System for determining a state of presence within a vehicle

Abstract

Title: System for Determining a Presence State in a Vehicle The invention relates to a system for determining a presence state in a motor vehicle by means of a presence sensor, the system being configured to implement the following steps: - generate a point cloud (50) from data from the presence sensor, in particular a radar, - detect, in the point cloud generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone (60). Figure for the abstract: Fig. 4

Description

Title of the invention: System for determining a state of presence within a vehicle

[0001] The invention relates to a system for determining a state of presence within a motor vehicle.

[0002] It is known to detect the presence of a living being, particularly a person inside a motor vehicle, followed by an alert if a person is left inside the vehicle. Such detection and alert enhance passenger safety, particularly by preventing the danger of leaving a child inside the vehicle. Such detection can, for example, utilize radar located inside the motor vehicle.

[0003] In particular, there is a need to prevent false alarms. Such a false alarm can occur when a person outside the vehicle passes near the radar and the radar detects this person outside the vehicle. This detection can cause confusion if the person, although outside the vehicle, is considered to be inside the vehicle. This is because the radar inside the vehicle can pick up reflections of waves from the person outside the vehicle.

[0004] A false alarm can also occur when strong wind creates vibrations inside the vehicle.

[0005] Such false alarms are, of course, undesirable.

[0006] The invention aims in particular to prevent such false alarms.

[0007] The invention thus relates to a system for determining a state of presence within a motor vehicle by means of a presence sensor, the system being configured to implement the following steps: - generate a point cloud from data coming from the presence sensor, in particular a radar, - detect, within the point cloud generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone, - in a first embodiment of the invention: removing from the point cloud points that are associated with the external target and that are located within the predetermined detection zone in order to provide a filtered point cloud, then processing the filtered point cloud in a presence detection step in the vehicle, to detect a possible presence of a target inside the motor vehicle, or - in a second embodiment of the invention: in the event of proven detection, in the point cloud, of points which are associated with the external target and which are located within the predetermined detection zone, forcefully generate a non-detection data of presence within the vehicle, in particular even if there is someone in the vehicle.

[0008] Thanks to the invention, it is possible to avoid false alerts when the sensor, in particular a radar, captures points originating, for example, from a person outside the vehicle. The invention makes it possible to disregard points related to the person outside the vehicle, either by removing these points from the detected point cloud or by forcing a non-detection message regardless of whether or not there is a person in the vehicle.

[0009] In the first embodiment of the invention, the invention makes it possible to somehow obscure parasitic points generated by the presence of a person outside the vehicle.

[0010] In the second embodiment of the invention, the points which are due to the presence of the person outside the vehicle are not obscured and a non-presence information is forced to avoid a false alarm in the case where no person is present inside the vehicle.

[0011] In general, the invention makes it possible to prevent false alarms due to the spurious detection of a person or object moving in the immediate vicinity of the vehicle.

[0012] According to one aspect of the invention, the sensor detection area corresponds to the interior of the vehicle's passenger compartment.

[0013] According to one aspect of the invention, the detection zone is delimited by a volume, for example a parallelepiped, in particular a rectangular parallelepiped, corresponding substantially to the contour of the vehicle's passenger compartment.

[0014] According to one aspect of the invention, the presence sensor, which is in particular a radar, is thus configured to detect a presence throughout the entire interior of the vehicle, in particular to detect the potential presence of people in the front seats and in the rear seats.

[0015] According to one aspect of the invention, presence detection is carried out on the basis of points located by the presence sensor and which are inside this detection zone.

[0016] It should be noted that a person outside the vehicle passing near the vehicle is likely to generate the appearance of spurious points captured by the radar, in the detection zone, due to a reflection of radar waves off the person towards inside the vehicle. It is the presence of these extraneous points that can generate a false alarm.

[0017] According to one aspect of the invention, the sensor, in particular a radar, is configured to detect in the detection area, points whose behavior is representative of a presence, in particular of breathing, of a person.

[0018] Undesirably, the spurious points related to the presence and movement of a person near the car may be interpreted by the presence state determination system as points within the detection zone.

[0019] According to one aspect of the invention, the presence state determination system is configured to generate a grouping (or "Clustering" in English) of certain points in the point cloud and their tracking (or "Tracking" in English), in particular to identify spurious points.

[0020] In the first embodiment of the invention (namely, removing from the point cloud points associated with the external target and located within the predetermined detection zone in order to provide a filtered point cloud), extraneous points in the point cloud appearing within the detection zone are removed during point cloud processing. Extraneous points outside the detection zone may, if desired, not be processed, i.e., are not removed, since they do not affect the determination of a presence state within the detection zone.

[0021] Thus, if no person is present inside the vehicle, then the image obtained by such point cloud processing is devoid of points of interest, leading to the conclusion that no presence is detected inside the vehicle. Therefore, there is no reason to trigger an alert.

[0022] The invention is advantageous because, thanks to this treatment of stray points, it is not necessary to have additional sensors to detect the presence of an external object / person (external target), for example, a person passing outside the vehicle, near the vehicle. The invention thus makes it possible to use the radar normally present in the passenger compartment to perform the suppression of stray points.

[0023] Tracking generally refers to a process of continuously monitoring a moving object or target in order to determine its position and orientation in three-dimensional space. This makes it possible, in particular, to track the movements and changes in position of an object in real time.

[0024] In the situation, for example, where a child is inside the vehicle and an outside person (external target) passes near the vehicle, the sensor generates a point cloud and the system allows the identification of two groups of points, of which One group of points is associated with the child (because these points have similar movements), and another group of points is associated with the outsider (because these points have similar movements). In this situation, the system triggers an alert.

[0025] According to one aspect of the invention, the system is configured to use artificial intelligence to determine whether a group of points, in particular on the basis of point movements, are linked to a person (a target) inside the vehicle or outside the vehicle.

[0026] For example, movements of points with particular vectors can be associated by artificial intelligence with a person outside the vehicle.

[0027] For example, a few points in a group of points very close to the edge of the detection zone may be considered as belonging to an external person. In particular, the system is configured to recognize spurious points generated by the reflection of waves from the sensor off the external person, spurious points which exhibit a particular approach and retreat movement that is characteristic of a person walking alongside the vehicle.

[0028] It is also possible to deduce that points are spurious points if they remain present on the image captured by the sensor for a period of time which is relatively short, for example less than 20 seconds or less than 10 seconds, meaning that the target appears and disappears relatively quickly, and allowing us to ensure that this target is outside the vehicle.

[0029] According to one aspect of the invention, the system uses artificial intelligence capable of learning to improve the identification of spurious points.

[0030] According to one aspect of the invention, the system, in particular via artificial intelligence, is configured to compare a cloud of detected points, in particular its dynamics, to a stored catalogue to determine whether the cloud of points or a grouping of points from the cloud of points belong to a target in the vehicle or not.

[0031] An alert is issued in the event of detection of a presence in the vehicle, whether or not parasitic points are detected.

[0032] In the second embodiment of the invention, in the event of proven detection, in the point cloud, of points which are associated with an external target and which are located inside the predetermined detection zone, forcefully generate a non-detection data of presence inside the vehicle, even if a person's presence in the vehicle is detected simultaneously with the detection of spurious points which are associated with an external target.

[0033] According to one aspect of the invention, a new presence detection post-processing step is repeated periodically so that the forcing ci- above prevents a potential alert only for a period of time when the external target interferes with detection.

[0034] According to one aspect of the invention, the system is arranged to generate an alert signal based on the detected state of presence and the signal is in particular an audible sound signal in the environment of the vehicle, or a message sent to a smartphone or an emergency center.

[0035] According to one aspect of the invention, the system includes a computer configured to process signals from the presence sensor.

[0036] According to one aspect of the invention, the computer is configured to operate an artificial intelligence to process the point cloud obtained through the presence sensor.

[0037] The invention also relates to a method for determining a state of presence within a motor vehicle by means of a presence sensor, the method comprising the following steps: - generate a point cloud from data coming from the presence sensor, in particular a radar, - detect, within the point cloud generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone, - in a first embodiment of the invention: removing from the point cloud points that are associated with the external target and that are located within the predetermined detection zone in order to provide a filtered point cloud, then processing the filtered point cloud in a presence detection step in the vehicle, to detect a possible presence of a target inside the motor vehicle, or - in a second embodiment of the invention: in the event of proven detection, in the point cloud, of points which are associated with the external target and which are located inside the predetermined detection zone, forcefully generate a non-detection data of presence inside the vehicle (even if there is someone in the vehicle).

[0038] Other features, details and advantages of the invention will become clearer upon reading the following description on the one hand, and several illustrative and non-limiting examples of embodiments given with reference to the accompanying schematic drawings on the other hand, in which:

[0039] [Fig-1] Fig. 1 is a schematic view of a vehicle comprising the system of determination of a state of presence according to an example of an embodiment of the invention, with an external person passing near the vehicle;

[0040] [Fig.2] Fig.2 is a block diagram representing the steps in determining a presence by the system of the [Fig.l];

[0041] [Fig.3] Fig.3 is a schematic representation of a detection zone used by the system of [Fig.1];

[0042] [Fig.4] Fig.4 is a representation of three images of the detection area of the [Fig.3], at three successive moments in a situation of approaching an outside person;

[0043] [Fig. 5] Fig. 5 is a block diagram representing steps in determining a presence according to another example of an embodiment of the invention.

[0044] The features, variants, and different embodiments of the invention can be combined in various ways, provided they are not incompatible or mutually exclusive. In particular, variants of the invention may be conceived comprising only a selection of features, described hereafter in isolation from the other described features, if this selection of features is sufficient to confer a technical advantage and / or to differentiate the invention from the prior art.

[0045] Figure 1 shows a system 1 for determining the presence state E of an occupant within a vehicle V, here a motor vehicle, according to the invention. The system 1 comprises a presence sensor 10 and a computer 20 connected to the presence sensor 10.

[0046] The presence sensor 10 is configured to determine the presence of an occupant 30 within the vehicle V.

[0047] The presence sensor 10 is placed in a central area of ​​the vehicle V, here in the center of the roof 2 of the vehicle V as shown in [Fig.1], the presence sensor 10 can then advantageously cover all the seats of the vehicle V.

[0048] The presence sensor 10 can be any instrument enabling the determination of the presence of an occupant 30 within the vehicle V. The presence sensor 10 can consist of a single sensitive element or a network of sensitive elements distributed throughout the vehicle V.

[0049] The presence sensor 10 can, for example, be a camera or a camera network or even an ultrasonic sensor or an ultrasonic sensor network.

[0050] The presence sensor 10 is here a device for transmitting and receiving a radar wave.

[0051] The presence sensor 10 here comprises at least one antenna (not shown) designed to emit an electromagnetic wave 11 and at least one sensor (not shown) designed to receive a reflected electromagnetic wave 12 after reflection of the electromagnetic wave 11, and in particular after reflection on a person 30 inside vehicle V or an outside person 40 passing near vehicle V.

[0052] The presence sensor 10 is here more specifically a radar (from the common English acronym "radio detection and ranging") using millimeter electromagnetic waves. The radar is, for example, a Doppler radar using sustained or continuous waves.

[0053] The control unit 20 comprises at least one memory and at least one processor. The control unit 20 may be, for example, the electronic control unit (ECU) of vehicle V. The control unit 20 may also be a control unit dedicated to system 1. Instructions enabling the determination of the presence state E are stored in memory and implemented by the processor. When implemented, these instructions enable the execution of the process described later.

[0054] The computer 20 is programmed in particular to control the presence sensor 10, in particular to control the emission of the electromagnetic wave 11.

[0055] The determination system 1 is configured to implement the following steps: - generate a point cloud of 50 from data coming from the presence sensor 10 (step 101 in [Fig.2]), - detect, in the point cloud 50 generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone 60 (step 102 in [Fig.2]),

[0056] In a first embodiment of the invention, the determination system 1 is configured to remove, from the point cloud, points which are associated with the external target and which are located inside the predetermined detection zone in order to provide a filtered point cloud 51 (step 103 in [Fig.2]), then process the filtered point cloud 51 in a presence state detection step E in the vehicle V, to detect a possible presence of a target inside the motor vehicle (step 104 in [Fig.2]).

[0057] Steps 101, 102 and 103 can be considered as pre-treatment and step 104 as post-treatment.

[0058] In the event of the actual presence of a person 30 in the vehicle V, the system 1 sees, for the detection area, only the points related to this person 30 actually in the vehicle V because the extraneous points which are related to the external person 40 have been removed from the points used in step 104 of detection of a presence state E.

[0059] Thanks to the invention, it is possible to avoid false alerts when the sensor 10, here a radar, captures points originating from a person 40 outside the vehicle V. The invention makes it possible to disregard points related to the person outside the vehicle, either by removing these points or by simply deleting them. of the detected point cloud. In this first example of an embodiment of the invention, the invention makes it possible to somehow obscure spurious points generated by the presence of a person outside the vehicle.

[0060] In general, the invention makes it possible to prevent false alarms due to spurious detection of a person 40 or an object moving in the immediate vicinity of the vehicle V.

[0061] As can be seen in [Fig. 3], the detection zone 60 corresponds to the interior of the vehicle's passenger compartment V. Thus, the system 1 relies on this detection zone 60 and the presence or absence of points within it to detect a presence state E. It is understood that it is important that there are no extraneous points in this detection zone 60 that would disrupt step 104. The points in the point cloud can be located in an x, z coordinate system.

[0062] The detection zone 60 is delimited by a volume, for example a rectangular parallelepiped, corresponding substantially to the contour of the passenger compartment of the vehicle V.

[0063] The presence sensor 10 is thus configured to detect a presence throughout the passenger compartment of the vehicle V, in particular to detect the potential presence of people in the front seats and in the rear seats.

[0064] The presence detection E is carried out on the basis of points located by the presence sensor 10 and which are inside this detection zone 60.

[0065] As can be seen in [Fig.4] (which represents three images of the detection zone 60, at three successive times in a situation of approaching an external person), a person 40 outside the vehicle who passes near the vehicle V is likely to generate the appearance of spurious points 61 captured by the radar 10, in the detection zone 60, due to a reflection of radar waves on the person 40 towards the inside of the vehicle V. It is the presence of these spurious points 61 that can generate a false alert.

[0066] On the image on the left of [Fig.4], the point cloud 50 is generated by the person 40 outside the vehicle V, this person being initially quite far from the detection zone 60 so that the reflections do not generate spurious points in the detection zone 60.

[0067] Then person 40 approaches vehicle V and reflections of waves on person 40 generate the appearance of parasitic points 61 captured by radar 10, in the detection zone 60 (see image in the center of [Fig.4]).

[0068] After the removal of the spurious points 61 by the system 1, there remains a filtered point cloud 51 (on the right image of [Fig.4]) which does not affect the presence detection E.

[0069] The sensor 10 and the system 1 are configured to detect in the detection zone 60, points whose behavior is representative of a presence, in particular of breathing, of a person.

[0070] In a known manner, system 1 is configured to generate a grouping (or "Clustering" in English) of certain points in the point cloud and their tracking (or "Tracking" in English).

[0071] The spurious points 61 in the point cloud appearing inside the detection zone 60 are removed during point cloud processing. The spurious points outside the detection zone may, if desired, not be processed, i.e., are not removed, since they do not affect the determination of a presence state inside the detection zone.

[0072] Thus, if no person is present inside the vehicle, then the image obtained by such point cloud processing is devoid of points of interest, leading to the conclusion that no presence is detected inside the vehicle. Therefore, there is no reason to trigger an alert.

[0073] The grouping (or "Clustering" in English) is carried out using a Kalman filter. Other types of filters or means can of course be used.

[0074] In general, grouping points by their velocity vector allows us to conclude that a group of points belongs to the same object / person (a target).

[0075] For example, in the situation where a child is inside the vehicle and an outside person (external target) passes near the vehicle V, the sensor generates a point cloud and system 1 identifies two groups of points: one group of points associated with the child (because these points have similar movements) and another group of points associated with the outside person (because these points have similar movements). In this situation, the system triggers an alert.

[0076] System 1 is configured to use artificial intelligence to determine whether a group of points, in particular on the basis of point movements, are linked to a person (a target) inside the vehicle or outside the vehicle.

[0077] For example, movements of points with particular vectors can be associated by artificial intelligence with a person outside the vehicle.

[0078] For example, a few points in a group of points very close to the edge of the detection zone may be considered as belonging to an external person. In particular, the system is configured to recognize spurious points generated by the reflection of waves from the sensor 10 on the external person 40, spurious points which exhibit a particular approach and retreat movement that is characteristic of a person 40 walking next to the vehicle V.

[0079] It is also possible to deduce that points are spurious points if they remain present on the image captured by the sensor for a period of time which is relatively short, for example less than 20 seconds or less than 10 seconds, meaning that the target appears and disappears relatively quickly, and allowing us to ensure that this target is outside the vehicle.

[0080] System 1 uses artificial intelligence capable of learning to improve the identification of spurious points.

[0081] The system 1 can in particular be configured to compare a cloud of detected points, including its dynamics, to a stored catalogue to determine whether the point cloud or a grouping of points from the point cloud belong to a target in the vehicle or not.

[0082] An alert is issued in the event of detection of presence E in vehicle V, whether or not parasitic points are detected.

[0083] In a second embodiment of the invention illustrated in [Fig.5], in the event of confirmed detection (in a step 202), in the point cloud previously generated in a step 201, of points which are associated with an external target and which are located inside the predetermined detection zone 60, the system 1 is configured to forcibly generate a non-detection of presence data within the vehicle (even if there is someone in the vehicle), bypassing a presence detection post-processing step E (step 203).

[0084] By forcing the bypass of the presence detection post-processing step E, no alert is generated by system 1 (step 204), even if a person's presence in the vehicle is detected simultaneously with the detection of spurious points that are associated with an outside person 40.

[0085] If no parasitic point is detected, the presence detection post-processing step 203 E is implemented normally, and an alert can be generated if a presence is detected in the vehicle V.

[0086] A new presence detection post-processing step is repeated periodically so that the above forcing prevents a potential alert only over a period of time when the external target 40 interferes with detection.

[0087] In general, the system is arranged to generate an alert signal based on the detected presence state and the signal is in particular an audible sound signal in the environment of the vehicle, or a message sent to a smartphone or an emergency center.

Claims

Demands

1. A system (1) for determining a state of presence within a motor vehicle (V) by means of a presence sensor (10), the system being configured to implement the following steps: - generate a point cloud (50) from data from the presence sensor, which is a radar, - detect, in the point cloud generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone (60), - in a first embodiment of the invention: remove, from the point cloud, points that are associated with the external target (40) and that are located inside the predetermined detection zone (60) in order to provide a filtered point cloud (51), then process the filtered point cloud in a step of detecting a state of presence in the vehicle, to detect a possible presence of a target (40) inside the motor vehicle.or - in a second embodiment of the invention): in the event of confirmed detection, in the point cloud, of points which are associated with the external target (40) and which are located within the predetermined detection zone (60), forcibly generate a non-detection data point for the presence within the vehicle.

2. System (1) according to the preceding claim, wherein the detection area (60) of the sensor (10) corresponds to the interior of the vehicle's passenger compartment (V), being delimited in particular by a volume, for example a parallelepiped, in particular a rectangular parallelepiped, corresponding substantially to the contour of the vehicle's passenger compartment.

3. System (1) according to any one of the preceding claims, wherein the sensor (10) is configured to detect in the detection area (60), points whose behavior is representative of the presence, in particular of breathing, of a person.

4. System (1) according to any one of the preceding claims, wherein the system (1) for determining a presence state is configured to generate a grouping of certain points in the point cloud and their monitoring, in particular to identify extraneous points.

5. System (1) according to any one of the preceding claims, wherein the spurious points related to an external target (40) in the point cloud (50) appearing inside the detection area (60) are removed in a point cloud processing.

6. System (1) according to any one of the preceding claims, wherein the system is configured to use artificial intelligence to determine whether a group of points, in particular on the basis of point movements, are linked to a person (a target) inside the vehicle or outside the vehicle, for example, point movements with particular vectors can be associated by the artificial intelligence with a person outside the vehicle.

7. System (1) according to any one of the preceding claims, wherein the system is configured to recognize spurious points generated by the reflection of waves from the sensor on the outside person (40), spurious points which exhibit a particular approach and retreat movement which are characteristic of a person walking beside the vehicle.

8. System (1) according to any one of the preceding claims, wherein the system, in particular via artificial intelligence, is configured to compare a cloud of detected points (50), in particular its dynamics, to a stored catalogue to determine whether the point cloud or a grouping of points from the point cloud belong to a target in the vehicle or not.

9. System (1) according to any one of claims 1 to 4, wherein, in the event of proven detection, in the point cloud (50), of points which are associated with an external target and which are located within the predetermined detection zone (60), forcefully generate a non-detection data of presence within the vehicle, even if a person's presence in the vehicle (V) is detected simultaneously with the detection of spurious points which are associated with an external target.

10. System (1) according to the preceding claim, wherein the system (1) is configured such that a new presence detection post-processing step is periodically repeated so that the above forcing prevents a potential alert only over a period of time when the external target is interfering with detection.

11. System (1) according to any one of the preceding claims, wherein the system is arranged to generate an alert signal based on the detected presence state and the signal is in particular an audible sound signal in the environment of the vehicle, or a message sent to a smartphone or emergency center.

12. A method for determining a presence state within a motor vehicle (V) by means of a presence sensor (10), being a radar, the method comprising the following steps: - generating a point cloud (50) from data from the presence sensor, - detecting, in the point cloud generated in the previous step, any points that are associated with an external target located outside a predetermined detection zone (60), - in a first embodiment of the invention: removing, from the point cloud, points that are associated with the external target and that are located inside the predetermined detection zone (60) in order to provide a filtered point cloud, then processing the filtered point cloud in a step of detecting a presence state in the vehicle, to detect a possible presence of a target inside the motor vehicle,or - in a second embodiment of the invention: in the event of confirmed detection, in the point cloud, of points which are associated with the external target and which are located within the predetermined detection zone (60), to forcibly generate a non-detection data point for the presence inside the vehicle, in particular even if there is someone in the vehicle.

Images