WO2025223820A1 - Method for identifying a person in the vicinity of a motor vehicle - Google Patents

Abstract

The invention relates to a method for identifying a person in the vicinity of a motor vehicle using sensors arranged on the vehicle, the method comprising the following steps when the person approaches the motor vehicle: a) detecting their gait by means of the sensors, analysing at least one gait-related parameter, comparing it with previously recorded and established gait models and calculating a gait recognition score; b) detecting the face of the person by means of the sensors, analysing at least one parameter of the face, comparing it with previously recorded and established face models and calculating a face recognition score; c) normalising the scores according to reference scores; d) fusing the normalised gait recognition and face recognition scores in order to identify whether or not the person is associated with the motor vehicle.

Description

Description

Title: Method for identifying an individual near a motor vehicle

technical field

The present invention relates to a method for identifying an individual in the vicinity of a motor vehicle, in particular for the purpose of authenticating him in order to determine whether the individual is associated or not with said motor vehicle.

Previous technique

Access to vehicles via connected devices is a technology deployed in automobiles. It emerged thanks to the evolution of wireless networks, Bluetooth connectivity, and Internet of Things (IoT) technologies. Vehicle access via connected devices is primarily used for simple functions, such as remotely unlocking vehicle doors using a smartphone or remote control.

Renault, through its Virtual Key project based on Bluetooth Low Energy (BLE) technology, offers this service for customer convenience. The Virtual Key solution relies on the creation of a physical device containing a BLE module that interacts with a wireless remote control from the Renault range. Access to the vehicle's doors and windows is via this remote control, which already has all the necessary pairing mechanisms with the vehicle, and in particular the module that operates the vehicle's doors and windows. This technology exists in the prior art.

It is also possible to manage access to your car with your smartphone rather than your key, with interoperability across the vast majority of smartphone models worldwide. Accessing vehicles via connected devices is therefore a mature method of vehicle access.

Access to the vehicle without a connected device could be considered, so that the vehicle owner would no longer need a key or code to unlock their vehicle. The vehicle would recognize people as they approach. identified as legitimate and can thus unlock the doors and then welcome them without any constraints for them.

Application KR 20220144186 describes a biometric authentication method that uses an individual's gait pattern as a means of identification, along with a two-dimensional composite multiplier neural network model for classification.

US application 2022/0201389 discloses a method for locating a primary user performing a remote control operation of a vehicle and uses amplitude-modulated (AM) ultrasonic communication via an ultrasonic beamforming device.

Application EP 4 139 177 describes a system for assessing the risk posed by a person approaching a vehicle equipped with surveillance cameras.

US application 2019/0051069 discloses a user recognition system for automated recognition in an autonomous vehicle that includes an environmental sensor, specifically to recognize a stop request gesture from the user.

International application WO 2017/176618 is known to use gesture biometrics and heart biometrics to determine whether a user is authorized to use the vehicle.

Application EP 3 342 097 describes a device managing biometric information, allowing the user to register new biometric information, such as a fingerprint and information from a connected device.

US patent application 2020/0193005 discloses a method for unlocking a vehicle when the person approaching the vehicle matches an authorized individual. Images are acquired, and the person's gait and facial features are determined and compared to stored data.

Description of the invention There is therefore a need to further improve the means of selecting and identifying an individual in the vicinity of a motor vehicle in order to facilitate possible access to said vehicle.

Summary of the invention

Identification method

The present invention addresses this need by means of, according to one of its aspects, a method for identifying an individual in the vicinity of a motor vehicle, using a plurality of sensors arranged on said vehicle, the method comprising at least the following steps when said individual approaches said motor vehicle: a) detecting his gait by means of at least one of said plurality of sensors, analyzing at least one parameter related to the gait, comparing it to previously established and recorded gait models and calculating a gait recognition score, b) detecting the face of said individual by means of at least one of said plurality of sensors, analyzing at least one parameter of said face, comparing it to previously established and recorded face models and calculating a facial recognition score, c) normalizing said scores according to reference scores, and d) merging said normalized gait recognition and facial recognition scores in order to obtain an identification score to identify whether or not the individual is associated with said motor vehicle.

Thanks to this invention, it is possible to select and detect a pedestrian walking towards the vehicle who could potentially be the vehicle's owner. The vehicle is equipped with a sophisticated system that allows it to estimate the user's trajectory and secure reliable biometric identification.

The invention makes it possible to identify an individual based in particular on their way of walking, that is to say their gait. This gait This can be considered a distinctive characteristic of the person, as it is unique and can be used to authenticate them. The identification and authentication process involves a thorough analysis of how the person moves, including foot placement, arm movements, walking speed, and pace.

The merging of the two recognition techniques improves the accuracy of person identification and authentication by using information from each technique to compensate for any limitations of either. This combination results in more reliable and precise person recognition, reducing the risk of errors or uncertainties in identification.

The fusion system ensures vehicle security and protection by guaranteeing access only to authorized personnel. To achieve this, a final identification and authentication system is implemented before each individual is allowed to enter their vehicle. This system utilizes facial and gait recognition technologies, which the fusion system uses to verify the individual's identity by comparing these elements to a pre-populated database. Once the double recognition is successfully completed, access to the vehicle is granted. This security procedure guarantees the confidentiality and security of vehicles, preventing unauthorized access.

A vehicle owner can access their vehicle freely, without needing any connected devices. Thanks to this invention, the user can enter their vehicle completely naturally, without encountering any constraints or difficulties. In other words, there's no need to carry a key, a badge, or any other unlocking device, as the access system is fully automated and personalized.

The process according to the invention makes it possible to minimize the power consumed necessary for the implementation of its steps.

When the vehicle's electronic system processes data, it can consume significant power. Therefore, to preserve battery range, it is important to reduce power consumption. This consumption is maximized. In the case of the invention, to optimize power consumption, data processing is only activated when truly necessary. Thus, each step of the process is only advantageously initiated if the preceding step has validated its necessity. This approach therefore saves battery energy by avoiding unnecessary processing. The vehicle's electronic system, particularly the multimedia system, only uses its resources at the appropriate time, selectively, according to the needs of the task at hand. This technique aims to optimize battery life, which is crucial.

Furthermore, optimizing power consumption improves the energy efficiency of the multimedia system, which can also have significant environmental and economic implications. Indeed, reducing energy consumption helps to lessen the environmental impact of energy production and transmission activities.

Step b) of detecting the individual's face advantageously uses the analysis of at least one image of said individual, taken by at least one of said plurality of sensors.

In a preferred embodiment, at least one machine learning algorithm, including at least one neural network, is used to establish gait models and face models.

Step a) can begin when an individual is detected near the vehicle at a predefined minimum distance, in particular between 2 and 5 meters.

The method according to the invention may further include an authentication step of the individual in order to authorize or deny them access to the vehicle, in particular by means of an authentication score.

Biometrics of the gait

Gait biometrics is a biometric technology that focuses on the unique characteristics of a person's gait. It is based on the fact that each person has a unique way of walking. Gait recognition biometrics is a subset of behavioral biometrics, in which subjects can be identified by their gait. The theory Behind this recognition system lies the fact that each person has a unique gait. It's common for a familiar person to be recognized from a distance by their walk. This is one of the few recognition methods capable of identifying people remotely.

A person's gait is as unique as the timbre of their voice. Thanks to this knowledge, gait recognition technology can be implemented using machine learning (ML) algorithms. Today, identifying individuals by their gait is being deployed for the following reasons:

- Recognition of the approach works well remotely,

- Gait recognition can be performed using low-resolution video and simple instrumentation.

- Recognition of the approach can occur without the cooperation of individuals.

- Gait recognition may work well while other features such as faces and fingerprints are not available.

- the characteristics of the approach are generally difficult to imitate.

The analysis of the approach focuses advantageously on a range of personal characteristics, at least as follows and in a non-exhaustive manner:

- Walking speed: Walking speed is the distance covered by an individual during a given period of time.

- Stride length: Stride length corresponds to the distance traveled by a foot between two successive contacts with the ground,

- Foot contact time with the ground: Foot contact time with the ground is the time during which a foot is in contact with the ground during a stride,

- Swing time: Swing time is the duration for which one foot is in the air during a stride.

- Stride length: stride length is the distance covered by both feet during one stride, - Range of motion: Range of motion is the distance traveled by a body limb during normal walking.

- Symmetry of gait: Symmetry of gait is the similarity between the movements of the left and right body during walking,

- ankle angle: the ankle angle is the angle formed between the foot and the leg during walking,

- Pelvic movement: pelvic movement is the amplitude of pelvic movement during walking,

- Walking cadence: Walking cadence is the number of steps taken by an individual during a given period of time.

- Knee angle: The knee angle is the angle formed by the leg and thigh during walking.

- Height: noting that the person being sought is very tall or very short can help with the selection process.

- Hip height: Hip height is the distance between the hip and the ground during walking.

These characteristics can then be analyzed using at least one machine learning-based pattern recognition algorithm to create a unique model of the individual's gait.

Such an algorithm can use a deep neural network architecture to identify patterns and relationships between gait data and individual characteristics. Once the individual's gait model has been created, it can be used to recognize that person by comparing their gait characteristics to those of the model, identifying them if their gait recognition score is above a predetermined threshold.

Facial biometrics

Facial recognition is an identification and authentication method that uses facial features to uniquely identify a person. It is based on the analysis of the physical characteristics of the facial features such as interocular distance, jaw shape, nose-to-mouth distance, and cheek and eyebrow shape.

The authentication process, for its part, consists of verifying that the identified person is indeed present and that, consequently, the system is not dealing with fraud by photo, video or mask, in particular.

Facial biometrics relies in particular on deep machine learning techniques that provide accurate and robust facial recognition. To achieve this, machine learning algorithms analyze a large number of facial images to identify common features and build a model that can detect and extract the same features in new images.

This model can then be used to compare an unknown face image to previously established and recorded face models, in order to identify the person when their facial recognition score is above a predetermined reference threshold.

The invention involves facial biometrics outside the vehicle. To achieve this, it must be able to function in a variety of environmental conditions, such as sunlight, wind, rain, or snow. It captures a person's face tracking in real-time video, even when that person is moving towards the vehicle.

Fusion

The standardized scores for gait recognition and facial recognition are merged.

Each biometric device behaves advantageously as a detection source, reporting whether a single hypothesis is true or false, and similarly for the opposing hypothesis. This means that these detectors produce a binary response based on a decision threshold: either they indicate detection (in our case, biometric detection), or identification (i.e., the hypothesis is true), or they indicate no detection (i.e., the hypothesis is false). They can therefore generate the following four situations: - A "true positive" is a situation in which a test or model correctly predicts the existence of a hypothesis. In other words, a true positive is when the test result is positive and the tested condition actually exists.

- A "false positive" occurs when a test hypothesis is wrongly rejected, i.e., it is reported as positive when it is actually negative.

- A "false negative" occurs when a test hypothesis is wrongly accepted, i.e., it is reported as negative when it is actually positive.

- A "true negative" occurs when a test hypothesis is correctly accepted as negative. This means that the test correctly indicates the absence of the targeted condition.

This explains why, in statistics and machine learning, various metrics such as accuracy, recall, and F1 score can be used to evaluate model performance measures.

We define in a well-known way:

- The false positive rate (FPR) is the probability that a detection test incorrectly indicates the presence of a condition that is actually absent. It represents the proportion of false positive errors in detection tests.

- The false negative rate (FNR) is the probability that the detection method fails to correctly identify a condition that is actually present. It represents the proportion of false negative errors in detection tests.

In the invention, the biometric detection system can be considered bimodal because it uses two biometrics that combine their detection scores. Therefore, score fusion applies.

A machine learning algorithm system for bimodal fusion biometric detection, and optionally more, can be used. This system is advantageously made to learn in order to improve itself. This minimizes the system's energy consumption.

In its general definition, score fusion is the process of combining the results of multiple classifiers or detectors to produce a more accurate and reliable estimate. The rationale for score fusion is that different models or classifiers utilize different characteristics. or different processing methods to represent the same object or phenomenon. Merging the scores of each model or classifier provides a more robust, complete, and accurate estimate.

Score fusion is the most common approach because scores generated by different detectors can be easily accessed and combined. This fusion is advantageously preceded by a score normalization step.

The established gait and facial patterns are preferably recorded beforehand in one or more reference databases. These patterns can then be used to compile statistics.

Distributions and histograms of scores from tests of legitimate and illegitimate or imposter applications are advantageously calculated.

According to the invention, the fusion of said standardized gait recognition and facial recognition scores is carried out by linear combination.

Preferably, the following function is applied to the sets of gait detection #1 and face detection #2 scores to obtain the fusion statistic S:

[Math 1]

With i :1..N, N the number of scores from detections #1 and #2, s1 the N normalized scores from detection #1, s2 the N normalized scores from detection #2, oc e [0, 1]

In this preferred mode of the invention, the fusion scores are thus constructed by linear combination.

This fusion is advantageously optimized by seeking the a that minimizes the associated Equal Error Rate (or EER):

[Math 2] oc e [0, 1], o min | EER( <K,Ù minimum

The gait recognition score and the facial recognition score are advantageously compared to predetermined reference thresholds. Sensors

The method according to the invention may use at least one video sensor, in particular one or more cameras, and may also utilize a radar sensor and/or an audio sensor. These sensors not only detect movements near the vehicle but also capture videos and images. This combination of sensors offers considerable potential for identifying a person outside the vehicle by providing precise information about their physical appearance, position, and movements.

Exterior vehicle video sensors are monitoring devices that capture images and videos of the surrounding environment. These sensors can be located in various places on the vehicle, such as the windshield, sides, and rear of the body. The primary purpose of exterior vehicle video sensors is to improve driver safety and visibility by detecting obstacles and potential hazards to help avoid collisions.

External radar sensors on a vehicle are sensors that use radio waves to detect objects and obstacles in the surrounding environment. The radar sensor works by establishing its Radar Cross Section (RCS), which is a measure of an object's ability to reflect the radio waves emitted by a radar. The RCS is determined by several factors, such as the object's size, shape, texture, and composition. In the context of automotive radar sensors, the RCS is used to assess an object's ability—potentially a pedestrian—to be detected by radar and to help determine the object's distance, speed, and direction relative to the vehicle. These sensors can be placed in various locations on the vehicle, such as the front or rear bumper. The primary benefit of external radar sensors on a vehicle is to improve safety by detecting obstacles and potential hazards on the road. These radar sensors are particularly useful in situations of limited visibility, such as at night, in fog, or in rain. Radar sensors can detect the presence of moving or stationary objects near the vehicle, such as cars, pedestrians, animals, or obstacles. This type of sensor can therefore be used to improve driver assistance features. Any type of video, radar or other sensors can be used.

Specific identification

The method according to the invention may include a specific identification or re-identification step, which consists of identifying an individual who has already been previously identified, and the time interval between these two identifications is sufficiently short to reasonably assume that their fundamental characteristics have not changed between them. These fundamental characteristics, considered as prior knowledge, can then significantly assist other identification systems. It is thus possible to locate a specific person in a scene using information previously collected during their initial identification. This technique relies on the use of highly discriminating criteria, including:

- Skin color: this information allows us to differentiate people according to their skin tone, which can be useful for identifying a person in a scene where several individuals look alike.

- Gender: a characteristic that allows us to distinguish people according to their male or female sex.

- Estimating a person's height: a criterion that can help identify a person within a group by taking into account their estimated height.

- the general shape of the person's morphology: this characteristic allows for the identification of individuals by taking into account their overall silhouette, their build or their posture, pregnant woman,

- Clothing worn: this information can help differentiate people based on their attire.

- Accessories worn, such as hats, bags, or glasses: these distinctive items can help identify a person in a scene,

- the person's distinctive movements or gestures: particularities such as a limp, a swaying gait or a specific gesture can also help to identify a person in a scene. Thanks to the combination of these discriminating criteria, the re-identification technique makes it possible to sort a specific person in a scene, even if they are not in the same position or context as during their first identification.

These criteria are advantageously associated with classes of a model based on deep machine learning techniques allowing the detection of discriminating criteria via different classes that we seek to extract from the person to be identified.

Other recognitions

The method according to the invention may further include a step of recognizing the individual's voice by means of at least one of said plurality of sensors, said voice being compared to previously recorded voice models. Identification and authentication are thus carried out by voice biometrics, for example by asking the user to pronounce a secret phrase that would have previously enrolled them.

At least one of the sensors can therefore be an external microphone to capture the individual's voice so that they can be identified by voice biometrics.

The method according to the invention may further include a step of recognizing a specific gesture of the individual. The user can perform a specific gesture known only to them, which would serve as additional authentication to secure access to the vehicle.

Estimating the trajectory of individuals

The method according to the invention may include a preliminary step of predicting the trajectory of individuals; if the detection is positive, indicating that the individual is approaching the vehicle, steps a) to d) are carried out.

This step involves estimating the pedestrian's trajectory towards the vehicle. This step triggers the more complex steps described earlier. This step must also be carried out with minimal energy consumption, as it operates in standby mode.

The pedestrian trajectory prediction problem consists of predicting where and in which direction the pedestrian will be in the future using information on the pedestrian and the environment. Predicting the trajectory of pedestrians is complex due to the uncertainty of their interaction with the environment.

There are two known methods for predicting pedestrian trajectories: methods that involve building a model of the pedestrian's kinematics to predict their trajectory, and more recently, prediction-based methods using deep learning. However, since this step may be the first to be implemented during the vehicle's standby cycle, it must consume minimal energy resources. The invention advantageously uses a simple kinematic trajectory estimation rather than deep learning processing, which would be too resource-intensive at this stage.

The pedestrian kinematics model is a physical method for predicting a pedestrian's trajectory. It uses information such as the change in the pedestrian's position, initial velocity, acceleration, and direction.

The method according to the invention advantageously calculates the estimation and confirmation of a trajectory. Simply establishing that a pedestrian is heading towards the vehicle is sufficient to implement steps a) to d) of the method.

Learning models

In one embodiment of the invention, the models of both gait and facial biometrics are updated in real time. Each time an individual is correctly identified and authenticated, the video data that enabled this decision is advantageously used to generate new feed data, such as validation or development data. This data is distinct from the training data used to calculate the reference model values and the test data used to measure the model's accuracy. The updated model can be evaluated against the test data after training is complete to measure its final performance.

Two machine learning algorithms can be used in parallel with those used to establish gait and facial models, in order to enhance future detections. The performance of the gait and facial biometric detection models of the invention is improved. continuously using this new power data from the experiences of already authenticated users.

Device

According to another aspect, the invention relates to a device for identifying an individual in the vicinity of a motor vehicle, the device comprising a plurality of sensors arranged on said vehicle and being configured to perform at least the following steps when said individual approaches said motor vehicle: a) detect his gait by means of at least one of said plurality of sensors, analyze at least one parameter related to the gait, compare it to previously established and recorded gait models and calculate a gait recognition score, b) detect the face of said individual by means of at least one of said plurality of sensors, analyze at least one parameter of said face, compare it to previously established and recorded face models and calculate a facial recognition score, c) normalize said scores according to reference scores, and d) fuse said normalized gait recognition and facial recognition scores in order to obtain an identification score to identify whether or not the individual is associated with said motor vehicle.

The device advantageously includes at least one video sensor, in particular one or more cameras. The device may also include a radar sensor and/or an audio sensor.

The characteristics stated in relation to the process apply to the device and vice versa.

Motor vehicle

According to another aspect, the invention relates to a motor vehicle comprising a powertrain and at least one identification device according to the invention.

The characteristics stated in relation to the process apply to the vehicle and vice versa. Brief description of the drawings

The invention will be better understood by reading the detailed description that follows, a non-limiting example of its implementation, and by examining the attached drawing, in which

[Fig 1] Figure 1 illustrates an example of the implementation of the invention,

[Fig 2] Figure 2 represents a flowchart illustrating the example of implementation of the invention of Figure 1,

[Fig 3] Figure 3 illustrates an example of the implementation of the invention on a motor vehicle,

[Fig 4] Figure 4 illustrates the recognition of the approach according to the invention,

[Fig 5] Figure 5 illustrates facial recognition according to the invention,

[Fig 6a] [Fig 6b] [Fig 6c] [Fig 6d] [Fig 6e] [Fig 6f] Figures 6a to 6f are histograms and curves illustrating the principle of fusion,

[Fig 7] Figure 7 illustrates an example of the implementation of part of the invention,

[Fig 8] Figure 8 represents a flowchart illustrating the example of implementation of the invention of Figure 7,

[Fig 9] Figure 9 illustrates another example of an implementation of the invention, and

[Fig 10] Figure 10 illustrates a variant embodiment of Figure 9.

Detailed description

Figure 1 illustrates an example of implementation of the invention.

In step 0, the vehicle is parked, its sensors are ready to operate. An individual approaches the vehicle. In step 1, where the individual is standing at a distance greater than 5 meters, the basic pedestrian trajectory estimation begins, as described previously.

In step 2, if the individual approaches within a distance of less than 5 meters, specific identification or re-identification is triggered.

In step 3, the individual's gait is detected by means of a plurality of sensors, visible in Figure 3. At least one parameter related to the The individual's gait is analyzed and compared to previously established and recorded gait models. A gait recognition score is calculated. These steps are also performed for facial biometrics, with a facial recognition score being calculated. These scores are then normalized.

In steps 4 and 5, the standardized gait and facial recognition scores are merged to identify and authenticate the individual as being associated with the vehicle. Identification and authentication scores are then calculated.

In step 6, specific identification or re-identification can be triggered again.

Step 7 corresponds to an optional step of recognizing the individual's voice by means of at least one of said plurality of sensors, said voice being compared to previously recorded voice models, and to an optional step of recognizing a particular gesture of the individual.

In step 8, access to the vehicle may be granted to the individual if they have been properly identified and authenticated.

Figure 2 is a flowchart illustrating the invention and is broken down as follows:

1) Original situation in which the vehicle is parked.

2) Estimation of the trajectory of a pedestrian moving towards the vehicle.

3) If a pedestrian is detected moving towards the vehicle, the processing continues; otherwise, it returns to trajectory estimation.

4) If it is the first time, there is no possible comparison for a new identification.

5) Otherwise, the specific identification module establishes the instantaneous current classes.

6) If there is consistency with the previously established classes, the person therefore has the same discriminating characteristics as in the previous identification, the processing can continue; otherwise, return to trajectory estimation.

7) Biometric processing by gait and facial recognition.

8) Merging of scores. 9) Reporting of the identification score and the authentication score.

10) If there is identification and authentication, the person is the expected one; otherwise, back to trajectory estimation.

11) The specific identification module establishes the instantaneous current classes and transmits them to the same pre-processing module to serve as reference classes.

12) Optional biometrics requiring user intervention such as voice or gesture biometrics.

13) If there is identification and authentication, the person is the expected one; otherwise, back to trajectory estimation.

14) The identified and authenticated person is legitimately entitled to access their vehicle.

Figure 3 shows the sensors and their location on the motor vehicle.

Figure 4 illustrates gait biometrics, and Figure 5 illustrates facial biometrics, as described previously. The internal architecture of the neural networks used is indicative and varies according to the needs and expected performance.

Figure 6a illustrates the possible responses provided by the sensors, as described previously. The established gait and facial models are preferably pre-recorded in one or more reference databases. These models can then be used to generate statistics.

Distributions and histograms of scores from tests of legitimate and illegitimate or imposter applications are advantageously calculated.

Figure 6b illustrates an example of a histogram for a first biometric system #1 that could be assigned to facial biometrics and its associated DET curve, with ‘tar’ for ‘target’ representing the histogram of legitimate test scores, and ‘nontar’ for ‘non-target’ representing the histogram of sham test scores. The red perpendicular bisector line intersects the EER.

Figure 6c illustrates an example of a histogram of a first biometric system #1 that could be assigned to the biometrics of gait and its associated DET curve, with the same notations as for Figure 6b. By combining the two DET curves on the same graph, as shown in Figure 6d, we observe that the DET curve for detection system #1 is lower than that of detection system #2, with EER#1 <EER#2. This is because the statistics for detection system #1 have a smaller overlap area between its two histograms than those for detection system #2. Therefore, detection system #1 is more selective than detection system #2.

Figure 6e shows the EER of the fusion of the two biometrics. In the illustrated example, a minimum EER is observed for an oc_min of 0.62. The linear combination fusion is therefore optimal for this a_min. As shown in Figure 6f, the optimal fusion produces a DET curve lower than the two DET curves of detection systems #1 and #2, with a lower minimum EER. The optimal fusion is therefore the most selective. This fusion with this oc_min generates a DET curve lower than each of the two DET curves of detection systems #1 and #2, but also lower than all those resulting from the same fusion with a different oc_min. This fusion with this oc_min is therefore the most selective.

Figure 7 illustrates an example of trajectory detection using the front left side radar. The ego vehicle is parked. The ego vehicle's front side radar receives the echo from the target person. If the person's trajectory simply converges towards the ego vehicle's radar, this triggers the more complex identification steps according to the invention.

Figure 8 is a flowchart illustrating the processing for a radar and is broken down as follows:

1) The radar in functional standby mode has a sampling rate of F=5 Hz.

2) The radar receives the "Radar Cross Section" or (RCS radar cross section)

3) The radar establishes the polar coordinates of visible targets for a maximum of 10.

4) For all target groups.

5) If the target is less than 5 meters away, it is of interest; otherwise, it is not of interest.

6) Calculation of the variation in its distance from the vehicle.

7 and 8) If the distance of the target to the vehicle decreases, it approaches the vehicle; otherwise, it moves away from the vehicle and is of no interest. 9) As the target approaches the vehicle, more complex processes are activated to identify the person.

Figure 9 illustrates another example of implementing the invention, using machine learning algorithms (3’) in parallel with those used to establish gait and face models. In step 8’, this data is transmitted to enhance subsequent detections.

In this example, since this processing is deferred, it is possible to perform it via the cloud, as illustrated in Figure 10.

The invention is not limited to the examples just described. In particular, the arrangement and number of sensors on the vehicle may differ.

Furthermore, the neural networks used may differ.

Claims

Demands 1. A method for identifying an individual in the vicinity of a motor vehicle, using a plurality of sensors arranged on said vehicle, the method comprising at least the following steps when said individual approaches said motor vehicle: a) detecting their gait by means of at least one of said plurality of sensors, analyzing at least one parameter related to the gait, comparing it to previously established and recorded gait models and calculating a gait recognition score, b) detecting the face of said individual by means of at least one of said plurality of sensors, analyzing at least one parameter of said face, comparing it to previously established and recorded face models and calculating a facial recognition score, c) normalizing said scores according to reference scores, and d) merging by linear combination said normalized gait recognition and facial recognition scores in order to obtain an identification score to identify whether or not the individual is associated with said motor vehicle. 2. A method according to claim 1, wherein step b) of detecting the individual's face uses the analysis of at least one image of said individual, taken by at least one of said plurality of sensors. 3. A method according to any one of the preceding claims, wherein step a) begins when an individual is detected near the vehicle at a predefined minimum distance, in particular between 2 and 5 meters. 4. A method according to any one of the preceding claims, using at least one video sensor, in particular one or more cameras, the method further using in particular a radar sensor and/or an audio sensor. 5. A method according to any one of the preceding claims, wherein at least one machine learning algorithm, in particular using at least one neural network, is used to establish gait models and face models. 6. A method according to any one of the preceding claims, wherein the gait recognition score and the facial recognition score are compared to predetermined reference thresholds. 7. A method according to any one of the preceding claims, further comprising an authentication step of the individual in order to authorize or deny them access to the vehicle, in particular by means of an authentication score. 8. A method according to any one of the preceding claims, further comprising a step of recognizing the individual's voice by means of at least one of said plurality of sensors, said voice being compared to previously recorded voice models. 9. A method according to any one of the preceding claims, further comprising a step of recognizing a particular gesture of the individual. 10. A method according to any one of the preceding claims, comprising a preliminary step of predicting the trajectory of individuals; if the detection is positive, indicating that the individual is approaching the vehicle, steps a) to d) are carried out. 11. A device for identifying an individual in the vicinity of a motor vehicle, the device comprising a plurality of sensors arranged on said vehicle and configured to perform at least the following steps when said individual approaches said motor vehicle: a) detect their gait by means of at least one of said plurality of sensors, analyze at least one parameter related to the gait, compare it to a) detect the face of said individual by means of at least one of said plurality of sensors, analyze at least one parameter of said face, compare it to previously established and recorded face models and calculate a facial recognition score, c) normalize said scores according to reference scores, and d) merge said normalized gait recognition and facial recognition scores by linear combination in order to obtain an identification score to identify whether the individual is associated with said motor vehicle or not. 12. Device according to the preceding claim, comprising at least one video sensor, in particular one or more cameras. 13. Device according to claim 11 or 12, further comprising a radar sensor and/or an audio sensor. 14. A motor vehicle comprising a powertrain and at least one identification device according to any one of claims 11 to 13.