least one embedded system are used to calculate the reward value. The reward is calculated, for example, proportionally to the number of adjustments made.

[0075] According to a particular and non-limiting embodiment of the present invention, the reward value is a sum of the adjustment amplitude of parameters relating to several (n) vehicle onboard systems multiplied by a scaling coefficient k.

[0076] For example, the reward value Reward(t) can be given by:

[0077] Reward(t) = k(amplitude of adjustment 1 + amplitude of adjustment 2 + ...+ amplitude of adjustment n).

[0078] According to a particular and non-limiting embodiment of the present invention, the reward value is a weighted sum (weight al, ..., an) of adjustment amplitude of several vehicle components.

[0079] For example, Reward(t) = al*amplitude of adjustment 1 + a2*amplitude of adjustment 2 + ...+ an*amplitude of adjustment n).

[0080] In this case, the scale factor k is then absorbed into the weights al,...,an.

[0081] In the sixth operation, the internal parameters of the neural network are adjusted during a reinforcement learning phase using a gradient method taking the calculated reward value as a parameter.

[0082] No user data is stored or transferred to a third party because user data is only used transiently during the reinforcement learning phase and is subsequently deleted.

[0083] The neural network's reinforcement learning (fourth, fifth, and sixth operations) can stop, for example, if a user has deactivated this learning, or if no configuration adjustment has been made by a user during a specified number of consecutive trips. Learning can then be deactivated by default; the driver is informed and confirms the training stoppage via a human-machine interface. The driver always has the option to reset the learning.

[0084] According to a particular and non-limiting embodiment of the present invention, the method used during the reinforcement learning phase is the method described by Baxter, J. and Bartlett, PL ((2001, Infinity-horizon policy-gradient estimation. Journal of Artificial Intelligence Research, 15:319350, Nov 2001. ISSN 1076-9757. doi: 10.1613 / jair.8O6).

[0085] Figure 3 schematically illustrates a device 3 configured to configure at least one embedded system in a vehicle that contributes to creating an atmosphere inside the vehicle's passenger compartment, according to a particular and non-limiting embodiment of the present invention. Device 3 corresponds, for example, to a device embedded in the vehicle, such as a computer.

[0086] Device 3 is, for example, configured to carry out the operations described opposite Figures 1 and 2 and / or the steps of the process described opposite [Fig. 4]. Examples of such a device 3 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 3, individually or in combination, can be integrated into a single integrated circuit, into several integrated circuits, and / or into discrete components. Device 3 can be implemented in the form of electronic circuits or software (or computer) modules, or a combination of electronic circuits and software modules.

[0087] The device 3 comprises one (or more) processor(s) 30 configured to execute instructions for carrying out the steps of the process and / or for executing instructions from the software embedded in the device 3. The processor 30 may include integrated memory, an input / output interface, and various circuits known to those skilled in the art. The device 3 further comprises at least one memory 31, corresponding, for example, to volatile and / or non-volatile memory, and / or includes a memory storage device which may include memory volatile and / or non-volatile, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic or optical disk.

[0088] 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 31.

[0089] According to various particular and non-limiting embodiments, the device 3 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.

[0090] According to a particular and non-limiting embodiment, the device 3 includes a block 32 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 the block 32 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").

[0091] According to another particular and non-limiting embodiment, the device 3 includes a communication interface 33 that enables communication with other devices (such as other computers in the embedded system or sensors) via a communication channel 34. The communication interface 33 corresponds, for example, to a transmitter configured to transmit and receive information and / or data via the communication channel 34. The communication interface 33 corresponds, for example, to a wired CAN (Controller Area Network) or CAN FD (Controller Area Network Flexible Data-Rate) type network. "Flexible data rate controller network"), FlexRay (standardized by ISO 17458) or Ethernet (standardized by ISO / IEC 802-3).

[0092] According to a particular and non-limiting embodiment, the device 3 can provide output signals to one or more external devices, such as a display screen 35, touch or not, one or more speakers 36 and / or other peripherals 37 (projection system) via output interfaces 38, 39, 40 respectively. According to a variant, one or more of the external devices is integrated into the device 3.

[0093] Figure 4 illustrates a flowchart of the different steps of a configuration process for at least one embedded system of a vehicle contributing to creating an atmosphere inside a vehicle's passenger compartment, according to a particular and non-limiting embodiment of the present invention.

[0094] The process is implemented for example by a device embedded in the vehicle or by device 3 of [Fig.3].

[0095] In a step 41, a neural network is obtained. The neural network includes internal parameters which are initially determined during a learning phase to provide as output of the neural network an output data representative of a configuration of said at least one embedded system when an input data representative of a characteristic associated with a user of the vehicle and / or with an environment inside and / or outside the passenger compartment of the vehicle is presented as input to the neural network.

[0096] In a step 42, it is checked whether a change condition is met.

[0097] If so, then in step 43, the configuration of at least one of said at less an embedded system is adjusted according to the output data.

[0098] In a step 44, it is checked whether the internal parameters of the neural network need to be adjusted.

[0099] If so, then the process further includes steps 45-47 before providing the output data.

[0100] In step 45, an adjustment data is obtained. The adjustment data is representative of a configuration adjustment of at least one of said at least one component of the vehicle.

[0101] In step 46, a reward value is calculated based on the adjustment data.

[0102] In step 47, the internal parameters of the neural network are adjusted during a reinforcement learning phase using a gradient method taking the calculated reward value as a parameter.

[0103] According to one variant, the variants and examples of the operations described in relation to [Fig.2] apply to the steps of the process in [Fig.4].

[0104] Of course, the present invention is not limited to the embodiments described above but extends to a method for configuring at least one vehicle-mounted system that contributes to creating an atmosphere inside the vehicle's passenger compartment, which would include secondary steps without falling outside the scope of the present invention. The same would apply to a device configured for implementing such a method.

[0105] The present invention also relates to a vehicle, for example an automobile or more generally an autonomous land-powered vehicle, comprising the device 3 of [Fig.3].

Claims

Demands

1. A method for configuring at least one vehicle-mounted system contributing to creating an ambiance inside a vehicle's passenger compartment, the method comprising the following steps: - obtaining (41) a neural network whose internal parameters are initially determined during a training phase to provide as output from the neural network an output data representative of a configuration of said at least one vehicle-mounted system when an input data representative of a characteristic associated with a vehicle user and / or an ambiance inside and / or outside the vehicle's passenger compartment is presented as input to the neural network; - If a change condition is met (42) then the method further comprises a step of adjusting (43) the configuration of at least one of said at least one vehicle-mounted system according to the output data;- If the internal parameters of the neural network need to be adjusted (44) then the method further includes the following steps before providing the output data to the output of the neural network: - obtaining (45) an adjustment data representative of a configuration adjustment of at least one of said at least one embedded system; - calculating (46) a reward value as a function of the adjustment data; and - adjusting (47) the internal parameters of the neural network during a reinforcement learning phase using a gradient method taking as a parameter the calculated reward value.

2. A method according to claim 1, wherein the change condition is satisfied if at least one of the following conditions is met: - a vehicle user changes a setting of said at least one on-board system; - change in the interior and / or exterior brightness level of the vehicle's passenger compartment; - change in weather conditions; - change in the internal and / or external temperature of the vehicle's passenger compartment.

3. A method according to claim 1 or 2, wherein the internal parameters of the neural network must be adjusted if at least one of the following conditions is met: - an adjustment of the internal parameters of the neural network is validated by the vehicle user; - a vehicle user modifies a setting of said at least one on-board system; - a new vehicle journey is initiated; - a number of journeys made by the user without modification of a configuration of said at least one on-board system is less than a threshold.

4. A method according to any one of the preceding claims, wherein the input data is representative of at least one of the following: - information representing an identifier of a driver of the vehicle; - information representing a facial emotion of the driver; - information representing the time at which the first data is obtained; - information representing a level of brightness outside the passenger compartment of the vehicle; - information representing weather conditions; - information representing an outside temperature of the vehicle; - information representing a temperature inside the passenger compartment of the vehicle.

5. A method according to any one of the preceding claims, wherein the output data is representative of at least one of the following information: - information representative of an adjustment of the position and / or tilt of a driver's seat of the vehicle; - information representative of an adjustment of the position and / or tilt of a steering wheel of the vehicle; - information representative of an adjustment of the position and / or tilt of at least one rearview mirror of the vehicle; - information representing an adjustment of a vehicle interior lighting system; - information representing an adjustment of an audio system.

6. A method according to any one of the preceding claims, wherein the neural network is a multilayer neural network.

7. Device (3) for configuring at least one vehicle-mounted system contributing to creating an ambiance inside a vehicle's passenger compartment, said device (3) comprising a memory (31) associated with at least one processor (30) configured for carrying out the steps of the method according to any one of claims 1 to 6.

8. Computer program comprising instructions for carrying out the method according to any one of claims 1 to 6, when such instructions are executed by a processor.

9. Computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the process according to any one of claims 1 to 6.

10. Vehicle comprising a device according to claim 7.

Images