Safety control method and system for electronic cigarette

Abstract

The invention relates to the technical field of electronic cigarettes, in particular to a safety control method and system for an electronic cigarette. The method comprises the steps of obtaining a pre-stored matching template and multi-modal data input during user verification; the multi-modal data comprises smoking behavior data, and the matching template comprises a smoking behavior template; performing identity verification on the user according to the multi-modal data and the matching template, and if the verification is passed, sending a starting instruction to the electronic cigarette; after the electronic cigarette is started, whether the current smoking behavior data are matched with the smoking behavior template or not is judged, and if the current smoking behavior data are not matched with the smoking behavior template, a locking instruction or an instruction for requiring a user to perform verification again is sent to the electronic cigarette. According to the method, the use safety of the electronic cigarette can be ensured in the whole process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic cigarettes, and in particular to a safety control method and system for an electronic cigarette. BACKGROUND

[0002] With the popularity of electronic cigarettes, the safe use of electronic cigarettes has become the focus of attention of regulatory agencies, manufacturers and consumers. At present, the main safety protection scheme for electronic cigarettes mainly relies on a single biometric authentication means, such as fingerprint recognition, voiceprint recognition or password lock, etc. However, such single-mode authentication methods are easy to be cracked: fingerprint recognition can be cracked by high-precision 3D printing molds or silica gel fingerprint films; voiceprint recognition is vulnerable to audio playback attacks, and may fail to authenticate in a noisy environment; when setting an unlock password, users often set a weak password, such as 1234, for ease of remembering. Once the electronic cigarette is cracked, it may be misused by minors and unauthorized users, posing a safety hazard. Moreover, most electronic cigarettes cannot continuously monitor the legitimacy of the user during subsequent use. In summary, the prior art cannot guarantee the safety of electronic cigarette use throughout the process.

[0003] Therefore, how to guarantee the safety of electronic cigarette use throughout the process is a technical problem to be solved at present. SUMMARY

[0004] To solve the above technical problem that the safety of electronic cigarette use cannot be guaranteed throughout the process, the present application provides solutions in the following aspects.

[0005] In a first aspect, the present application provides a safety control method for an electronic cigarette, comprising: obtaining a pre-stored matching template and multi-modal data input by a user during verification; the multi-modal data comprises smoking behavior data, and the matching template comprises a smoking behavior template; performing identity verification on the user according to the multi-modal data and the matching template, and sending a start instruction to the electronic cigarette if the verification is passed; after the electronic cigarette is started, determining whether the current smoking behavior data matches the smoking behavior template, and sending a lock instruction or an instruction requiring the user to re-verify to the electronic cigarette if the current smoking behavior data does not match the smoking behavior template.

[0006] Further, performing identity verification on the user according to the multi-modal data and the matching template comprises: inputting the multi-modal data into a trained multi-modal fusion neural network to obtain a joint recognition score; determining that the user has failed to verify if the joint recognition score is less than a first pass threshold; and determining that the user has passed the verification if the joint recognition score is greater than or equal to the first pass threshold.

[0007] Further, the multi-modal data further comprises voiceprint data and fingerprint data; before the multi-modal data is input into the trained multi-modal fusion neural network, the multi-modal data is further subjected to feature extraction to extract a first feature vector from the fingerprint data, a second feature vector from the voiceprint data, and a third feature vector from the smoking behavior data.

[0008] Further, the feature extraction on the multi-modal data comprises: extracting a plurality of fingerprint features from the fingerprint data by minutia matching or image correlation method; converting the plurality of fingerprint features into the first feature vector; filtering and denoising the voiceprint data and performing endpoint detection to extract a plurality of acoustic features; converting the plurality of acoustic features into the second feature vector; extracting a plurality of behavior features from the smoking behavior data; converting the plurality of behavior features into the third feature vector; wherein the fingerprint features comprise endpoints, bifurcation points and ridge lines, the acoustic features comprise mel-frequency cepstral coefficients and mel-frequency spectrum, and the behavior features comprise inhalation duration and holding force.

[0009] Further, the determination of whether the current smoking behavior data matches the smoking behavior template comprises: calculating a similarity between the current smoking behavior data and the smoking behavior template; if the similarity is less than a second passing threshold, determining that the current smoking behavior data does not match the smoking behavior template; if the similarity is greater than or equal to the second passing threshold, determining that the current smoking behavior data matches the smoking behavior template.

[0010] Further, the calculation of the similarity between the current smoking behavior data and the smoking behavior template comprises: inputting the smoking behavior template into a trained LSTM model to obtain a behavior template vector; the smoking behavior template represents the smoking behavior data input by the user during registration; inputting the current smoking behavior data into the LSTM model to obtain a target feature vector; and calculating the similarity between the target feature vector and the behavior template vector based on cosine similarity or Euclidean distance.

[0011] Further, after the electronic cigarette is started, the method further comprises: obtaining environmental voiceprint data; determining whether any keyword in a preset keyword library matches the environmental voiceprint data; if so, sending a locking instruction to the electronic cigarette and / or sending an alarm information to a target contact, the alarm information comprising an audio clip.

[0012] Further, the matching template includes a voiceprint template, the voiceprint template characterizing voiceprint data input by the user during registration; and the sending of the locking instruction to the electronic cigarette and / or the sending of the alarm information to the target contact includes: pre-processing and feature calculation on the voiceprint template of the user to obtain a fundamental frequency baseline; calculating a fundamental frequency standard deviation and a fundamental frequency mean value of current voiceprint data; dividing the fundamental frequency standard deviation by the fundamental frequency mean value to obtain a fluctuation coefficient; and if the fluctuation coefficient is greater than the fundamental frequency baseline, sending the locking instruction to the electronic cigarette and / or sending the alarm information to the target contact.

[0013] Further, the obtaining of the pre-stored matching template and the multi-modal data input during user verification includes: obtaining the multi-modal data and the matching template from the electronic cigarette through a communication mode of Bluetooth, WiFi or USB.

[0014] In a second aspect, the present application provides a safety control system of an electronic cigarette, including a processor and a memory, the memory storing computer program instructions, when the computer program instructions are executed by the processor, realizing the safety control method of the electronic cigarette of the first aspect.

[0015] The present application has the beneficial effect that, compared with single biometric authentication, cross verification based on multi-modal data of the user avoids the risk that the single biometric feature may be forged, thereby improving the safety of the electronic cigarette start; and by controlling the start and locking of the electronic cigarette according to the similarity between the smoking behavior data and the pre-stored smoking behavior template during the use of the electronic cigarette, the safety during the use of the electronic cigarette is improved, thereby ensuring the safety of the use of the electronic cigarette throughout the whole process. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a flowchart schematically showing a safety control method of an electronic cigarette according to an embodiment of the present application;

[0017] Figure 2 is a structural block diagram schematically showing a safety control system of an electronic cigarette according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0019] The specific implementation of the present application will be described in detail below with reference to the drawings.

[0020] Figure 1is a flow chart schematically showing a safety control method of an electronic cigarette according to an embodiment of the present application.

[0021] In a first aspect, the present application provides a safety control method of an electronic cigarette. It should be noted that in the present embodiment, a terminal sends a starting instruction and a locking instruction to the electronic cigarette to control the starting and locking of the electronic cigarette. The terminal can be a mobile phone, and the mobile phone communicates with the electronic cigarette through a communication protocol such as Bluetooth, WiFi or USB. In alternative embodiments, a chip built-in the electronic cigarette can also be used to control the starting and locking of the electronic cigarette. Specifically, as shown in Figure 1 The method of the present application comprises the following steps.

[0022] S101, obtaining a pre-stored matching template and multi-modal data input by a user during verification.

[0023] In the present embodiment, the multi-modal data includes fingerprint data, voiceprint data and smoking behavior data; the matching template includes a fingerprint template, a voiceprint template and a smoking behavior template. The fingerprint data can be collected by a fingerprint sensor, such as an optical sensor, a capacitive sensor or an ultrasonic sensor; the voiceprint data can be collected by a microphone; and the smoking behavior data can be recorded by airflow and pressure sensors in the form of a waveform graph of 5 puffs. In addition, an IMU (Inertial Measurement Unit) can be additionally added to collect the smoking behavior data.

[0024] After the electronic cigarette collects the data, it sends the processed data through a communication protocol such as Bluetooth, WiFi or USB to a terminal, so that the terminal can obtain the data input by the user during verification in real time.

[0025] It should be noted that when the user uses the electronic cigarette for the first time, the collection of biological characteristics, i.e. the collection of the matching template, needs to be completed. Specifically, during registration, the user presses his / her finger on the fingerprint sensor several times, for example 3 times, and the complete fingerprint data obtained after processing is taken as the fingerprint template; the user reads a plurality of pre-set phrases, for example the phrase "safe start", and the voiceprint data collected at this time is taken as the voiceprint template; and the user uses the electronic cigarette to smoke / puff normally for several times, for example 5 times, and the smoking behavior data collected at this time is taken as the smoking behavior template. The smoking behavior data records the breathing rhythm and holding habit of the user when using the electronic cigarette.

[0026] S102, verifying the identity of the user according to the multi-modal data and the matching template, and controlling the starting and locking of the electronic cigarette according to the verification result.

[0027] In the present embodiment, the verification result includes: verification passed and verification failed.

[0028] Specifically, the multi-modal data is input into the trained multi-modal fusion neural network to obtain a joint recognition score, the value range of the joint recognition score being 0 to 1. In the process of multi-modal feature fusion, the multi-modal fusion neural network adopts an attention mechanism for weighting. After the multi-modal fusion neural network is trained using a large amount of data, it has learned how to weigh the contribution degree of each biological feature. For example, since a fingerprint is usually the most reliable, the multi-modal fusion neural network determines that the contribution degree of the fingerprint can be 0.6; since a voiceprint is susceptible to interference and has low reliability in a relatively noisy environment, the multi-modal fusion neural network determines that the contribution degree of the voiceprint can be 0.1. One biological feature corresponds to one modality data, and the contribution degree is also the weight.

[0029] In one embodiment, before the multi-modal data is input into the trained multi-modal fusion neural network, the method of the present application further comprises: performing feature extraction on the multi-modal data to extract a first feature vector, i.e., a fingerprint feature vector, from the fingerprint data, a second feature vector, i.e., a voiceprint feature vector, from the voiceprint data, and a third feature vector, i.e., a behavior feature vector, from the smoking behavior data.

[0030] Specifically, the minutia matching or image correlation method is used to extract fingerprint features such as ridges, valleys, endpoints, and bifurcation points from the fingerprint data; the fingerprint features are converted into the first feature vector; the voiceprint data is filtered and denoised and endpoint detection is performed to extract acoustic features that can reflect the physical structure of the sound-producing organ, such as MFCC (Mel Frequency Cepstral Coefficient) and Mel spectrum, and then the acoustic features are converted into the second feature vector; similarly, behavior features such as inhalation duration, inhalation intensity curve, interval time, holding angle, and holding pressure are extracted from the waveform graph obtained by 5 puffs, and then the behavior features are converted into the third feature vector. The first feature vector, the second feature vector, and the third feature vector are used as input data of the multi-modal fusion neural network.

[0031] Further, it is determined whether the joint recognition score is less than a first passing threshold. If the joint recognition score is less than the first passing threshold, it is determined that the user fails the verification, no start instruction is sent to the electronic cigarette, i.e., the electronic cigarette remains locked, and a failure log is recorded. If the joint recognition score is greater than or equal to the first passing threshold, it is determined that the user passes the verification, and a start instruction is sent to the electronic cigarette to control the atomizer to start working according to the start instruction.

[0032] In the embodiment, the first passing threshold is set to 0.9, and in other optional embodiments, the skilled in the art can set it according to actual needs, for example, set to 0.8. It can be understood that the joint recognition score is the result of cross-validation of multiple biometric characteristics, and the higher the joint recognition score, the higher the possibility that the user is the owner of the electronic cigarette, that is, the higher the possibility of passing the identity verification. Therefore, when the joint recognition score is greater than or equal to 0.9, the electronic cigarette is started, which can ensure that the current user is the owner of the electronic cigarette, and the owner referred to is the person who inputs the matching template.

[0033] By fusing fingerprints, voiceprints, behaviors and other multi-dimensional biometric characteristics to verify the identity of the user, the risk that a single biometric characteristic is easily cracked can be avoided, thereby avoiding the problem that the electronic cigarette is used by minors or non-owners, and improving the safety of the use of the electronic cigarette.

[0034] S103, after the electronic cigarette is started, the current puffing behavior data of the user is analyzed, and the starting and locking of the electronic cigarette are controlled according to the analysis result.

[0035] Specifically, after the electronic cigarette is started, the current puffing behavior data is sent to the terminal in real time, and the terminal judges whether the current puffing behavior data matches the pre-stored puffing behavior template. If the current puffing behavior data does not match the pre-stored puffing behavior template, a locking instruction or an instruction requiring the user to re-verify is sent to the electronic cigarette.

[0036] In the embodiment, whether the current puffing behavior data matches the pre-stored puffing behavior template can be judged by calculating the similarity between the current puffing behavior data and the pre-stored puffing behavior template. Specifically, the puffing behavior template is input into a trained LSTM (Long Short-Term Memory, long short-term memory network) model to obtain a behavior template vector; and the current puffing behavior data is input into the above-mentioned LSTM model to obtain a target feature vector; then the similarity between the target feature vector and the behavior template vector is calculated by cosine similarity or Euclidean distance; it is judged whether the similarity is less than a second passing threshold. If the similarity is less than the second passing threshold, it is determined that the current puffing behavior data does not match the pre-stored puffing behavior template; if the similarity is greater than or equal to the second passing threshold, it is determined that the current puffing behavior data matches the pre-stored puffing behavior template. It should be noted that before the judgment, the similarity needs to be normalized to the range of 0 to 1.

[0037] In the embodiment, the second passing threshold can be set to 0.8 or 0.9, and in subsequent use, the second passing threshold can also be adjusted. Specifically, the similarity when the user passes the verification is periodically counted, and the mean and standard deviation of the similarity are calculated. The mean minus the standard deviation obtains a new second passing threshold. Wherein, the periodic counting can be once a week.

[0038] In use, the behavior characteristics of the user can fluctuate over time, for example, fluctuations caused by changes in mood, attention, health status, etc. These fluctuations are normal fluctuations, and if a fixed threshold is always used, such fluctuations may be misjudged as non-personal operation, resulting in locking or frequent verification failure of the electronic cigarette. Therefore, by dynamically adjusting the second passing threshold, a fault tolerance space can be reserved, allowing normal short-term fluctuations to still pass the verification, thereby improving the smoothness of the electronic cigarette use and user experience.

[0039] By performing identity verification based on the current smoking behavior data of the user during use of the electronic cigarette, the user's use experience can be unaffected, and unauthorized use during use can also be avoided, for example, the holder places the electronic cigarette on the table after completing the start verification and smoking for a period of time, which may be used by unauthorized users such as minors when they see it; for example, the non-holder requests the holder to complete the verification before starting the electronic cigarette, and then uses it by himself.

[0040] By comparing the smoking behavior data of the actual user with the pre-stored smoking behavior template, when the similarity is found to be less than the second passing threshold, secondary verification is triggered, for example, requiring fingerprint confirmation again, or directly locking the electronic cigarette, which avoids the situation of being used by unauthorized users in the middle, and further improves the safety of the electronic cigarette use.

[0041] By performing identity verification on the user based on multi-modal data before starting, and performing identity verification on the user based on smoking behavior data during use, the safety of the electronic cigarette use is improved compared to single biometric authentication, and compared to one-time identity verification only at the start, the user's legality can be continuously monitored during use of the electronic cigarette, thereby ensuring the safety of the electronic cigarette use throughout the process.

[0042] In one embodiment, after starting the electronic cigarette, the method of the application further comprises: acquiring environmental voiceprint data; determining whether any one of the keywords in the preset keyword library matches the environmental voiceprint data, and if so, sending a locking instruction to the electronic cigarette and / or sending an alarm information to the target contact, the alarm information containing an audio clip. The keywords in the keyword library can be set by the holder, for example, set to "give me the electronic cigarette", or use the system default keyword.

[0043] Exemplarily, the judgment can be made by DTW (Dynamic Time Warping). Specifically, a feature sequence is extracted from the environmental voiceprint data, a distance matrix of the feature sequence of the environmental voiceprint data and the feature sequence of any one keyword in the keyword library is calculated, an accumulated distance is obtained by dynamic programming, and it is judged whether the accumulated distance is less than a preset distance threshold. If yes, it is determined that the environmental voiceprint data matches the keyword in the keyword library. The extracted feature sequence is an MFCC feature.

[0044] In addition, whether the preset keyword is contained in the environmental voiceprint data can also be detected by machine learning. In an optional embodiment, if the preset keyword is detected in the environmental voiceprint data, the emotion of the user is identified, and if it is identified that the user's emotion is nervous, an alarm information is sent to the target contact person. Specifically, the voiceprint template of the user is preprocessed and feature calculation is performed to obtain a fundamental frequency baseline, that is, the mean value and the standard deviation of the fundamental frequency of the voiceprint template, the standard deviation of the fundamental frequency is divided by the mean value of the fundamental frequency to obtain the fundamental frequency baseline, and the maximum value in the multiple fundamental frequency baselines is taken as the final fundamental frequency baseline.

[0045] Further, the standard deviation and the mean value of the fundamental frequency of the current voiceprint data are calculated, the standard deviation of the fundamental frequency of the current voiceprint data is divided by the mean value of the fundamental frequency to obtain a fluctuation coefficient, and if the fluctuation coefficient is greater than the fundamental frequency baseline, it is determined that the user's emotion is nervous.

[0046] In an optional embodiment, the fluctuation of the fundamental frequency and the decrease of the energy entropy can also be considered comprehensively to trigger the locking instruction and / or the alarm instruction.

[0047] For example, in public places such as subways and shopping malls, when the user is forced to use an electronic cigarette by a coercer who orders the user to "quickly puff", the system detects the keyword "puff" and identifies the user's nervous performance: the fundamental frequency fluctuates and the energy entropy decreases, and then immediately sends a locking instruction to the electronic cigarette, so that the coercer cannot continue to use the electronic cigarette or force the user to use the electronic cigarette.

[0048] For another example, when the user is restricted in personal freedom, the user can seek help from an emergency contact person by intentionally triggering a keyword, for example, the user intentionally says "give me an electronic cigarette", and the mobile phone immediately sends a help-seeking information to the target contact person. In this process, the electronic cigarette also sends an encrypted audio segment to the mobile phone, which contains background environmental sound such as quarrel sound, which can provide evidence for subsequent evidence collection, and assist the police to respond quickly through the mobile phone positioning function.

[0049] In summary, by fusing the multi-dimensional features such as fingerprints, voiceprints, and behavior patterns, the anti-fraud performance is enhanced, for example, if the fingerprints are forged, the system can require the user to perform voiceprint verification and puffing behavior simulation at the same time; by continuously monitoring whether there is abnormal use according to the user's puffing behavior, for example, abnormal use such as rapid and continuous puffing by children, if there is, the electronic cigarette lock is immediately triggered, so that the attack success rate is reduced by 2-3 orders of magnitude; through multi-modal signal cross-validation, such as voiceprint emotion analysis and keyword matching, the coercion state can be detected, and illegal operations can be actively prevented.

[0050] Figure 2 is a structural block diagram schematically showing a safety control system of an electronic cigarette according to the present embodiment.

[0051] In a second aspect, the present application further provides a safety control system of an electronic cigarette. As shown in the accompanying drawings, the system comprises a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions, when executed by the processor, implement the safety control method of the electronic cigarette according to the first aspect of the present application. Figure 2

[0052] The system further comprises a communication interface and other components well known to those skilled in the art, the settings and functions of which are known in the art, and thus will not be described here.

[0053] In the present application, the aforementioned memory can be any tangible medium containing or storing a program, which can be used by or in conjunction with an instruction execution system, device or apparatus. For example, the computer readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module or both. Any such computer storage medium can be part of the device or accessible or connectable to the device. Any application or module described in the present application can be implemented using computer readable / executable instructions that can be stored or otherwise held by such computer readable medium.

[0054] ​In the description of the present specification, the meaning of "a plurality of" is at least two, such as two, three or more, and the like, unless otherwise explicitly specifically limited. In addition, the division of steps of the above method is only for the purpose of clear description, and when implemented, it can be combined into one step or some steps can be split and decomposed into multiple steps, as long as the same logical relationship is included.

[0055] While the present specification has shown and described several embodiments of the present application, it is to be understood that the same are presented by way of example only and not limitation. As such, numerous changes, modifications and substitutions can be made by one having ordinary skill in the art without departing from the spirit and scope of the present application. It should be understood that in the process of practicing the present application, various alternatives to the embodiments of the present application described herein can be employed.

Claims

1. A safety control method of an electronic cigarette, characterized by, The method comprises the following steps: obtaining pre-stored matching templates and multi-modal data input during user authentication; the multi-modal data comprises smoking behavior data, and the matching templates comprise smoking behavior templates; authenticating the user according to the multi-modal data and the matching templates, and sending a start instruction to the electronic cigarette if the authentication is passed; after the electronic cigarette is started, determining whether the current smoking behavior data matches the smoking behavior template, and sending a lock instruction or an instruction requiring the user to re-authenticate if the current smoking behavior data does not match the smoking behavior template.

2. The safety control method of an electronic cigarette according to claim 1, characterized in that, The method for authenticating the user according to the multi-modal data and the matching templates comprises the following steps: inputting the multi-modal data into a trained multi-modal fusion neural network to obtain a joint recognition score; if the joint recognition score is less than a first passing threshold, determining that the user fails to pass the authentication; if the joint recognition score is greater than or equal to the first passing threshold, determining that the user passes the authentication.

3. The safety control method of the electronic cigarette according to claim 2, characterized in that, The multi-modal data further comprises voiceprint data and fingerprint data; before the multi-modal data is input into the trained multi-modal fusion neural network, the method further comprises the following steps of: performing feature extraction on the multi-modal data to extract a first feature vector from the fingerprint data, a second feature vector from the voiceprint data, and a third feature vector from the smoking behavior data.

4. The safety control method of an electronic cigarette according to claim 3, characterized in that, The method for performing feature extraction on the multi-modal data comprises the following steps: extracting a plurality of fingerprint features from the fingerprint data by using minutia point matching or image correlation; and converting the plurality of fingerprint features into the first feature vector; performing filtering and endpoint detection on the voiceprint data to extract a plurality of acoustic features; and converting the plurality of acoustic features into the second feature vector; extracting a plurality of behavior features from the smoking behavior data; and converting the plurality of behavior features into the third feature vector; The fingerprint features comprise endpoints, bifurcation points, and ridge lines, the acoustic features comprise mel-frequency cepstral coefficients and mel-frequency spectra, and the behavior features comprise inhalation duration and holding force.

5. The safety control method of an electronic cigarette according to claim 1, characterized in that, The method for determining whether the current smoking behavior data matches the smoking behavior template comprises the following steps: calculating the similarity between the current smoking behavior data and the smoking behavior template; if the similarity is less than a second passing threshold, determining that the current smoking behavior data does not match the smoking behavior template; if the similarity is greater than or equal to the second passing threshold, determining that the current smoking behavior data matches the smoking behavior template.

6. The safety control method of an electronic cigarette according to claim 5, characterized in that, The method for calculating the similarity between the current smoking behavior data and the smoking behavior template comprises the following steps: inputting the smoking behavior template into a trained LSTM model to obtain a behavior template vector; the smoking behavior template represents the smoking behavior data input by the user during registration; inputting the current smoking behavior data into the LSTM model to obtain a target feature vector; calculating the similarity between the target feature vector and the behavior template vector based on cosine similarity or Euclidean distance.

7. The safety control method of an electronic cigarette according to claim 1, characterized in that, After the electronic cigarette is started, the method further comprises the following steps: obtaining environmental voiceprint data; determining whether any keyword in the preset keyword library matches the environmental voiceprint data, and if so, sending a locking instruction to the electronic cigarette and / or sending an alarm message to the target contact, the alarm message including an audio clip.

8. The safety control method of an electronic cigarette according to claim 7, characterized in that, The matching template includes a voiceprint template representing voiceprint data entered by the user during registration. The sending of the locking instruction to the electronic cigarette and / or the sending of the alarm message to the target contact includes: preprocessing and feature calculation of the voiceprint template of the user to obtain a fundamental frequency baseline; calculating a fundamental frequency standard deviation and a fundamental frequency mean value of the current voiceprint data; dividing the fundamental frequency standard deviation by the fundamental frequency mean value to obtain a fluctuation coefficient; if the fluctuation coefficient is greater than the fundamental frequency baseline, sending a locking instruction to the electronic cigarette and / or sending an alarm message to the target contact.

9. The safety control method of an electronic cigarette according to claim 1, characterized in that, obtaining a pre-stored matching template and multi-modal data input by the user during verification, including obtaining the multi-modal data and the matching template from the electronic cigarette through a Bluetooth, WiFi or USB communication mode.

10. A safety control system for an electronic cigarette, characterized in that, including a processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the safety control method of the electronic cigarette of any one of claims 1-9.

Images