GB2540663A

GB2540663A - Method for obtaining at least one biometric template for biomeric sensing

Info

Publication number: GB2540663A
Application number: GB1609736.2A
Authority: GB
Inventors: Hill Andrew; Burr John; Weiss Thomas
Original assignee: Daimler AG
Current assignee: Mercedes Benz Group AG
Priority date: 2016-06-03
Filing date: 2016-06-03
Publication date: 2017-01-25
Also published as: GB201609736D0

Abstract

A method for obtaining at least one biometric template for biometric sensing by a detection device, 22, in a vehicle (e.g. car), 10. The biometric template is capable of being used in a comparison with at least one biometric feature of a person so as to effect at least one function of the vehicle (e.g. locking, unlocking, operating the ignition). The method comprises detecting the biometric feature of the person via a detecting device, providing biometric training data and obtaining the biometric template on the basis of said training data. The biometric feature is detected by a plurality of sensors (24, 26, 28, 30), each being positioned apart from each other. The sensors may be in the interior of the vehicle. The sensors may be fingerprint sensors arranged on different components of the vehicle (e.g. steering wheel, handbrake, gearstick). The sensors may be configured for facial recognition, voice recognition or iris recognition.

Description

Method for Obtaining at least one Biometric Template for Biometric Sensing

The invention relates to a method for obtaining at least one biometric template for biometric sensing. US 2014/0359757 A1 shows a method of biometric authentication in which one or more icons are presented on a display. The method further comprises a step of receiving an indication that a user is interacting with at least one of the icons presented. In a further step, biometric information is acquired from a digit during the user interaction with the icon, when the digit is positioned in a fingerprinting sensing area. The method comprises a further step of invoking a function based, at least in part, on the acquired biometric information.

Moreover, DE 10 2011 054 449 A1 shows a procedure for access control for at least one user to an entry field. Furthermore, DE 199 38 064 A1 shows an identification device for access control to an object.

It has been shown that biometric sensing has been adopted in both mobile and consumer markets. Until recently the technology has mainly been showcased in spy movies and, more practically, in use for border control applications. The addition of fingerprint sensors in mobile phones for simply unlocking a mobile phone has brought a great awareness to its practicality to consumers. Unfortunately, initial enrollments are necessary which are usually very inconvenient and can require the dedication of a lot of time by the user in order to train sensors used in the context of biometric sensing to recognize just a single fingerprint. This is a very frustrating experience and decreases a user’s willingness to use a system that involves biometric sensing.

It is therefore an object of the present invention to provide a method by means of which enrollment for biometric sensing can be simplified or avoided.

This object is solved by a method having the features of patent claim 1. Advantageous embodiments with expedient developments of the invention are indicated in the other patent claims.

The invention relates to a method for obtaining at least one biometric template for biometric sensing by means of a detecting device of a vehicle such as, for example, a passenger vehicle. Said biometric template is capable of being used in a comparison of the biometric template and at least one biometric feature of a person so as to effect at least one function of the vehicle on the basis of the comparison. For example, in the context of biometric sensing, said biometric feature of a person being a user of the vehicle is detected by means of the detecting device. On the basis of the detected biometric feature of the person, biometric actual data are provided by the detecting device, said biometric actual data being indicative of the actually detected biometric feature. This means the biometric actual data represent the actually detected biometric feature of the person. In said comparison, the actually detected biometric feature or the provided biometric actual data are compared with the biometric template by, for example, the detecting device or an electronic control unit of the detecting device. If the biometric actual data match the template, said at least one function of the vehicle will be effected by the detecting device. If, for example, the biometric actual data do not match the biometric template, the at least one function will not be effected. Thus, basically, a particularly comfortable operation of the vehicle can be realized by means of biometric sensing.

In order to realize a particularly easy enrollment for realizing biometric sensing or to avoid enrollment for biometric sensing, the method according to the present invention comprises a first step in which the biometric feature of the person is detected by means of the detecting device. The method according to the present invention further comprises a second step in which biometric training data are provided by means of the detecting device on the basis of the detected biometric feature. Moreover, the method according to the present invention comprises a third step in which a said biometric template is obtained by, for example, the detecting device on the basis of the biometric training data, wherein the biometric feature is detected by means of a plurality of sensors of the detecting device, said sensors being arranged at a distance from each other. In comparison with conventional methods using biometric sensing, the method according to the present invention can simplify or avoid enrollment by learning the person’s biometric or biometric feature slowly over time. This allows the invention to provide a truly delightful experience with biometric sensing and biometric systems using biometric sensing.

The idea behind the present invention is that biometric sensing such as fingerprint sensing has recently become reality in both mobile and consumer markets. The addition of fingerprint sensors in mobile terminals such as mobile phones has led to greater acceptance of fingerprint sensing and a change in perspective from it representing tracking of criminals to a method for personal security and convenience in consumer electronics.

In the context of biometric sensing such as fingerprint sensing there are a variety of ways to match fingerprints - specifically either match-on-chip or match-on-host. In the match-on-chip scenario, a fingerprint being a biometric feature of a person is detected and matched to the person all within the same chip, so that no personal information is transmitted off the chip. In contrast to this, some systems function based on match-on-host system, where the information from the sensor is sent to a host processor perform the final match. Regardless of the system used, there is a concept of templates that describe the fingerprint by features rather than the entire image. Both systems have the capability to export this profile from the chip, and to load a profile from offboard systems to add or refine existing profiles.

Usually, when a user first starts to use a biometric identification device, said device must first gather enough information to be able to correlate the user with images that it sees such as fingerprint attributes or facial features. It is not sufficient to simply take a single image, as variations will necessarily exist between readings. These variations could be due to environmental effects such as different amounts of pressure on fingerprint sensors, background noise for voice identification or minor changes in lighting or position for facial recognition, or with irregularities in the relative positioning of the user and the recognition device. For this reason, usually, enrollment consists of taking a plurality of images before the first use. Usually, this must be repeated for each finger so enrollment can be a lengthy procedure.

Thus, usually, enrollment is a form of iterative training. Even after a successful enrollment, biometric identification methods remain sensitive to environmental variation. For example, fingerprint sensors can fail in the presence of moisture or when hand lotion is used. Likewise, facial recognition is sensitive to ambient light, makeup and facial hair. A match is not made with absolute certainty; rather a threshold must be passed for it to be a correct identification. Flowever, if a match is made, the additional information form the new image can be used to improve the model to make future matches more robust to these environmental variations. It is this feature that the method according to the present invention exploits to implement zero enrollment or simplifying enrollment.

Zero enrollment functions by passively receiving training data from biometric sensors without explicit user interaction, said training data being said biometric training data. This can be achieved by placing fingerprint sensors and/or other sensors of the detecting device in convenient locations to harvest data, in particular biometric training data. These locations are not necessarily locations that are convenient locations to interact with for biometric authentication, merely somewhere the user or person will touch frequently in order to detect at least one fingerprint which is a said biometric feature of the person. In other words, the biometric sensors are placed in order to maximize touches by a user during normal and generally natural use of the vehicle and vehicle components, as opposed to placing sensors in an area that is only specific to the activity of biometric authentication. Thus, said sensors are used as harvest biometric sensors which, for example, build a user profile over time. That is, for example, when pressed, the respective sensor first looks for a match, and if one is found, it updated its template to improve its idea of the user. This is the interaction that is happening when one goes through the lengthy and conventional enrollment process of today. However, in the method according to the present invention, instead of taking several images all in one session, the method according to the present invention examines these images over time as, for example, the fingers of the person are serendipitously placed on the sensors, until eventually the detecting device has built a complete model of the user or the at least one biometric feature.

At the point where a complete model of the user or the at least one biometric feature exists, at least one of said sensors can then be used for authentication. This means that, for example, after the complete model of the biometric feature was generated, at least one of said sensors or the sensors of the detecting device can be used to detect the biometric feature whose model had been generated before. On the basis of the actually detected biometric feature, the respective sensor and, thus, the detecting device can provide biometric actual data indicative of the actually detected biometric feature. Then, the biometric actual data are compared with the generated model, i.e. the generated biometric template. If, for example, the biometric actual data at least partially match the template, the at least one function is effected by means of the detecting device and performed by the vehicle.

However, more importantly, information from one sensor can be used to train other sensors that otherwise might not receive such serendipitous touches. For instance, a steering wheel of the vehicle receives a large number of touches and can harvest a large amount of fingerprint data. However, a start/stop button of the vehicle will receive only two touches per trip. Therefore, the template generated from touches on the fingerprint sensors on the steering wheel can be transferred to fingerprint sensors on the start/stop button via, for example, CAN (controller area network) or another in-car interface, allowing touch recognition without any enrollment on that sensor. This technique described with respect to fingerprints and fingerprint sensors can be applied to sensors other than fingerprint sensors, such as facial, iris or voice recognition sensors.

The method according to the present invention is particularly advantageous in fleet and shared car situations where the cost of enrollment is highest. In contrast to a time-consuming and tedious enrollment process, biometric information is learned over time in the method according to the present invention, using multiple sensors to train each other. These sensors are placed strategically so that they harvest biometric information without requiring explicit input from the user so enrollment can be avoided or simplified.

Further advantages, features, and details of the invention derive from the following description of a preferred embodiment as well as from the drawing. The features and feature combinations previously mentioned in the description as well as the features and feature combinations mentioned in the following description of the figure and/or shown in the figure alone can be employed not only in the respectively indicated combination but also in other combinations or taken alone without leaving the scope of the invention.

The drawing shows in the only Fig. part of a schematic and sectional side view of an interior of a vehicle configured to perform biometric sensing on the basis of at least one biometric template which is obtained by means of a plurality of sensors being arranged at a distance from each other.

The only Fig. shows a schematic and sectional side view of an interior 10 of a vehicle which is configured as, for example, a passenger vehicle. As will be described in greater detail in the following, the vehicle is configured to perform biometric sensing by means of which at least one function of the vehicle can be effected by a user of the vehicle in a particularly comfortable way. In the context of biometric sensing, at least one biometric feature of said user being a person is detected. If the detected biometric feature of the user matches at least one biometric template, said at least one function is effected so that the vehicle performs said function.

The vehicle comprises a plurality of components which are arranged in the interior 10. A first one of said components is a center console 12 which is arranged between front seats in the transverse direction of the vehicle, wherein said front seats are not shown in the Fig. A second one of said components is a first control element 14 being, for example, a knob which can be rotated and/or translationally moved in relation to the center console 12 in order to operate the vehicle. A third one of said components is a dash board 16 on which a second control element 18 is arranged. For example, said control element 18 is a start/stop button by means of which a drive motor such as an internal combustion engine configured to drive the vehicle can be activated and deactivated. The control element 18 is a fourth one of said components. A fifth one of said components is a steering wheel 20 by means of which the driver of the vehicle can steer the vehicle by turning the steering wheel 20. The vehicle further comprises a detecting device 22 by means of which said biometric feature can be detected so that said biometric sensing is performed by means of the detecting device 22. Moreover, said biometric template is generated by means of the detecting device 22.

For example, after generating and storing the biometric template in a memory of the detecting device 22, in particular in a memory of an electronic control unit of the detecting device 22, the at least one biometric feature of the user can be detected by means of the detecting device 22. Biometric actual data are provided by the detecting device 22 on the basis of the actually detected biometric feature, wherein the provided biometric actual data are indicative of the actually detected biometric feature. The actually detected biometric feature, i.e. the biometric actual data, is or are compared with the stored template. If the actually detected biometric feature, i.e. the biometric actual data, matches or match the stored template, the at least one function will be effected by means of the detecting device 22, in particular by the electronic control unit so that the vehicle will perform said at least one function.

Usually, a time-consuming and tedious enrollment is necessary to obtain and, thus, generate the said biometric template. However, in the present case, the vehicle, in particular the detecting device 22, is configured in such a way that the enrollment can be simplified or avoided. The detecting device 22 and, thus, the vehicle overall are configured to perform a method for obtaining the biometric template. Said method comprises a first step of, by means of the detecting device 22, detecting the biometric feature of the person being in the interior 10. The method further comprises a second step of, by means of the detecting device 22, providing biometric training data on the basis of the detected biometric feature. The method further comprises a third step of obtaining the biometric template on the basis of the biometric training data, wherein the biometric feature is detected by means of a plurality of sensors 24, 26, 28 and 30 of the detecting device, wherein the sensors 24, 26, 28 and 30 are arranged at a distance from each other in the interior 10. Thus, in the method, the biometric training data and, thus, the biometric template are learned over time using multiple sensors 24, 26, 28 and 30 to train each other. The sensors 24, 26, 28 and 30 are placed strategically so that they harvest biometric information, i.e. the biometric training data, without requiring explicit input from the person.

For example, the biometric feature to be represented by the biometric template is a fingerprint of said person being a user of the vehicle. Thus, the respective sensors 24, 26, 28 and 30 are fingerprint sensors configured to detect the said fingerprint of the person. The control elements 14 and 18 on the steering wheel 20 are components which are touched several times by the person over time. Every time the person touches the control elements 14 and 18 or the steering wheel 20 in such a way that a finger of the person touches the respective sensor 24, 26, 28 or 30, the fingerprint of that finger is detected and biometric training data on the basis of which the biometric template can be generated are provided by the detecting device 22. After obtaining, generating and storing the biometric template, the respective sensors 24, 26, 28 and 30 can be used to detect the finger and, thus, the fingerprint so as to provide biometric actual data indicative of the actually detected fingerprint. The provided biometric actual data are compared to the biometric template that had been obtained, generated and stored before. If, as explained above, the biometric actual data match the stored template, the at least one function will be effected and, thus, performed by the vehicle so that the person can effect the said function by means of the fingerprint. Thus, by means of the said method a time- consuming and tedious enrollment can be avoided. Hence, the said biometric sensing can be performed as a zero enrollment biometric sensing.

List of reference signs 10 interior 12 center console 14 control element 16 dash board 18 control element 20 steering wheel 22 detecting device 24 sensor 26 sensor 28 sensor 30 sensor

Claims

1. A method for obtaining at least one biometric template for biometric sensing by means of a detecting device (22) of a vehicle, the biometric template being capable of being used in a comparison of the biometric template and least one biometric feature of a person so as to effect at least one function of the vehicle on the basis of the comparison, the method comprising: - by means of the detecting device (22): detecting the biometric feature of the person; - by means of the detecting device (22): providing biometric training data on the basis of the detected biometric feature; and - obtaining the biometric template on the basis of the biometric training data, wherein the biometric feature is detected by means of a plurality of sensors (24, 26, 28, 30) of the detecting device (22), the sensors (24, 26, 28, 30) being arranged at a distance from each other.

2. The method according to claim 1, wherein the sensors (24, 26, 28, 30) are arranged in the interior (10) of the vehicle.

3. The method according to claim 1 or 2, wherein the template is used for a plurality of sensors (24, 26, 28, 30) of the detecting device (22) so as to effect at least one function or different functions of the vehicle by means of the sensors (24, 26, 28, 30) on the basis of the same template.

4. The method according to any one of the preceding claims, wherein the sensors (24, 26, 28, 30) are configured as fingerprint sensors arranged at different locations and/or on different components in the interior (10) of the vehicle.

5. The method according to any one of the preceding claims, wherein the sensors (24, 26, 28, 30) are configured for facial and/or iris and/or voice recognition.