Method and system for detecting a state of mind wandering in a vehicle driver

Abstract

Method and system for detecting a driver's mind-wandering state. The present invention relates to a method (100) for detecting a driver's mind-wandering state, comprising a reception step (104) of: - data representative of the driver's physiological data, - data representative of the vehicle's kinematics and / or the driver's controls, and - at least one data point representative of a driving context, the detection method (100) being characterized in that it comprises a quantification step (108) of the driver's mind-wandering state as a function of the received data. (Figure 2)

Description

Title of the invention: Method and system for detecting a state of mind wandering in a vehicle driver

[0001] The present invention relates to the field of the automotive industry, and more specifically concerns a method and a system for detecting a state of mind wandering in a vehicle driver.

[0002] A substantial percentage of traffic accidents are due to states of drowsiness, distraction or inattention, to such an extent that the new European safety rules GSR 2 (for the English "General Safety Regulations" 2), or road safety organizations such as EuroNCAP (for the English "European New Car Assessment Program"), require new vehicles to have systems for detecting such states.

[0003] In fact, there are many types or causes of distraction or inattention, related to listening to music, a passenger, handling objects in the passenger compartment, consulting a screen, this consultation being related to driving or not, ... These types or causes of distraction can be analyzed by sensors present in the vehicle, which can detect in particular the movements of the driver.

[0004] Another type of distraction, much less easily detectable, is mind-wandering, defined in the scientific literature as a shift in the content of thoughts away from the driving task, towards self-generated thoughts and feelings. Mind-wandering is therefore a form of endogenous or internal distraction stemming from the driver's concerns or thoughts. Its occurrence is facilitated by a monotonous driving environment, a familiar route, or a driving task that does not require sustained attention, thus freeing up attentional resources to think about something else. It is also encouraged by negative emotions, particularly sadness or depression.

[0005] Mind wandering is therefore a type of purely cognitive distraction, not linked to a physical, visual and / or auditory distraction, and therefore cannot be detected solely by observing, for example, the direction of the driver's gaze.

[0006] However, when the driver is in such a state, he sees without looking, reduces his distance from other vehicles, drives faster, reacts more slowly, and has difficulty maintaining a straight trajectory, hence a significant number of accidents related to this type of distraction, which can occupy between one third and two thirds of the driver's driving time.

[0007] It is therefore difficult on the one hand to detect a state of mind wandering in a driver, and on the other hand unproductive to warn the driver too frequently of this state when no imminent risk is present, given that such warnings would no longer have any effect on him, and could make the driver insensitive to other more priority alerts.

[0008] The present invention aims to remedy at least in part the aforementioned drawbacks by providing a method and a system for detecting a state of mind-wandering, which makes it possible to prevent accidents by making effective the alerts notifying the driver of such a state.

[0009] To this end, the invention proposes a method for detecting a state of mind wandering in a vehicle driver, comprising a reception step of: - at least one piece of data representative of the driver's physiological data, - at least one piece of data representative of the vehicle's kinematics and / or the driver's controls, and - at least one piece of data representative of a driving context, the detection process being characterized in that it includes a step of quantifying a state of mind wandering of the driver according to the data received.

[0010] Vehicle kinematics refers to the movements of the vehicle and its dynamic parameters. Data representing vehicle kinematics may therefore include the vehicle's position, speed, acceleration or deceleration, trajectory, radius of curvature of its trajectory, etc.

[0011] The detection method is implemented in the vehicle, in particular in a detection system implemented in a computer and using various physiological sensors.

[0012] Thanks to the invention, since the state of wandering is quantified, the driver can, for example, only be alerted if their wandering state is high and if the driving context presents a high risk of an accident in such a state. Thus, the alert is more effective because it is only triggered when relevant, which increases the driver's acceptability of the alert and their confidence in the usefulness of the function implemented by the detection method according to the invention.

[0013] In one embodiment of the invention, the quantification step includes a step of associating the received data, by calculating at least one distance, between the received data and a group among a set of groups associated with levels of mind-wandering, the mind-wandering state of the driver being quantified by the level of the group to which the received data are associated.

[0014] In this embodiment of the invention, the groups are formed, for example, using the K-means algorithm. However, alternatively, The quantification step uses, for example, a neural network that has previously undergone a supervised learning step.

[0015] It is understood that each group is associated with a single level of mind-wandering, but that the set of groups is representative of the different levels of a scale of levels of mind-wandering, as stored in the computer implementing the detection method according to the invention.

[0016] In this embodiment of the invention, the detection method according to the invention includes, for example, a preliminary learning step comprising sub-steps of: - collection, over each driving period of a plurality of driving periods, of at least one data point representative of the physiological data of a driving subject over the driving period, of at least one data point representative of the kinematics of the vehicle driven by the driving subject or of the driving subject's controls over the driving period, and of at least one data point representative of a driving context over the driving period, the collected data being represented over each driving period by a multidimensional reference vector, - grouping of the multidimensional reference vectors into a plurality of groups each associated with a level of thought wandering, the quantification step including an assignment of a multidimensional vector representative of the received data, called the current multidimensional vector, to one of the groups in the plurality of groups,by calculating distances between, on the one hand, the current multidimensional vector and, on the other hand, each of the groups in the plurality of groups, the current multidimensional vector being associated with the group in the plurality of groups for which the calculated distance is the smallest.

[0017] The data collection substep is, for example, carried out on a set of journeys representative of different driving contexts, including a monotonous driving context, for example on an open highway, and a driving context requiring sustained attention, for example in a city. During each driving period in this data collection substep, the driver is questioned to self-assess their level of mind wandering. This assessment method is called "probe-caught." However, other methods for assessing the driver's level of mind wandering for each driving period can be used in this data collection substep.

[0018] In one embodiment of the invention, the data collected during a driving period, or the data received, includes at least one of the following: - the driver's heart rate, - an opening of the driver's pupils, - the direction of the driver's gaze, - the driver's blink rate, - the vehicle's speed - a demand for engine torque from the vehicle, - an angle at the steering wheel, - a distance between vehicles, - brightness, - the duration for which the driver has been in a state of mind-wandering, - the positions of surrounding obstacles or other road users, and - a type of road.

[0019] Low light levels particularly encourage entering a state of mind-wandering. The received data may include other meteorological data, for example, a temperature reading. The received data may also include other data representative of driver commands, for example, pressing a pedal, a number of driving actions such as turning, activating a turn signal, etc.

[0020] Each multidimensional reference vector includes, for example, one dimension per data point collected or received.

[0021] On the other hand, the distance calculation includes, for example, for each group in the plurality of groups, a calculation of a weighted sum of differences between coordinates of the current multidimensional vector and coordinates of one of the groups in the plurality of groups. This optional feature allows for greater weight to be given, in the distance calculation, to distances relative to certain data, for example, data representing physiological measurements and / or driver commands, depending on contextual data, such as the type of road.

[0022] Thus, to better account for the impact of road type on the values ​​present in the current multidimensional vector, distance calculations will give more weight to the difference between the vehicle's speed and the average speed of the group with which a distance is being calculated, and to the difference between the heart rate and the average heart rate of that group, when the road type is a long straight road or a highway, than when the road type is urban, for example. This allows the current multidimensional vector to be linked more accurately to a group representative of the driver's state of thought wandering.

[0023] In another example, to take into account the impact of meteorological data on the values ​​present in the current multidimensional vector, the calculation of distances will also give more weight to the difference between the speed of the vehicle and the average speed of the group with which a distance is calculated, and The difference between heart rate and the average heart rate of this group is significant in low light conditions compared to high light conditions. Since weather data has a less pronounced impact on the driver's cognitive state than road type, the influence of weather data will be less significant than that of road type.

[0024] According to an optional feature of the detection method according to the invention, it further includes a driver alert step via a human-machine interface, of a high level of mind wandering, when the level of mind wandering from the quantification step is strictly greater than a predetermined level of mind wandering.

[0025] The data reception step may then include receiving a history of alerts previously triggered during earlier stages of alerting the driver to a high level of mind-wandering. Thus, the current alerting step will take previous alerts into account to determine whether or not to trigger a first alert or a new alert. Indeed, alerts that are too close together are counterproductive.

[0026] Furthermore, the detection method according to the invention includes, for example, a step of estimating the reliability of at least one of the received data points, and when the estimated reliability of said at least one received data point is less than a predetermined reliability threshold, the distance calculation does not use, in the current multidimensional vector, a dimension corresponding to said at least one received data point. As an alternative embodiment of this optional step of estimating the reliability of at least one of the received data points, when the estimated reliability of said at least one received data point is less than a predetermined reliability threshold, the current multidimensional vector includes, instead of the received data point, a default value, which is taken into account in the distance calculation.

[0027] The invention also relates to a system for detecting a state of mind wandering of a vehicle driver, comprising means for receiving data representative of physiological data of the driver, data representative of a kinematic of the vehicle and / or of driver controls, and at least one data representative of a driving context, the detection system being characterized in that it comprises means for quantifying a state of mind wandering of the driver as a function of the data received by the receiving means.

[0028] The detection system according to the invention includes means for implementing the detection method according to the invention, and has advantages similar to those of the detection method according to the invention.

[0029] In one embodiment of the invention, the detection system according to the invention includes means for alerting the driver when the quantification means provide a level of driver mind wandering strictly greater than a predetermined level of mind wandering.

[0030] The detection system according to the invention may also include means for estimating the reliability of the quantification means, and means for inhibiting the alerting means based on the reliability estimated by the estimation means.

[0031] Other features and advantages of the invention will become apparent from the following description on the one hand, and from several illustrative and non-limiting examples of embodiments given with reference to the accompanying schematic drawings on the other hand, in which:

[0032] [Fig-1] represents a vehicle incorporating a detection system according to the invention, from a state of mind-wandering of a vehicle driver, in a mode of embodiment of the invention,

[0033] [Fig.2] represents steps of a detection method according to the invention, of a state of mind wandering of the driver of the vehicle of [Fig.1], in this embodiment of the invention, and

[0034] [Fig.3] represents groups of collected data, each corresponding to a level of thought wandering.

[0035] According to an embodiment of the invention represented in [Fig.1], a vehicle 1 includes a system for detecting 2 a state of mind wandering of a driver of the vehicle 1.

[0036] The detection system 2 is here a vehicle computer connected to a computer bus 10 of the vehicle 1, such as a CAN (Controller Area Network) or Ethernet bus. The detection system 2 is also connected to various physiological sensors 12, such as an eye tracker and / or a heart rate monitor. Finally, the detection system 2 is connected to an environmental data acquisition module 14, for example, a computer of an advanced driver assistance system (also called ADAS), implementing a module for modeling the driving environment of the vehicle 1. Alternatively, this computer is identical to the one implementing the detection system 2.

[0037] The modeling module uses data from the various cameras, radars and LiDARs (from the English "Light Detection And Ranging") of vehicle 1, to identify objects in the immediate environment of vehicle 1, for example within a radius of about 200 meters around vehicle 1, these objects being the road, road markings and signs, obstacles on the road, other vehicles, pedestrians, cyclists, etc.

[0038] Alternatively, the environmental data acquisition module 14 can retrieve data from the driving environment from an analysis of vehicle dynamics (speed allows, for example, the type of road to be deduced) and / or from an analysis of images acquired by one or more cameras.

[0039] The detection system 2 includes means for receiving data 20 from the computer bus 10, physiological sensors 12 and the environmental data acquisition module 14.

[0040] The detection system 2 also includes means for quantifying 22 a state of mind-wandering of the driver as a function of the data received by the receiving means 20. These quantification means 22 include at least a processor, a random access memory and a read-only memory in which is stored a computer program whose instructions, when executed on the processor, make it possible to implement the detection method 100 (referenced [Fig.2]) according to the invention of a state of mind-wandering of the driver.

[0041] According to one embodiment of the invention, the detection system 2 further comprises a means for storing a history of the data used or provided by the quantification means 22, and / or the driver's alert means 26, consisting here of a software or hardware port capable of enabling the detection system 2 to communicate with a human-machine interface of the vehicle 1, this human-machine interface comprising, for example, a screen and / or a loudspeaker, and / or a haptic device, for example integrated into the steering wheel of the vehicle 1.

[0042] We now describe, in relation to [Fig.2], the method of detecting 100 of a state of mind-wandering of the driver, according to the invention.

[0043] The detection method 100 according to the invention includes a prior learning step 102, itself comprising a first sub-step 1022 of data collection over different driving periods representative of different driving contexts.

[0044] During this first data collection substep 1022, a driving subject, who may be the driver of vehicle 1 or another driver of a different vehicle, travels on various types of roads with different characteristics, such as highways, city streets, straight or winding roads, with heavy or light traffic, pedestrians, or other obstacles, and with varying weather conditions. These different characteristics of the journey—the type of road, its congestion, and the weather conditions—form a driving context. Each driving period is associated with a driving context.

[0045] During each driving period, physiological measurements are taken on the driver using, for example, an eye tracker and a heart rate monitor, so as to obtain, for each driving period, a heart rate, a Respiratory rate and ocular and pupillometric data are collected. Ocular data relates to eye movement and includes, for example, gaze fixation time and the number of blinks per minute, while pupillometric data includes, for example, the driver's pupil diameter. Data such as age may also be collected, as younger drivers are more likely to be distracted.

[0046] During each driving period, representative environmental data are also collected, such as the position of surrounding obstacles or other road users, the type of infrastructure, or the weather conditions.

[0047] In addition, during each driving period, data are taken from the vehicle's computer bus, such as vehicle speed and / or engine torque demand, and pedal usage.

[0048] Each driving period is further associated with a level of mind-wandering in the driver, assessed by the driver themselves through questioning during that driving period, for example, via a human-machine interface. In this embodiment of the invention, the state of mind-wandering is quantified on a scale of 1 to 4 levels, with level 1 corresponding to the absence of mind-wandering, level 2 to a slight state of mind-wandering, level 3 to a well-established state of mind-wandering, and level 4 to an intense state of mind-wandering. Of course, the invention applies identically regardless of the size of the scale used.

[0049] For each driving period, the data collected during that driving period are then represented in the form of a multidimensional reference vector. Each dimension of the multidimensional reference vector corresponds to one of the collected data points. The multidimensional reference vector therefore comprises: - one or more physiological data measured on the driver, for example heart rate in beats per minute, pupil opening in millimeters, blink rate in number per minute, and possibly other physiological data such as the driver's age and gender, - one or more data points taken from the vehicle's computer bus and representative of vehicle kinematics and / or driver commands, such as speed in meters per second, engine torque demand in Newton meters, steering wheel angle in degrees, turn signal activation, etc. - one or more data points representative of a driving context, such as the inter-vehicle distance in meters, the type of road (examples: monotonous road - motorway type, demanding road - city or mountain type), lighting conditions, the position of surrounding obstacles or other road users, etc. - possibly historical data, such as the duration for which the driver has been in a state of mind-wandering, etc.

[0050] The preliminary step 102 includes a second substep 1024 of grouping the multidimensional reference vectors recorded during the first substep 1022 of collection. This grouping substep 1024 uses a grouping algorithm, here the so-called K-means algorithm, but other grouping algorithms can be used such as HCA (for Hierarchical Ascending Classification) or DBSCAN (for Density-Based Spatial Clustering of Applications with Noise).

[0051] The application of the K-means algorithm makes it possible to group the multidimensional reference vectors from the first sub-step 1022 of collection, into K groups. Figure 3 schematically represents four groups Gl, G2, G3 and G4 from the K-means algorithm, assuming for simplicity that the multidimensional vectors only include a number of heartbeats in bpm (beats per minute), a pupil diameter in mm (millimeters) and a vehicle speed in km / h (kilometers per hour).

[0052] Then, a level of mind-wandering is assigned to each group formed by the K-means algorithm, based in particular on the self-assessments made by the driving subject(s) during the driving periods associated with the group's multidimensional vectors, and on the driving contexts of these driving periods. For example, the average of the self-assessments associated with the group's multidimensional vectors is used to assign a level of mind-wandering to the group in question.

[0053] For example, the first group Gl corresponds to an average driver speed of 50 km / h, small pupil dilation, and a high heart rate, indicating sustained driver attention. This first group is therefore associated with level 1 of mind-wandering. This association is consistent with the fact that the corresponding driving contexts in this first group Gl are demanding for the driver.

[0054] The second group, G2, corresponds to an average driver speed of 30 km / h, a fairly average pupil dilation, and a relatively low heart rate. This second group is associated with level 2 of mind-wandering, which is consistent with the demanding driving contexts corresponding to this second group, G2.

[0055] The third group, G3, corresponds to an average driver speed of 70 km / h, relatively large pupil dilation, and an average heart rate. This third group is associated with level 3 of mind-wandering, which is consistent with the less demanding driving contexts associated with this third group, G3.

[0056] Finally, the fourth group, G4, corresponds to an average driver speed of 130 km / h, large pupil dilation, and a low heart rate. This fourth group is associated with level 4 of mind-wandering, which is consistent with the less demanding driving contexts associated with this fourth group, G4.

[0057] It is therefore understood that the attribution of a level of mind-wandering to each group depends on the self-assessments of the driving subject(s) corresponding to the multidimensional vectors of that group, the values ​​present in these multidimensional vectors, but also on contextual data associated with the driving periods represented by these reference multidimensional vectors.

[0058] Once the groups have been formed and each associated with a level of thought wandering, the learning step 102 is completed and the data associated with the different groups are recorded in a memory of the detection system 2.

[0059] Following the learning step 102, it is assumed that the vehicle 1 is being driven, triggering a first step 104 of the detection process 100, which is the reception by the receiving means 20 of physiological data from the physiological sensors 12, data from the computer bus 10 and data provided by the environmental data acquisition module 14. This latter data are for example objects detected in the immediate environment of the vehicle 1, allowing the quantification means 22 to determine a driving context, or directly a driving context determined by the environmental data acquisition module 14.

[0060] The received data allows the formation of a current multidimensional vector V (referenced [Fig.3]) in the same multidimensional space as the reference multidimensional vectors used in the training step 102 to form the groups, for example groups G1 to G4. In other words, at least part of the received data corresponds to the same types of data as those used to form the groups resulting from the training step 102.

[0061] Following the first reception step 104, the detection process 100 may include a second estimation step 106 of the reliability of the received data. This estimation step 106 uses, for example, reliability data from physiological sensors 12. For example, if the eye tracker is not functional, it sends a confidence estimate in its measurement close to 0 (the measurement being (The closer the confidence estimate is to 1, the more reliable the estimate is, and the closer the confidence estimate is to 0, the less reliable it is). This information is recorded, for example, during this estimation step 106 to implement the next, third step 108 of the detection process 100 in a degraded manner, or to halt the detection process 100 while awaiting repair of the eye tracker. However, in this example of use of the invention, it is assumed that the physiological sensors provide reliable measurements.

[0062] The third step of the detection process 100 is the quantification 108 of the driver's mind-wandering state.

[0063] This third quantification step 108 may include a prior substep of normalization of the current multidimensional vector V. In this case, the reference multidimensional vectors of each group G1 to G4 are also normalized, in the same way, only once during the learning step 102.

[0064] This normalization uses, for example, the following formula, which allows all coordinates between 0 and 1 to be normalized:

[0065] xfmin(x) with x; a coordinate of the current multidimensional vector V or of the niax(x)-min(x) multidimensional reference vector considered, corresponding to a data type coded on one dimension, min(x) the minimum value of the coordinates of the multidimensional reference vectors relative to this dimension, and max(x) the maximum value of the coordinates of the multidimensional reference vectors relative to this dimension.

[0066] We can of course choose another normalization method.

[0067] In the following, the current multidimensional vector V and the groups G1 to G4 considered are normalized, that is to say, we work on coordinates between 0 and 1.

[0068] Then, in this third quantization step 108, the quantization means 22 calculate, for each k-th group of reference multidimensional vectors constituted during the learning step 102, a distance Dk between the current multidimensional vector V and this k-th reference group. This distance Dk is, for example, calculated as a distance between the current multidimensional vector V having coordinates (xb ..., xN), and the center of the reference group, with coordinates (xmi,...,xmN), the coordinates of the current multidimensional vector V and the center of the reference group being weighted by weights Wi to wN, in this distance calculation:

[0069] Dk =

[0070] The weights Wi to wN can, for example, all be equal to one. In this case, the distance Dk is a Euclidean distance. In this embodiment of the invention, some weights can be strictly greater than or less than one, so as to give more importance, in the evaluation of the driver's level of mind wandering, to certain data, such as, for example, physiological data, environmental data, or historical data.

[0071] The values ​​of the weights Wi to wN are, for example, modified according to the environment in which the vehicle is traveling. Thus, when the driver is on a highway, the physiological and behavioral data (from the computer bus, such as engine torque demand, distance to the vehicle in front) can be weighted by a weight strictly greater than one, for example, a value of two or three. This makes it possible to model the fact that reaching a high level of mind-wandering is faster in a highway environment.

[0072] On the contrary, since meteorological data have little influence on entering a state of heightened mind-wandering in a motorway environment, they can, for example, be weighted by a weight strictly less than one, for example 0.5, when the driver is operating in such an environment.

[0073] The smallest calculated distance Dk is between the current multidimensional vector V and the group to which the current multidimensional vector V is assigned. In this example of use of the invention, the smallest calculated distance is D4, therefore the current multidimensional vector V is assigned to group G4.

[0074] The driver's mind-wandering state is quantified at level n of the group to which the current multidimensional vector V is assigned, in other words at level 4, in this example of an embodiment of the invention.

[0075] It should be noted that when one of the physiological sensors is not reliable, the dimension corresponding to this sensor is not used, for example, in the calculation of the distances between the current multidimensional vector V and each of the groups G1 to G4 from the learning step 102, which makes it possible to obtain a level of mind wandering of the driver, independently of the feedback from this physiological sensor, or the weight value associated in the distance calculation can be adjusted according to the reliability of the measurement.

[0076] The quantification step 108 is followed by a fourth comparison step 110 of the level n of the mind-wandering state as quantified by the quantification means 22 in the previous step, with a predetermined level ns of mind-wandering, with for example ns equal to 2.

[0077] When the level n is strictly greater than this predetermined level ns (branch Y), therefore equal to three or four, the detection system 2 alerts the driver, in a Fifth step 112 of the detection process, by means of an alert signal indicating a high state of mind-wandering in the driver. For example, the detection system 2 uses a haptic device to vibrate the steering wheel, and / or displays a message on a screen in vehicle 1, and / or emits an audible message indicating to the driver that they are in such a state.

[0078] On the contrary, when level n is less than or equal to this predetermined level ns (branch N), no alert is issued.

[0079] The quantification step 108 can be followed by a recording step 114 of the driver's thought wandering level n, and a reliability level for this quantification, in the storage means 24. The reliability level of this quantification can be linked to the reliability of the sensors used. The recording step 114, for example, takes place in parallel with the comparison step 110.

[0080] Then the detection process 100 loops back to the reception step 104 periodically, for example every thirty seconds.

[0081] It is therefore understood that the detection system 2 is capable of quantifying a state of mind-wandering of the driver, of estimating the reliability of this quantification, of evaluating the duration during which the driver is in this state, and of alerting the driver in a timely manner, depending on the level n of mind-wandering of the driver, the evaluated duration and the estimated reliability.

[0082] Moreover, this embodiment of the invention is easily implementable in the vehicle, allows for a rapid and real-time calculation of the driver's mind-wandering state, and a personalization of an advanced driver assistance system incorporating the invention, which facilitates the acceptability of the alerts of the detection system according to the invention.

[0083] Of course, the invention is not limited to the examples just described and many modifications can be made to these examples without departing from the scope of the invention.

Claims

Demands

1. A method for detecting (100) a state of mind-wandering of a vehicle driver, comprising a receiving step (104): - of at least one data point representing physiological data of the driver, - of at least one data point representing vehicle kinematics and / or driver controls, and - of at least one data point representing a driving context, the detection method (100) being characterized in that it comprises a quantification step (108) of a state of mind-wandering of the driver as a function of the received data.

2. Method for detecting (100) a mind-wandering state according to claim 1, wherein the quantification step (108) includes a step of associating the received data, by calculating at least one distance (Dk) between the received data and a group (G4) from a set of groups (G1, G2, G3, G4) associated with levels of mind-wandering, the mind-wandering state of the driver being quantified by the level (n) of the group (G4) to which the received data are associated.

3. A method for detecting (100) a state of mind-wandering according to claim 2, comprising a preliminary learning step (102) comprising substeps of: - collecting (1022), over each driving period of a plurality of driving periods, at least one data point representative of physiological data of a driving subject over the driving period, at least one data point representative of the kinematics of the vehicle driven by the driving subject and / or of the driving subject's controls over the driving period, and at least one data point representative of a driving context over the driving period, the collected data being represented over each driving period by a multidimensional reference vector, - grouping (1024) the multidimensional reference vectors into a plurality of groups (G1, G2, G3, G4) each associated with a level of mind-wandering,the quantification step (108) involving the assignment of a multidimensional vector representative of the received data, said, current multidimensional vector (V), to one of the groups (G4) of the plurality of groups (Gl, G2, G3, G4), by calculating distances (Dk) between on the one hand the current multidimensional vector (V) and on the other hand each of the groups (Gl, G2, G3, G4) of the plurality of groups (Gl, G2, G3, G4), the current multidimensional vector (V) being associated with the group (G4) of the plurality of groups (Gl, G2, G3, G4) for which the calculated distance (D4) is the smallest.

4. Method for detecting (100) a mind-wandering state according to claim 3, wherein the data collected during a driving period, or the data received, includes at least one of the following: - driver's heart rate, - driver's pupil opening, - driver's gaze direction, - driver's eyelid blink rate, - vehicle speed, - vehicle engine torque demand, - steering wheel angle, - inter-vehicle distance, - brightness, - duration for which the driver has been in a mind-wandering state, - positions of surrounding obstacles or other road users, and - road type.

5. Method for detecting (100) a state of mind-wandering according to claim 3 or 4, wherein the calculation of distances comprises, for each group of the plurality of groups (G1, G2, G3, G4), a calculation of a weighted sum of differences between coordinates of the current multidimensional vector (V) and coordinates of one of the groups (G1, G2, G3, G4) of the plurality of groups.

6. Method for detecting (100) a mind-wandering state according to any one of claims 1 to 5, comprising an alert step (112) to the driver via a human-machine interface, of a high level of mind-wandering, when the mind-wandering level from the quantification step (108) is strictly greater than a predetermined level (ns) of mind-wandering.

7. Method for detecting (100) a mind-wandering state according to claim 6, wherein the data reception step (104) comprises receiving a history of alerts previously triggered during previous alert steps (112) to the driver of a high level of mind-wandering.

8. A system for detecting (2) a state of mind-wandering of a vehicle driver (1), comprising means for receiving (20) data representative of physiological data of the driver, data representative of vehicle kinematics and / or driver controls, and at least one data representative of a driving context, the system for detecting (2) being characterized in that it comprises means for quantifying (22) a state of mind-wandering of the driver as a function of the data received by the receiving means (20).

9. A detection system according to claim 8, comprising driver alerting means (26) when the quantification means (22) provide a level (n) of driver mind-wandering strictly greater than a predetermined level (ns) of mind-wandering.

Images