JP2023529560A - Calibration of behavioral biometrics in aerosol-generating devices - Google Patents

Abstract

The present invention provides a technical solution that can establish the identity of authorized users of aerosol generating devices. This technical solution advantageously guarantees that the aerosol-generating device can be fully calibrated and authenticated to the user on the device, ensuring that security is not compromised while also considering the technical limitations inherent in the aerosol-generating device. Behavioral biometrics are used to ensure that any restrictions placed on the user of the device cannot be easily breached or circumvented. Measured behavioral biometrics can include movement of the device, position of the device, pressure applied to the device, time of use of the device, how the device is held, or use of one or more switches of the device. Features extracted from biometrics can be aggregated to create datasets that can be combined to form user profiles.

Description

The present disclosure relates to aerosol-generating devices configured to heat an aerosol-generating substrate to generate an aerosol. Such devices can heat or vaporize tobacco or other suitable aerosol-generating substrate materials by conduction, convection, and/or radiation rather than burning them to generate an aerosol for inhalation. The present invention relates generally to controlling access to aerosol-generating devices, and more particularly to calibrating user behavioral biometrics in aerosol-generating devices.

The popularity and use of risk-reducing or risk-modifying devices (also known as vaporizers) assist habitual smokers who wish to quit smoking traditional tobacco products such as cigarettes, cigars, cigarillos, and cigarettes. It has grown rapidly in the last few years as an aid to Various devices and systems are available for heating or warming an aerosolizable substance, as opposed to burning tobacco in conventional tobacco products.

Commonly available risk reduction or risk modification devices are substrate-heated aerosol-generating devices or heat-not-burn devices. However, such devices should be restricted to specific users for safety and regulatory reasons.

Accordingly, there is a need for techniques particularly suited to aerosol generating devices for establishing the identity of authorized users. In addition, calibration techniques are also required for effective operation.

For data protection purposes, it is desirable that such aerosol-generating devices not communicate with remote devices, and that authentication and calibration occur entirely at the aerosol-generating device. Therefore, there is a need for suitable certification and calibration methods that take into account the form factor and technical limitations inherent in aerosol generating devices, such as small size, limited processing power, and limited number of control elements.

According to a first aspect, the present disclosure is a method for calibrating user identification of an aerosol-generating device comprising one or more sensors and a memory, comprising: initiating a calibration mode of the device; measuring one or more behavioral biometrics of the user with one or more sensors; calculating a user profile based on the measured one or more behavioral biometrics of the user; A method is provided that includes storing in memory and exiting a calibration mode of the device.

In general, the present invention provides a technical solution in which the aerosol-generating device has means enabling calibration and identification of behavioral biometrics of the user of the device. This technical solution advantageously ensures that unauthorized users cannot use the aerosol-generating device, while also allowing any method for user calibration or identification to be done entirely on the aerosol-generating device. do. Furthermore, the use of behavioral biometrics in the present invention ensures that any restrictions placed on the user of the device cannot be easily breached or circumvented.

Using more than one behavioral biometric of the user to create a user profile improves the accuracy of the user profile, but increases the number of sensors required, thus increasing product cost. Simultaneously measuring multiple behavioral biometrics can improve the success rate.

By integrating the full calibration hardware, such as sensors and memory, into the device, calibration can be safely done only on the device.

At a high level, in a first embodiment, the invention includes a method for calibrating user identification of an aerosol generating device comprising one or more sensors and a memory.

The method includes initiating a calibration mode of the device; measuring one or more behavioral biometrics of the user with one or more sensors; measuring one or more behavioral biometrics of the user; calculating a user profile based on the metrics; storing the user profile in memory; and exiting calibration mode of the device.

Optionally, calculating a user profile based on one or more measured behavioral biometrics of the user comprises extracting one or more specific features from each behavioral biometric; aggregating one or more particular features from the behavioral biometrics to create respective behavioral biometric datasets; combining each behavioral biometric dataset to create a user profile; including.

This has the advantage that a large amount of measurement data from multiple sensors and/or behavioral biometrics can be reduced to a much smaller set while still preserving the characteristics of the user contained within the data. This greatly reduces the amount of data that needs to be stored in the device's memory.

Optionally, extracting one or more specific features comprises calculating statistical features.

This has the advantage that the statistical features are simple to compute and the processing requirements of the device are reduced.

Optionally, measuring the one or more behavioral biometrics includes movement of the device, absolute localization of the device to determine the position of the device in space, pressure applied to the device, timing of using the device, Including measuring one or more of the holding method of the device or the use of one or more switches of the device.

By measuring multiple behavioral biometrics simultaneously, the success rate of calibration or identification can be improved. Further, the interaction between multiple behaviors provides a combined behavior identification. This can be beneficial as it is specific to individual users and not just one behavior.

Optionally, the device further comprises one or more outputs indicating to the user the end of measuring one or more behavioral biometrics. This allows the user to know when the measurement is finished.

If the user has not performed the requested behavior during calibration, the output can indicate to the user that the measurement is not finished yet and calibration should continue. This prevents the user from ending calibration without determining a user profile.

Optionally, exiting calibration mode returns the aerosol generating device to normal operating conditions. Optionally, the normal operating state of the device is indicated to the user by one or more outputs of the device.

Optionally, the first set of sensors can be used to measure multiple behavioral biometrics. By measuring multiple behavioral biometrics simultaneously, the success rate of calibration and subsequent user identification can be improved.

Optionally, the method comprises initiating an identification mode of the device and determining, by the one or more sensors, one or more behavioral biometrics of the user based on a calibrated user profile stored on the device. measuring the metric; calculating a user profile based on one or more measured behavioral biometrics of the user; and comparing the calculated user profile to user profiles stored in memory. and further including.

Optionally, the method further comprises determining whether the calculated user profile and the user profile stored in memory match. Unlock the device if the profiles match, keep the device locked if the profiles do not match.

Optionally, the method further comprises exiting an identification mode of the device.

According to a second aspect, the present disclosure provides one or more sensors for calibrating and identifying a user of the device, a microprocessor for controlling the one or more sensors, and calibrating the device. and a software controlled device management system for performing calibration and subsequent identification of users of the device.

Optionally, the aerosol generating device of the second embodiment is configured to perform the method of any one of the first embodiment and optional functions thereof.

According to a third aspect, the present disclosure provides a computer-readable medium containing instructions that, when executed by a computer, cause the computer to perform the method of the first embodiment and optional functions thereof.

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings.

1 shows a schematic block diagram of an aerosol generating device; FIG. FIG. 4 is a flow diagram illustrating a calibration method performed by an aerosol generating device according to one embodiment; FIG. 4 is a flow diagram illustrating a calibration method performed by an aerosol generating device according to one embodiment; FIG. 4 is a flow diagram illustrating an identification method performed by an aerosol-generating device using stored profiles;

FIG. 1 shows a schematic block diagram of an aerosol-generating device 100 suitable for implementing the method of the invention. Aerosol generation device 100 includes sensor 102 , microprocessor 106 , memory 104 and software controlled device management system 108 . Sensor 102 is communicatively coupled to memory 104 , microprocessor 106 and software controlled device management system 108 .

Sensor 102 is configured to calibrate and identify a user of aerosol-generating device 100 . By way of example, sensors 102 may take the form of acceleration sensors, gravity sensors, motion sensors, pressure sensors, momentary switches, touch screen sensors, timing devices, light sensitive switches, and/or impedance sensors, among others.

Microprocessor 106 is configured to control sensor 102 . Microprocessor 106 may also be configured to execute control software to control various functions within device 100 . This may include implementing and controlling sensors 102, output elements, and/or input elements.

Memory 104 is configured to store one or more user profiles created by calibrating aerosol-generating device 100 .

Software controlled device management system 108 is configured to perform calibration and subsequent identification of a user of aerosol generating device 100 .

FIG. 2 illustrates a method for calibrating user identification of an aerosol-generating device 100 comprising one or more sensors 102 and memory 104 . The method of FIG. 2 can be performed by the aerosol generation device 100 of FIG.

At step 202 , the user of aerosol-generating device 100 initiates calibration mode of device 100 . Initiating the calibration mode interrupts the normal operating state of the device 100, stops the aerosol-generating operation of the device 100, and prevents the user from using the device 100 in the normal manner. The calibration mode of device 100 occurs entirely on device 100 . By integrating complete calibration hardware into device 100 , calibration can be safely performed on device 100 . A user may initiate a calibration mode of device 100 by engaging one or more input elements of device 100 . These input elements may be one or more buttons, one or more switches, one or more embedded safety buttons, one or more motion sensors, or similar inputs. The user can initiate the calibration mode, for example, by pressing and holding the on/off button of device 100 or repeatedly pressing the on/off button of device 100 several times. Alternatively, a dedicated calibration button can be provided on device 100 . This could be a small button that is operated using a tool like the tip of a ballpoint pen, or it could be a button that is not easy to actuate unintentionally, such as a recessed safety button. Alternatively, the motion sensor can detect motion along with the push button. Calibration mode can be initiated by first pressing a button, which initiates the measurement process described below. When calibration mode is initiated by a user, device 100 can use one or more output elements to indicate to the user that calibration mode is active. These output elements can communicate to the user using light, vibration, or sound.

At step 204 , one or more sensors 102 of aerosol-generating device 100 measure one or more behavioral biometrics of the user of device 100 . Behavioral biometrics are defined as what the user's body "does." This is different from physiological biometrics, such as face or fingerprint recognition, which are defined as "what" the user's body is.

Device 100 is provided with a program stored in memory 104 of device 100 that performs steps 204, 206, 208, and 210 to perform calibration. The program is launched at step 202 by starting the calibration mode of device 100 . During the measurement of one or more behavioral biometrics, device 100 can indicate to the user that this is occurring through one or more output elements of device 100 . One or more of the output elements can indicate the status and/or progress of the measurement cycle, for example by flashing lights or increasing the number of lights and turning lights on. A blinking light can change blink rate or change color over time. One or more output elements can also indicate to the user the end of the measurement.

One or more sensors 102 of device 100 can measure one or more behavioral biometrics of a user of device 100 . By measuring multiple behavioral biometrics simultaneously, the success rate of calibration and subsequent user identification can be improved. Behavioral biometrics that are measured are only those that can be measured by one or more sensors 102 on the aerosol-generating device 100 . It does not rely on ancillary sensors, devices or networks. The one or more sensors 102 of the device 100 that can detect one or more behavioral biometrics of the user may be acceleration, gravity, and motion sensors for detecting movement of the device 100 . They can both measure large movements of the device 100 or be used to detect small movements and impacts such as tapping. Gravity sensors can also measure absolute localization measurements to determine the position of device 100 in space. One or more sensors 102 may be pressure sensors. A pressure sensor can measure pressure applied to the sensor by a user's hand and/or fingers. The one or more sensors 102 may also be used in conjunction with other sensors such as momentary switches that can measure actuation, touch screen sensors that measure position, movement and/or pressure within the touch area, motion sensors or switches. a timing device, a light sensitive switch that detects when a particular area of the device 100 is covered by the user, and/or an impedance sensor that determines how the device 100 is being gripped by the user. good.

These one or more of the sensors 102 described above enable the device 100 to measure one or more behavioral biometrics of the user, such as when in the user's hands, among other measurable behavioral biometrics. Measure movement of the device 100, button pressure by a user's finger when pressing a button, grip pressure on the device when handling the device 100, and/or tapping in a fixed pattern on a touch screen display or on the device 100 itself, etc. can do.

In some scenarios, the same sensor or sensors 102, eg, motion sensors, can be used to detect multiple behavioral biometrics. This requires processing the measurements to isolate individual behavioral biometrics. This can be, for example, the entire user's hand gesture combined with the tap. Smaller tapping impulses can be separated from larger motions of the whole device and two measurements are generated from the same data set. This can reduce the number of sensors 102 used, but comes at a cost in terms of data processing power.

An example of a combination of behavioral biometrics is the detection of a hand movement pressing a button. This method uses detection of hand movements in combination with button presses on the device 100 . For hand motion, a motion sensor detects motion of the device 100 and a button detector measures the time between button actuations. Alternatively or additionally, force applied over time can also be measured. Another example of a combination of behavioral biometrics is the combination of button presses and force sensors. In this example, the number and timing of button presses are determined along with the force sensor during the press. A force sensor can measure the pressure applied to the button and/or device 100 .

Interactions between multiple behaviors are beneficial. In the example above, pressing the button affects both the applied pressure and the movement of the device 100 . The interaction between both behaviors provides a combined behavior identification. This can be more beneficial than testing only individual data streams as it is specific to individual users.

Measurements from one or more sensors 102 can be taken at the same time or immediately after each other so that individual behaviors affect each other. Alternatively, measurements can be alternated to conserve battery, but still switch quickly (in the range of 1-10 microseconds) to capture the synchronicity of detected behavior.

At step 206, a user profile is calculated based on one or more measured behavioral biometrics of the user. Device 100 may be provided with a program and one or more processors to complete the calculations.

Calculating a user profile based on one or more measured behavioral biometrics of the user includes extracting one or more specific features from each behavioral biometric, then extracting one or more specific features from each behavioral biometric, It may involve aggregating one or more specific characteristics. Each behavioral biometric data set can be created by aggregating one or more specific features. Each biometric data set can then be combined to create a user profile based on one or more measured behavioral biometrics of the user.

Extraction of one or more particular features may include computation of statistical features, normalization, data filtering, denoising, and/or high-pass/low-pass filtering followed by processing to extract particular features. can.

Feature extraction is a method of reducing large amounts of measurement data into a much smaller set while still preserving the user characteristics contained within the data. This greatly reduces the amount of data that needs to be stored as user identification data. Using statistical features is advantageous because they are simple to compute and reduce processing requirements. For example, from the particular measurement data set available, statistical features can include the mean, standard deviation, kurtosis, and skewness of each data stream. Advantageously, the data features are stored in an N×M feature matrix, where N is the number of features extracted from the data and M is the number of data streams received from the measurements. In general, the more features extracted, the higher the accuracy achieved. In practice, a maximum of 4 features is sufficient. For example, a typical motion sensor produces measurements in four data streams, typically a continuous data stream of x, y, z measurements as well as magnitude values. This results in a 4×4 feature matrix (N=4, M=4). If we omit the magnitude values, this becomes a 4x3 (N=4, M=3) matrix. If more features are desired, these can include, for example, x, y, z, etc. (maximum, average) velocities.

For behavioral biometric measurements that do not generate a continuous data stream, such as button presses (on/off), care must be taken to define an appropriate data stream to determine the features to be extracted later. For example, time between presses, time from on to off (for momentary switches), pressure level over time, touch coordinates over time (for touch screens), etc. can be used.

Whatever feature extraction is used for each behavioral biometric, the results can be normalized so that individual modes yield comparable results. In the previous example, the mean value should be normalized to the unit value of 1, for example.

One or more specific features extracted from each behavioral biometric are aggregated to create a dataset. Aggregation of data combines aspects of behavior and captures interactions between them, making them more user-specific and therefore much more difficult to imitate by unauthorized users. Data aggregation can be performed, for example, by matrix multiplication of feature matrices derived from the first behavioral biometric and the second behavioral biometric. Other data aggregation methods are possible.

Alternatively, the procedure can be performed using a single behavioral biometric, but this does not have the advantage of increasing the accuracy of multimodal calibration.

At step 208 , the calculated user profile is stored in memory 104 of device 100 . User profiles are stored for later use during user identification. A user profile can be stored as a variation of the standard profile by storing only delta values. This reduces memory requirements for storing user profiles.

The calculated user profile can be compared to a standard profile of measured behavioral biometrics. This standard profile may have been determined based on a range of different user profiles. A computed user profile can be tested to be within the limits of a standard user profile to determine whether the behavior requested for calibration has actually been performed. If the user did not perform the requested behavior during calibration (e.g., did not move the device 100 where the movement was required, or performed a movement that was very different from what was requested), the calibration Results are not stored and calibration ends without determining a user profile. This prevents the user from storing inappropriate and non-reproducible movements and thus increasing the risk of being unable to unlock or being inadequately protected (e.g. calibration without movement (static device 100 ), which results in easily hackable behavior). This option can be used with one or more behavioral biometrics. For example, the maximum percentage deviation between the standard and the actually measured features, or maximum and minimum thresholds for each feature can be used.

At step 210, the calibration mode of device 100 ends. After calibration is complete, device 100 exits calibration mode and returns to neutral, off, or normal operating state. This change may be indicated by one or more output elements of device 100 .

FIG. 3 is a flow diagram illustrating a calibration method performed by the aerosol-generating device 100 according to one embodiment. Especially in situations where two behavioral biometrics are measured and used to calculate the user profile. Measurements of two behavioral biometrics of the user can be made by one or more sensors 102 . In situations where the first behavioral biometric is measured by more than one of the sensors 102, one or more specific features are extracted from the data of each of the two or more sensors 102 and then all Aggregated when certain features are aggregated.

At step 302 , the user of aerosol-generating device 100 initiates calibration mode of device 100 . Initiating the calibration mode interrupts the normal operating state of the device 100, stops the aerosol-generating operation of the device 100, and prevents the user from using the device 100 in the normal manner. The calibration mode of the device takes place entirely on device 100 .

At step 304 , one or more sensors 102 of aerosol-generating device 100 measure first behavioral biometrics of the user of device 100 .

At step 306, one or more specific features are extracted from the first behavioral biometrics. Extraction of one or more particular features may include computation of statistical features, normalization, data filtering, denoising, and/or high-pass/low-pass filtering followed by processing to extract particular features. can.

At step 308, one or more specific characteristics are aggregated to create a first data set of first behavioral biometrics.

Steps 310 , 312 and 314 may be performed concurrently with steps 304 , 306 and 308 . Alternatively, they may be performed before or after steps 304 , 306 , 308 .

At step 316, the first and second datasets of behavioral biometrics are combined to create a user profile. This user profile contains data related to two behavioral biometrics. However, the same process can be used for more than two behavioral biometrics.

At step 318 , the calculated user profile is stored in memory 104 of device 100 . User profiles are stored for later use during user identification.

At step 320, the calibration mode of device 100 ends. After calibration is complete, device 100 exits calibration mode and returns to neutral, off, or normal operating state.

The methods of FIGS. 2 and 3 demonstrate that one or more behavioral biometrics can be used to compute a user profile. FIG. 4 shows how the calculated user profile is used to identify the user of device 100 .

FIG. 4 is a flow diagram illustrating an identification method performed by aerosol-generating device 100 using stored profiles.

At step 402, the identification mode of device 100 is initiated. This may be initiated by the user activating a switch, by the user picking up or moving the device 100 , or by the user holding the device 100 , for example.

At step 404 , one or more behavioral biometrics of the user are measured by one or more sensors 102 based on the calibrated user profile stored in memory 104 of device 100 . Only behavioral biometrics used in calibrating the user and creating a stored calibrated user profile are measured by one or more sensors 102 . This is so that the current user's behavioral biometrics can be directly compared with the behavioral biometrics of the stored user profile.

Measurements from one or more sensors 102 can be taken at the same time or immediately after each other so that individual behaviors affect each other. Alternatively, measurements can be alternated to conserve battery, but still switch quickly (in the range of 1-10 microseconds) to capture the synchronicity of detected behavior. Measurements can be taken at specified intervals or, alternatively, until sufficient data is measured.

At step 406 , a user profile for the current user is calculated based on one or more measured behavioral biometrics of the user of device 100 . This calculation is done in the same manner as the user profile calculations stored during calibration mode. This is so that the current user's user profile can be directly compared to the stored user profiles.

At step 408 , the calculated used profile is compared with the user profile stored in memory 104 of device 100 .

At step 409, it is determined whether the calculated user profile and the user profile stored in memory 104 of device 100 match.

If the profiles match, the method proceeds to step 410 and device 100 is unlocked for use by the user.

If the profiles do not match, the method proceeds to step 412 where device 100 remains locked and cannot be used by unauthorized users.

At step 414, the identification mode of device 100 ends after the device is unlocked or remains locked.

It will be apparent to those skilled in the art that all of the methods described herein are suitable for implementation by the aerosol generating device 100.

Having described aspects of the present disclosure in detail, it will be apparent that modifications and variations are possible without departing from the scope of aspects of the present disclosure as defined in the appended claims. Since various changes may be made in the above structures, products and methods without departing from the scope of aspects of this disclosure, all matter contained in the foregoing description and shown in the accompanying drawings may be changed from , are intended to be construed as illustrative and not in a limiting sense.

Although the disclosure has been described in terms of various specific embodiments, those skilled in the art will recognize that the disclosure can be practiced with modification within the scope of the claims.

Claims (15)

A method for calibrating user identification of an aerosol generation device comprising one or more sensors and a memory, comprising:
initiating a calibration mode of the device;
measuring one or more behavioral biometrics of the user with the one or more sensors;
calculating a user profile based on the measured one or more behavioral biometrics of the user;
storing the user profile in the memory;
exiting the calibration mode of the device;
A method, including calculating a user profile based on the measured one or more behavioral biometrics of the user;
extracting one or more specific features from each behavioral biometric;
aggregating the one or more specific characteristics from each behavioral biometric to create a dataset for each behavioral biometric;
Combining the datasets of respective behavioral biometrics to create a user profile;
2. The method of claim 1, comprising: 3. The method of claim 2, wherein extracting one or more particular features comprises computing statistical features. measuring the one or more behavioral biometrics;
movement of the device;
absolute localization of the device to determine the position of the device in space;
pressure on the device;
when to use said device;
a method of holding the device; or
use of one or more switches of said device;
A method according to any one of claims 1 to 3, comprising measuring one or more of The method of any preceding claim, wherein the device further comprises one or more outputs indicating to the user the end of measuring one or more behavioral biometrics. A method according to any preceding claim, wherein exiting the calibration mode returns the aerosol generating device to normal operating conditions. 7. The method of claim 6, wherein the normal operating state of the device is indicated to the user by one or more outputs of the device. A method according to any preceding claim, wherein the first set of sensors can be used to measure a plurality of behavioral biometrics. initiating an identification mode of the device;
measuring one or more behavioral biometrics of the user with the one or more sensors based on a calibrated user profile stored on the device;
calculating a user profile based on the measured one or more behavioral biometrics of the user;
comparing the calculated user profile to the user profile stored in the memory;
The method of any one of claims 1-8, further comprising 10. The method of claim 9, further comprising determining whether the calculated user profile and the user profile stored in the memory match. if the profiles match, unlock the device;
if the profiles do not match, keep the device locked;
11. The method of claim 10. A method according to any one of claims 9 to 11, further comprising terminating said identification mode of said device. An aerosol generating device comprising:
one or more sensors for calibrating and identifying a user of said device;
a microprocessor for controlling the one or more sensors;
a memory for storing one or more user profiles created by calibrating the device;
a software controlled device management system for performing the calibration and subsequent identification of the user of the device;
an aerosol-generating device, comprising: Aerosol-generating device according to claim 13, adapted to perform the method according to any one of claims 1-12. A computer-readable medium containing instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1-12.

Images