NL2022916B1 - Interaction Method and Interaction System of Smart Watch - Google Patents

Interaction Method and Interaction System of Smart Watch Download PDF

Info

Publication number
NL2022916B1
NL2022916B1 NL2022916A NL2022916A NL2022916B1 NL 2022916 B1 NL2022916 B1 NL 2022916B1 NL 2022916 A NL2022916 A NL 2022916A NL 2022916 A NL2022916 A NL 2022916A NL 2022916 B1 NL2022916 B1 NL 2022916B1
Authority
NL
Netherlands
Prior art keywords
signal
vibration signals
smartwatch
result
interaction method
Prior art date
Application number
NL2022916A
Other languages
Dutch (nl)
Inventor
Wu Kaishun
Chen Wenqiang
Wang Lu
Qiu Minghui
Original Assignee
Univ Shenzhen
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Univ Shenzhen filed Critical Univ Shenzhen
Application granted granted Critical
Publication of NL2022916B1 publication Critical patent/NL2022916B1/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/1613—Constructional details or arrangements for portable computers
    • G06F1/163—Wearable computers, e.g. on a belt
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
    • G06F3/014—Hand-worn input/output arrangements, e.g. data gloves

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • User Interface Of Digital Computer (AREA)
  • Electric Clocks (AREA)
  • Telephone Function (AREA)

Abstract

The present invention discloses an interaction method of a smart watch, including the following steps: 81, transmitting vibration signals based on human body, and collecting vibration signals of an accelerometer and a gyroscope of a smart watch; SZ, recognizing the vibration signals using an anomaly detection algorithm; S3, pre-processing the Vibration signals, and further classifying and recognizing the Vibration signals using an algorithm modified by k-Nearest Neighbor, S4, analyzing a feedback to a result from a user, and correcting timely to keep a stable recognition precision. Further, an interaction system of a smart watch is disclosed, including a signal detection module, a recognition and classifying module, and a real-time feedback module. The Vibration signals are transmitted based on human body, body parts are used as virtual screen, and the modified machine learning algorithm is combined, which actually broadens the interaction means of watches, and improves the user experience. The interaction method of the present invention is novel and interesting, and capable of meeting user demands and being widely used in text input and watch-games.

Description

Interaction Method and Interaction System of Smart Watch
Technical Field
The present invention relates to the field of interaction method of a smart device, specifically to an interaction method and an interaction system of a smart watch.
Background
Currently, wearable smart sensing devices are developing rapidly, and smart watches are especially popular. However, since the smart watches are worn on the wrist without the capability of being equipped with sufficiently large screen, people cannot input as they do on mobile phones. There are three main input modes of smart watches, namely, single touching, finger tracking and voice recognition. Single touching and finger tracking are limited by screen, and voice recognition is even more restricted due to information sensitivity. In order to get rid of situation of difficulty in providing inputs to smart watch, many research teams have performed related research. Mostly, additional devices are required for this purpose, and due to the purchase cost and learning cost of the additional devices, they are not extensively accepted.
Summary
In view of the above technical problems, the present application proposes an interaction method and an interaction system of a smart watch transmitting vibration signal based on human body. On the premise of fitting user’s usage habits, a new interaction means of smart watch is developed, thereby solving the problem of lacking interaction means of smart watch. The following technical solution is used in the present invention.
An interaction method of a smart watch includes the following steps:
51, transmitting vibration signals based on a human body, and collecting vibration signals of an accelerometer and a gyroscope of a smart watch;
52, recognizing the vibration signals using an anomaly detection algorithm;
53, pre-processing the vibration signals, and further classifying and recognizing the vibration signals using an algorithm modified by k-Nearest Neighbor;
54, analyzing a feedback to a result from a user, and correcting timely to keep a stable recognition precision.
Further, vibration signals along x-axis, y-axis, and z-axis of the accelerometer and the gyroscope are collected.
Further, step S2 of recognizing the vibration signals using the anomaly detection algorithm includes:
521, collecting z-axis data of the accelerometer;
522, fdtering the z-axis data of the accelerometer using a high-pass fdter;
523, setting a threshold value of an effective knock signal and a threshold value of a noise signal;
524, reading and selecting a signal with an amplitude smaller than the threshold value of the noise signal as a first state;
525, continuing to monitor to wait for a signal with an amplitude larger than the threshold value of the effective knock signal, recording a position of the signal with the amplitude larger than the threshold value of the effective knock signal as X, and setting an initial position of the signal to a position L before the position X, i.e., X-L;
526, continuing to monitor to wait for a continuous signal with an amplitude smaller than the threshold value of the noise signal, and when the continuous signal with the amplitude smaller than the threshold value of the noise signal, setting an end position of the signal as a current position;
527, acquiring signal data through the initial position and the end position, and judging whether a length of the signal meets a length range, if no, returning to S25, if yes, entering the next step;
528, performing a high-pass fdtering on the data, respectively calculating energies of the first m signal points and signal points after the mth signal point of the signal after filtering, judging whether the signal is larger than a threshold value of a signal-to-noise ratio, if yes, determining the signal as an effective signal, otherwise, determining as a noise signal, and returning to S25.
Further, step S3 includes:
531, pre-processing the signal with normalization, and subtracting a mean value from the signal and dividing by a variance;
532, in an initialization stage of a training model, storing the data processed in step S31 as a training sample in a database; in an actual use stage, classifying and recognizing the signal using the algorithm modified by k-Nearest Neighbor.
Further, the algorithm modified by k-Nearest Neighbor is specifically as follows: based on the dynamic time warping algorithm, the actual signal and the training signal are matched in a unit of frame, the shortest Manhattan distance between them is calculated, and used as the basis for classification and recognition of the k-Nearest Neighbor.
Further, step S4 includes:
541, after collecting a classifying result obtained in step S3, correcting an input result of user;
542, after the correcting, performing a certain extent of correction on the training sample, thereby keep the stability of the precision.
Further, in step S41, the actual input is corrected by providing a candidate key or through an associative result of an input method.
Further, step S42 specifically includes:
5421, when a corrected result coincides with the classifying result, performing no operation;
5422, when the corrected result does not coincide with the classifying result, for a sample of a same category as the classifying result in the training sample, deleting a sample with a largest distance calculated by the algorithm modified by k-Nearest Neighbor, and then replacing a current sample to a position of a deleted sample.
An interaction system of a smart watch includes:
a signal detection module, transmitting vibration signals based on a human body, and collecting vibration signals of an accelerometer and a gyroscope of a smart watch;
a recognition and classifying module, recognizing the vibration signals using an anomaly detection algorithm; pre-processing the vibration signals, and further classifying and recognizing the vibration signals using an algorithm modified by k-Nearest Neighbor; and a real-time feedback module, analyzing a feedback to a result from a user, and correcting timely to keep a stable recognition precision.
A program, executing the interaction method of the smart watch of the present invention.
Compared with the prior art, the advantages of the present invention are as follows. The vibration signals are transmitted based on human body, body parts (such as the back of hand) are used as virtual screen, and the modified machine learning algorithm is combined, which actually broadens the interaction means of watches, and improves the user experience. The interaction method of the present invention is novel and interesting, and capable of meeting user demands and being widely used in text input and watch-games.
Brief Description of the Drawings
FIG. 1 is a flow chart of an interaction method of the present invention;
FIG. 2 is a workflow chart of a signal detection module of the present invention;
FIG. 3 is a workflow chart of a feedback system of the present invention;
FIG. 4 is a structural diagram of the present invention;
FIG. 5 is a result of signal matching of an original dynamic time warping algorithm of the present invention; and
FIG. 6 is a result of signal matching of a dynamic time warping algorithm in a unit of frame after modification (a frame shift is 1, and a frame length is 3).
Detailed Description of the Embodiments
In order to make the objective, the technical solution and the advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.
The preferred embodiments of the present invention are further described in detail below with reference to the accompanying drawings.
The present invention discloses an interaction method and an interaction system of a smart watch based on vibration signals transmitted by human body and machine learning. As shown in FIG. 1, the interaction method of the present invention includes the following steps:
51, vibration signals are transmitted based on a human body, and an accelerometer and a gyroscope of a smart watch is controlled by a program to collect vibration signals of the accelerometer and the gyroscope of the smart watch;
Specifically, vibration signals along x-axis, y-axis, and z-axis of the accelerometer and the gyroscope are respectively collected;
52, the vibration signals are recognized using an anomaly detection algorithm;
53, the vibration signals are pre-processed, and the vibration signals are further classified and recognized using an algorithm modified by k-Nearest Neighbor;
54, a feedback to a result from a user is analyzed, and timely corrected to keep a stable recognition precision.
As shown in FIG. 2, step S2 of recognizing the vibration signals using the anomaly detection algorithm includes the following specific steps:
z-axis data of the accelerometer is first collected, a signal of 40 Hz obtained by high-pass filtering is selected, meanwhile, a signal with an amplitude smaller than the threshold value of the noise signal is read and selected, a signal length is preferably 10 signal points, and the threshold value of the noise signal is 0.015, at this time, a first state of signal detection is obtained; monitoring is continued to wait for a signal with an amplitude larger than the threshold value of the effective knock signal, a position of the signal with the amplitude larger than the threshold value of the effective knock signal is recorded as X, and an initial position of the signal is set to be a position L before the position X, i.e., X-L, and the threshold value of the effective knock signal is preferably 2; after receiving target signal, a continuous signal with an amplitude smaller than the threshold value of the noise signal is waited to be read, when the continuous signal with the amplitude smaller than the threshold value of the noise signal, an end position of the signal is set as a current position. Preferably, the signal length is 10 signal points, and the threshold value of the noise signal is 0.015. After detecting the obtained signal segment between the threshold values of the noise signal and the effective knock signal, the signal length and the signal-to-noise ratio are constrained. Preferably, the signal length L satisfies 37 < L < 60. After the signal length satisfies the constraint condition, energies of the first m signal points and signal points after the mt!1 signal point of the signal after filtering are respectively calculated to judge whether the signal is larger than a threshold value of a signal-to-noise ratio, if yes, the signal is determined as an effective signal, otherwise, determined as a noise signal. The threshold value of the signal-to-noise ratio is 10, until now, the signal is detected.
In the present embodiment, the step S3 of pre-processing the vibration signals, and further classifying and recognizing the vibration signals using the algorithm modified by kNearest Neighbor specifically includes the following steps:
first, the vibration signals along x-axis, y-axis, and z-axis of the accelerometer and the gyroscope of each sample are spliced by sensor category, and normalization is performed on three-axis data of corresponding sensor. Specifically, a mean value is subtracted from the data and divided by a variance of the data; then, in an initialization stage of a training model, the normalized data is stored as a training sample in a database, in an actual use stage, the signal is classified and recognized using the algorithm modified by k-Nearest Neighbor, specifically, based on the dynamic time warping algorithm, a distance between a test/input sample and the training sample is calculated, and a classifying result is obtained according to the distance.
Specifically, the dynamic time warping algorithm is an idea based on a dynamic programming. The object of the dynamic time warping is extended from the original one dimensional point to the three-dimensional (three-axis) frame, and the distance between each other is calculated, which can more accurately measure the similarity of two signals. At the same time, the power consumption of the algorithm is allowed to be reduced by adjusting the frame length and frame shift according to the actual sampling frequency and the demand, so as to obtain the desired performance. The problem that two signals cannot be compared due to the timing misalignment is solved, and the difference between the two is quantified. The distance is not limited to a Manhattan distance or an Euler distance.
FIG. 5 is a result of signal matching of an original dynamic time warping algorithm of the present invention; and FIG. 6 is a result of signal matching of a dynamic time warping algorithm in a unit of frame after modification (a frame shift is 1, and a frame length is 3). It can be seen that after increasing the constraint of the frame length and the frame shift, the signal matching mode of the dynamic time warping algorithm has been changed.
As shown in FIG. 3, in the step S4 of the present embodiment, after obtaining the classifying result in the step S2, the result is output to an application, and at the same time, the new sample X and the distance from the training sample obtained in the algorithm of the step S3 are recorded, and the feedback of the application is monitored. After receiving the feedback on the classifying result, operation is performed on the training sample according to the predetermined sample replacement strategy, thereby obtaining higher robustness. Specifically, after collecting the classifying result obtained in step 3, the user input result is corrected, and the actual input is corrected by providing a candidate key or by an associative result of the input method; after the correction, the training sample is corrected to a certain extent, thereby keeping the stability of the precision. Specifically, when the corrected result is consistent with the classifying result, no operation is performed; when the corrected result is inconsistent with the classifying result, for the sample of the same category as the classifying result in the training sample, the sample with the largest distance calculated by the algorithm modified by k-Nearest Neighbor is deleted, and then the current sample is replaced to the position of the deleted sample.
As shown in FIG. 4, a structure of the present embodiment includes three modules, namely, a signal detection module, a recognition and classifying module, and a real-time feedback module. The signal detection module detects the signal, and then normalizes the signal, the mean value is subtracted from the data and divided by the variance of the data, which is used as the input of the recognition and classifying module; the training (initialization) stage of the recognition and classifying module is simple signal storage operation, while put into use after the training is completed, and a modified classification algorithm is executed. The classifying result will be transmitted into the real-time feedback module.
The above descriptions are further detailed illustrations of the present invention in combination with specific/preferred embodiments, and it should not be understood that the specific embodiments of the present invention are limited to these descriptions. For those skilled in the art, several substitutions or modifications may be made to these described embodiments without departing from the inventive concept, all of which should be considered as falling within the protective scope of the present invention.

Claims (11)

ConclusiesConclusions 1. Een interactiemethode voor een smartwatch, omvattende de volgende stappen:An interaction method for a smartwatch, comprising the following steps: 51, het uitzenden van vibratiesignalen gebaseerd op het menselijk lichaam en het verzamelen van vibratiesignalen van een versnellingsmeter en een gyroscoop van een smartwatch;51, transmitting vibration signals based on the human body and collecting vibration signals from an accelerometer and a gyroscope from a smartwatch; 52, het herkennen van de vibratiesignalen gebruikmakend van een anomalie detecti ealgoritme;52, recognizing the vibration signals using an anomaly detection algorithm; 53, het voorverwerken van de vibratiesignalen, en het verder classificeren en herkennen van de vibratiesignalen gebruikmakend van een algoritme gemodificeerd door k-Nearest Neighbor;53, pre-processing the vibration signals, and further classifying and recognizing the vibration signals using an algorithm modified by k-Nearest Neighbor; 54, het analyseren van een terugkoppeling naar een resultaat van een gebruiker, en het tijdig corrigeren om een stabiele precisie voor herkenning te behouden.54, analyzing a feedback to a user's result, and correcting it in time to maintain a stable precision for recognition. 2. De interactiemethode voor de smartwatch volgens conclusie 1, waarbij vibratiesignalen langs x-as, y-as en z-as van respectievelijk de versnellingsmeter en de gyroscoop worden verzameld.The interaction method for the smartwatch according to claim 1, wherein vibration signals along x-axis, y-axis and z-axis are collected from the accelerometer and the gyroscope, respectively. 3. De interactiemethode voor de smartwatch volgens conclusie 2, waarbij de stap S2, van het herkennen van de vibratiesignalen gebruikmakend van een anomalie detectiealgoritme, omvat:The smartwatch interaction method according to claim 2, wherein the step S2 of recognizing the vibration signals using an anomaly detection algorithm comprises: 521, het verzamelen van z-as data van de versnellingsmeter;521, collecting z-axis data from the accelerometer; 522, het filteren van z-as data van de versnellingsmeter gebruikmakend van een hoogdoorlaatfilter;522, filtering z-axis data from the accelerometer using a high pass filter; 523, het instellen van een drempelwaarde voor een effectief klopsignaal en een drempelwaarde voor een ruissignaal;523, setting a threshold value for an effective knock signal and a threshold value for a noise signal; 524, het lezen en selecteren van een signaal met een amplitude kleiner dan de drempelwaarde van het ruissignaal als een eerste toestand;524, reading and selecting a signal with an amplitude smaller than the threshold value of the noise signal as a first state; 525, het blijven monitoren om te wachten op een signaal met een amplitude groter dan de drempelwaarde van het effectieve klopsignaal, het opnemen van een positie van het signaal met de amplitude groter dan de drempelwaarde van het effectieve klopsignaal als X, en het instellen van een initiële positie van het signaal als een positie L voor de positie X, i.e. X-L;525, continuing to monitor for a signal with an amplitude greater than the threshold value of the effective knock signal, recording a position of the signal with the amplitude greater than the threshold value of the effective knock signal as X, and setting a initial position of the signal as a position L for the position X, ie XL; 526, het blijven monitoren om te wachten op een continu signaal met een amplitude kleiner dan de drempelwaarde van het ruissignaal, en, wanneer het continue signaal met de amplitude kleiner is dan de drempelwaarde van het ruissignaal, het instellen van een eindpositie van het signaal als de huidige positie;526, continuing to monitor for a continuous signal with an amplitude smaller than the threshold signal of the noise signal, and, when the continuous signal with the amplitude is smaller than the threshold signal of the noise signal, setting an end position of the signal as the current position; 527, het verwerven van signaaldata door de initiële positie en de eindpositie, en het beoordelen of een lengte van het signaal voldoet aan een lengterange, het terugkeren naar S25 indien neen, het doorgaan naar de volgende stap indien ja;527, acquiring signal data by the initial position and the end position, and judging whether a length of the signal satisfies a length range, returning to S25 if no, continuing to the next step if yes; 528, het uitvoeren van een hoogdoorlaatfiltering op de data, respectievelijk het berekenen van energieën van de eerste m signaalpunten en signaalpunten na het signaalpunt m van het signaal na filtering, het beoordelen of het signaal groter is dan een drempelwaarde van een signaal-ruisverhouding, het vaststellen van het signaal als een effectief signaal indien ja, indien anders, het vaststellen als een ruissignaal en terugkeren naar S25.528, performing high pass filtering on the data, respectively calculating energies of the first m signal points and signal points after the signal point m of the signal after filtering, judging whether the signal is greater than a threshold value of a signal-to-noise ratio, determining the signal as an effective signal if yes, if different, determining as a noise signal and returning to S25. 4. De interactiemethode voor de smartwatch volgens conclusie 1, waarbij de stap S3 specifiek omvat:The interaction method for the smartwatch according to claim 1, wherein the step S3 specifically comprises: 531, het voorverwerken van het signaal met een normalisatie, en het aftrekken van een gemiddelde waarde van het signaal en het delen door een variantie.531, pre-processing the signal with a normalization, and subtracting an average value from the signal and dividing by a variance. 532, in een initialiseringsfase van een oefenmodel, het opslaan van de data verwerkt in stap S31 als een steekproef in een database; in een werkelijk gebruiksfase, het classificeren en herkennen van het signaal gebaiikmakend van het algoritme gemodificeerd door k-Nearest Neighbor.532, in an initialization phase of a practice model, storing the data processed in step S31 as a sample in a database; in an actual use phase, classifying and recognizing the signal using the algorithm modified by k-Nearest Neighbor. 5. De interactiemethode voor de smartwatch volgens conclusie 1, waarbij het algoritme gemodificeerd door k-Nearest Neighbor specifiek het volgende is: gebaseerd op het dynamic time warping algoritme worden het werkelijke signaal en het oefensignaal in overeenstemming gebracht in een eenheid van frame, een kortste afstand tussen het werkelijke signaal en het oefensignaal wordt berekend, en wordt gebruikt als een basis voor het classificeren en herkennen van de k-Nearest Neighbor.The smartwatch interaction method according to claim 1, wherein the algorithm modified by k-Nearest Neighbor is specifically the following: based on the dynamic time warping algorithm, the actual signal and the training signal are matched in a unit of frame, a shortest distance between the actual signal and the practice signal is calculated, and is used as a basis for classifying and recognizing the k-Nearest Neighbor. 6. De interactiemethode voor de smartwatch volgens conclusie 5, waarbij de afstand de Manhattan-afstand of de Euler-afstand is.The interaction method for the smartwatch according to claim 5, wherein the distance is the Manhattan distance or the Euler distance. 7. De interactiemethode voor de smartwatch volgens conclusie 1, waarbij de stap S4 omvat:The interaction method for the smartwatch according to claim 1, wherein the step S4 comprises: 541, na het verzamelen van een geclassificeerd resultaat verkregen in stap S3, het corrigeren van een inputresultaat van gebruiker;541, after collecting a classified result obtained in step S3, correcting an input result from user; 542, na het corrigeren, het uitvoeren van een correctie tot op zekere hoogte op de steekproef, waarbij een stabiliteit van een precisie behouden blijft.542, after correcting, performing a correction to a certain extent on the sample, while maintaining a stability of a precision. 8. De interactiemethode voor de smartwatch volgens conclusie 7, waarbij in de stap S41, de werkelijke invoer gecorrigeerd wordt door het voorzien van een kandidaatsleutel of door een associatief resultaat van een invoennethode.The interaction method for the smartwatch according to claim 7, wherein in step S41, the actual entry is corrected by providing a candidate key or an associative result of a check-in method. 9. De interactiemethode voor de smartwatch volgens conclusie 8, waarbij de stap S42 specifiek omvat:The interaction method for the smartwatch according to claim 8, wherein the step S42 specifically comprises: 5421, wanneer een gecorrigeerd resultaat samenvalt met het geclassificeerd resultaat, het niet uitvoeren van een bewerking;5421, when a corrected result coincides with the classified result, failure to perform an operation; 5422, wanneer het gecorrigeerde resultaat niet samenvalt met het geclassificeerde resultaat, voor een steekproef uit dezelfde categorie als het geclassificeerde resultaat in de steekproef, het verwijderen van een steekproef met een grootste afstand berekend door het algoritme gemodificeerd door de k-Nearest Neighbor, en het vervolgens vervangen van een positie van een verwijderde steekproef met een huidige steekproef.5422, when the corrected result does not coincide with the classified result, for a sample from the same category as the classified result in the sample, the removal of a sample with a largest distance calculated by the algorithm modified by the k-Nearest Neighbor, and the then replace a position of a deleted sample with a current sample. 10. Een interactiesysteem voor een smartwatch, omvattend:10. An interaction system for a smartwatch, comprising: een signaaldetectiemodule, voor het uitzenden van vibratiesignalen gebaseerd op het menselijk lichaam, en het verzamelen van vibratiesignalen van een versnellingsmeter en een gyroscoop van een smartwatch;a signal detection module, for transmitting vibration signals based on the human body, and collecting vibration signals from an accelerometer and a gyroscope from a smartwatch; een herkennings- en classificeringsmodule voor het herkennen van de vibratiesignalen gebruikmakend van een anomalie detectiealgoritme; het voorverwerken van de vibratiesignalen, en het verder classificeren en herkennen van de vibratiesignalen gebruikmakend van een algoritme gemodificeerd door k-Nearest Neighbor; en een real-time terugkoppelingsmodule voor het analyseren van een terugkoppeling naar een resultaat van een gebruiker, en het tijdig corrigeren om een stabiele precisie voor herkenning te behouden.a recognition and classification module for recognizing the vibration signals using an anomaly detection algorithm; pre-processing the vibration signals, and further classifying and recognizing the vibration signals using an algorithm modified by k-Nearest Neighbor; and a real-time feedback module for analyzing a feedback to a user's result, and correcting it in time to maintain a stable precision for recognition. 11. Een programma voor het uitvoeren van de interact emethode voor de smartwatch volgens één van de conclusies 1-9.A program for performing the interactive method for the smartwatch according to any of claims 1-9.
NL2022916A 2019-01-04 2019-04-10 Interaction Method and Interaction System of Smart Watch NL2022916B1 (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910013634.3A CN109840480B (en) 2019-01-04 2019-01-04 Interaction method and interaction system of smart watch

Publications (1)

Publication Number Publication Date
NL2022916B1 true NL2022916B1 (en) 2020-01-31

Family

ID=66476790

Family Applications (1)

Application Number Title Priority Date Filing Date
NL2022916A NL2022916B1 (en) 2019-01-04 2019-04-10 Interaction Method and Interaction System of Smart Watch

Country Status (3)

Country Link
CN (1) CN109840480B (en)
FR (1) FR3091606B1 (en)
NL (1) NL2022916B1 (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110169723A (en) * 2019-06-27 2019-08-27 九阳股份有限公司 A kind of food processor
CN111221420B (en) * 2020-01-13 2021-07-30 深圳大学 2D movement track identification method and system based on smart watch
CN111752388A (en) * 2020-06-19 2020-10-09 深圳振科智能科技有限公司 Application control method, device, equipment and storage medium
CN113741703A (en) * 2021-11-08 2021-12-03 广东粤港澳大湾区硬科技创新研究院 Non-contact intelligent earphone or glasses interaction method

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105810214B (en) * 2014-12-31 2019-11-05 展讯通信（上海）有限公司 Voice-activation detecting method and device
EP3065043A1 (en) * 2015-03-02 2016-09-07 Nxp B.V. Mobile device
CN107300971B (en) * 2017-06-09 2019-04-02 深圳大学 The intelligent input method and system propagated based on osteoacusis vibration signal

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
CHEN WENQIANG ET AL: "ViType: A Cost Efficient On-Body Typing System through Vibration", 2018 15TH ANNUAL IEEE INTERNATIONAL CONFERENCE ON SENSING, COMMUNICATION, AND NETWORKING (SECON), IEEE, 11 June 2018 (2018-06-11), pages 1 - 9, XP033367932, DOI: 10.1109/SAHCN.2018.8397098 *
HONGYI WEN ET AL: "Serendipity", HUMAN FACTORS IN COMPUTING SYSTEMS, ACM, 2 PENN PLAZA, SUITE 701 NEW YORK NY 10121-0701 USA, 7 May 2016 (2016-05-07), pages 3847 - 3851, XP058257548, ISBN: 978-1-4503-3362-7, DOI: 10.1145/2858036.2858466 *
MAOTIAN ZHANG ET AL: "iDial", PROCEEDINGS OF THE ACM ON INTERACTIVE, MOBILE, WEARABLE AND UBIQUITOUS TECHNOLOGIES, ACM, 2 PENN PLAZA, SUITE 701NEW YORKNY10121-0701USA, vol. 2, no. 1, 26 March 2018 (2018-03-26), pages 1 - 20, XP058396468, DOI: 10.1145/3191787 *
WENQIANG CHEN ET AL: "Virtual Keyboard for Wearable Wristbands", EMBEDDED NETWORK SENSOR SYSTEMS, ACM, 2 PENN PLAZA, SUITE 701NEW YORKNY10121-0701USA, 6 November 2017 (2017-11-06), pages 1 - 2, XP058396661, ISBN: 978-1-4503-5459-2, DOI: 10.1145/3131672.3136984 *

Also Published As

Publication number Publication date
FR3091606A1 (en) 2020-07-10
CN109840480B (en) 2021-08-13
FR3091606B1 (en) 2021-09-24
CN109840480A (en) 2019-06-04

Similar Documents

Publication Publication Date Title
NL2022916B1 (en) Interaction Method and Interaction System of Smart Watch
US10339371B2 (en) Method for recognizing a human motion, method for recognizing a user action and smart terminal
US20190170783A1 (en) Fall detection system using a combination of accelerometer, audio input and magnetometer
US10083397B2 (en) Personalized intelligent wake-up system and method based on multimodal deep neural network
US20120323521A1 (en) System and method for recognizing gestures
KR102036987B1 (en) Method and apparatus for sensing sleep state using biometric data and motion information
CN110970130A (en) Data processing method for attention defect hyperactivity disorder
KR101967342B1 (en) An exercise guide system by using wearable device
KR101579585B1 (en) User Action Monitoring Method, Apparatus, System And Recording Medium for Performing The Method
US10241583B2 (en) User command determination based on a vibration pattern
US20190343440A1 (en) System and method for measuring reaction time of a subject
CN110460809A (en) A kind of vagitus method for detecting, device and intelligent camera head apparatus
Günther et al. Activity recognition in manual manufacturing: Detecting screwing processes from sensor data
JPWO2014147785A1 (en) Motion detection device and motion detection method
KR102018416B1 (en) Device and system for monitoring micturition or ejaculation based on user&#39;s posture or change in posture, method of monitoring micturition or ejaculation, and computer-readable recording medium for implementing same method
WO2018036023A1 (en) Text input method and device for smart watch
Patterson et al. Sensor-based change detection for timely solicitation of user engagement
US11270109B2 (en) Interactive method and interactive system for smart watch
US20180353125A1 (en) Method and apparatus for determining sleep state using biometric information and motion information
Anju et al. Recent survey on Parkinson disease diagnose using deep learning mechanism
Kim et al. DeepSchema: Automatic Schema Acquisition from Wearable Sensor Data in Restaurant Situations.
Jagedish et al. Wearable Devices with Recurrent Neural Networks for Real-Time Fall Detection
CN113520339A (en) Sleep data validity analysis method and device and wearable device
Sao et al. An experimental analysis of human motion based data collected from smartphones' accelerometers
Kunst et al. Daily physical activities recognition using wearable devices: a systematic mapping