IT202000016267A1 - AASWE – ANTI-ASSAULT SMART WEARABLE EQUIPMENT AN ONLINE SYSTEM BASED ON INTELLIGENT GARMENTS TO ALERT THE AUTHORITIES IN THE EVENT OF AGAINST - Google Patents

Description

ASWE ? Anti-assault smart wearable equipment

A system based on smart clothing to alert the authorities? in case of aggression

Abstract

This document refers to intelligent anti-aggression garments connected online to an operations centre, designed to provide protection for a female user in the event of attempted physical aggression. The fields of the reference technique are that of intelligent clothing and personal defense technologies based on the use of alarms. The system consists of a complete intimate interweaving of sensors connected to a controller capable of transmitting an alert signal to an operations centre, in turn capable of sending an immediate distress call to the authorities, providing the exact location of the user . The user can request help in two ways: directly, with a special electronic key, by pressing a button located on it to signal a high-risk situation or, where possible, an aggression in progress. Alternatively, the system pu? decide independently in the event of a sudden attack suffered by the user, therefore unable to use the key. Automatic activation? based on input data provided by specific sensors of various types: audio, biometrics, specific location (exact place with an approximation of meters) and mechanical stress, ie? capable of sensing the stress inflicted on clothing. This data is processed by an automatic learning algorithm, possibly also connected to the Internet of Things (IoT), and trained to recognize dangerous situations, and therefore able to take the user's place in an emergency.

Pictures (2)

Worldwide applications

Year 2020 country Italy

[0001] PRESENTATION OF THE INVENTION AND ITS FIELD OF APPLICATION

The present work refers to an intelligent wearable equipment, conceived with finality? of self-defense and connected to an operations center capable of immediately alerting the authorities? in case of physical assault. Pi? specifically, the present invention refers to a sensor system woven into an underwear set, consisting of a slip and a bra capable of emitting a signal addressed to an operations center capable of alerting the police in the event of physical aggression , managed automatically by machine learning software, which receives the detection of various types of parameters as input from the sensors incorporated in the garments. ? also foreseen the possibility manual activation by means of an electronic key equipped with a button which, when pressed, alerts the operations center in the same way as the automatic system.

[0002] ? MOTIVATION OF THE INVENTION

Personal safety? a big problem all over the world. In particular, crimes with a sexual background are among the most serious and odious among the existing ones. Crimes against women are a major and growing concern of governments around the world.

[0003] NEED AND UNIQUENESS? OF THE INVENTION

Not ? immediately build a passive defense system that simultaneously possesses a high level of safety for the user and, at the same time, minimizes false alarms due to possible interpretation errors when the system automatically takes over the user.

The safety in the use of the system must be guaranteed by a series of conditions which minimize the margin of error in the activation of the same. These conditions are of a physical type (vocal inputs, biometric data, mechanical stress on the fabric), of an IT type (geolocation, interaction with other equipment such as video cameras, computers, ...) and of a behavioral type (selection of a certain mode of 'use by the user). Given what since [0001-0002], does the opportunity/need exist? of an intelligent anti-aggression garment cos? as described and claimed herein.

BRIEF SUMMARY OF THE INVENTION

[0004] OBJECT OF THE PATENT

The object of the patent? a system to prevent physical attacks, and in particular sexual ones, based on garments capable of alerting the authorities in real time? under suitable conditions.

The system consists of a combination of a manual warning device that can be activated by the user and an automatic activation system.

The manual activation system ? a key with a button that can be activated by pressing: when the button is pressed, the operations center immediately alerts the authorities providing the exact physical location of the user through an integrated geolocation system.

The system provides for: the detection of different types of parameters: vocals (screams, suffocation noises, scuffles, shots); biometrics (heart rate, sweating, speed changes, muscle tone); location (geographical location, specific location such as parking lots, streets, home); of tissue stress (tear attempts or violent removal), and provides for the possibility? of connection to the IoT (urban, domestic cameras, etc., ...) according to modality? specified by the user (shutdown, custom partial activation).

[0005] FUNCTIONS OF THE PROPOSED INVENTION ACTIVE AUTOMATICALLY

The system provides a detection tool which allows to: (1) perform accurate measurements of the parameters indicated in point [0004] such as, for example: abnormal sounds, heart rate, perspiration, muscle tone, tissue stress from traction or torsion; (2) process the measurements via an intelligent controller and check whether a dangerous situation exists (3) send an alarm to the operations center; and (4) receive immediate feedback from the control panel which automatically responds with a detection of location parameters and alerts the authorities.

[0006] FUNCTIONS OF THE PROPOSED INVENTION WHICH CAN BE ACTIVATED MANUALLY

The system also allows you to manage an emergency that has not been activated by a direct call from the user by pressing an alarm button located on the customized alarm key.

[0007] THE INTEGRATED LEARNER

The features? automatic are managed by a learner (machine learning algorithm), designed to detect the parameters considered in a normal situation? and to identify by learning the dangerous situations from the others.

[0008] SUMMARY DESCRIPTION OF THE STRUCTURE OF THE SYSTEM

The system includes: a controller/transmitter, a network of sensors dedicated to detecting: biometric parameters, physical location, stress produced on clothing, input from any connection to the IoT, cameras and other input peripherals for the control unit to identify with certainty the aggression. The system ? completed by the manual activation electronic key, and by a machine learning algorithm wired into the controller.

[0009] SYSTEM OPERATION PHASES

The system works according to the following phases:

1. monitor user parameter measurements;

2. receive user input;

3. process the data with a machine learning algorithm in order to intervene in the event of an aggression that is reputed to be certain even without input from the user, so ? foreseen a continuous monitoring and learning circuit;

4. memorize the information acquired in phase 2 to improve phase 3 in a constant learning process.

BRIEF DESCRIPTION OF THE FIGURES

[0010] Figure 1 illustrates the basic scheme of the system: the underwear set in its upper and lower parts, on which to position the vocal (1), biometric (2), location (3) and tissue stress (4) sensors ); the database (6) which collects this data, sent constantly by the controller/transmitter (5); the machine learning algorithm (7) which interprets the data to understand whether the user's situation is normal or an intervention is necessary, in which case it alerts the operations center (9); the electronic key (8) with which the user can? raise the alarm directly.

[0011] Figure 2 illustrates the basic algorithm followed by the automatic alarm system in its operation.

DETAILED DESCRIPTION OF THE OPERATION

[0012] The system must provide protection to the user through a continuous collection of data of different types, which define a certain configuration of the user?s security state, composed starting from the interpretation of the same data by an integrated learner such as set of programmable instructions in a control device that receives data and pu? interpret them by means of the aforementioned learner. For this purpose ? It is possible to use a suitably programmed standard FPGA board. The system must also provide that the activation can be controlled by the user who perceives the danger, cos? as from an automatic monitoring system that can take over the user if the latter, attacked by surprise, does not have the time and means to activate it manually.

[0013] The sensors positioned on the lower and upper parts of the suit represent the input of the autonomous monitoring system. The vocal sensors are in charge of detecting the noises and voices around the user: they must, in particular, pay attention to specific noises such as screams, muffled cries, noises relating to fights. The biometric sensors must listen to some parameters (the set of which can be variable) such as: perspiration, heart rate, muscle tone. Localization sensors should allow the user to be referenced with a precision in the order of meters, adding, integrated in a geo-localization system such as Google Maps, to know if the user is ? in a particular place outside the home (car parks, city streets, countryside, ...) or inside your own home. The definition of the activation thresholds ? made in the calibration of the system, in which ? supplied to the system itself an example of a dangerous situation to be identified as a threshold (such as, for example, the level in dB of the voice). The complex of these data pu? be treated by the machine learning algorithm according to a series of basic configurations (e.g.: risk zone: isolated parking lot at night; relatively safe zone: user's home in the absence of sounds and muscle relaxation, etc. ... ).

[0014] The basic configuration of the inputs can? be integrated into the Internet of Things (IoT) to enhance the correctness of the interpretation of the danger situation by the automatic system. This indicates that the finished product can? interact with public and domestic surveillance systems (cameras, alarm systems), in order to increase the degree of accuracy of the evaluation underlying the activation of the alarm.

[0015] The machine learning algorithm which presides over the danger assessment learns from the data. This involves an interaction with the user. In the early stages, this could involve requesting the same algorithm for feedback from the user, who could occasionally be called upon to confirm or deny a warning from the algorithm.

[0016] The danger configuration is detected if more? important parameters converge simultaneously to indicate it: physical alteration, combined with location outside the home at night, and a tensile stress of the garment, could indicate, for example, an aggression by the user, who could have been silenced.

[0017] The basic operating scheme of the invention ? the following:

If the button located on the electronic key (8) is pressed, the system alerts the operations centre, which warns the authorities giving the user's location to the authorities.

In the presence of torsion of the fabric detected by the traction sensor (4) beyond a pre-set traction threshold, which will have to? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not;

In the presence of anomalous sounds detected by the vocal sensor (1) beyond a pre-set threshold in dB, which should? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not;

In the presence of altered biometric parameters detected by the biometric sensors (2) beyond a pre-set stress threshold (heart rate and perspiration), which should? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not;

Does the locator (3) have a list of types of place-time considered at risk, such as covered parking? after 21, configurable in a personalized way and of the necessary localization resolution to determine if the user? in these areas and at these times: the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as a danger configuration or not;

The basic operating scheme of the automatic control algorithm? the following.

there is an initial set of sensors given by:

speech sensors (SV) from 1 to n

biometric sensors (SB) from 1 to m

location sensors (SL) 1 to p

tissue stress sensors (SST) 1 to q

functional forms and thresholds are given for each sensor, for example for vocal sensors the level in decibels of human voices including that of the user, fixed at SV1=80db;

all the other specifications of the other sensors are given (e.g.: SL1, physical location with approximation of meters, SL2, type of location (covered garage, suburban road, etc.?)

? It is possible to have a danger configuration such as, for example:

SV1 voice >80db

SL2 covered garage

Time: 02:00 AM

SST: stress beyond threshold

The general functional form ?:

Input configuration at time t: it starts from the Speech Sensor Input 1 (ISV1) up to the Tissue Stress Sensor Input q initialized to an initial value of the time t at the start of operation

ISV1 t value

?

ISSTq t-value

Current configuration: during operation the values change and the system records these changes continuously for any m greater than t

SV1P t+m value

?

SST1P t+m value

Differential

D-SV1 t-value - t+m value

D-ST1 t-value - t+m-value

If the values of one or more? sensors exceed the danger threshold defined for them, the system acts as follows:

if the sensors that exceed the danger threshold are of three or more types, whatever they are, the alarm is triggered

if the sensors that exceed the danger threshold are between two, the system sends a warning to the user who, to prevent activation of the alarm, must press the manual key button twice within 5 minutes, otherwise the alarm will go off.

if the sensors are of only one type, the automatic system does not start but sends a warning to the user who must press the manual key button twice within an hour, otherwise the alarm will go off.

Every time a danger signal takes place, whether it is a false signal, deactivated by the user, or a real one, the record made up of the values detected by the sensors on that occasion ? stored in the form of a danger signal, respectively, FALSE or TRUE.

The relative differentials associated with the sensors with respect to the initial configuration are also stored in the FPGA card used for storage and control.

Over time, the system hones capabilities? predictive, based on continuous detection of sensor values and comparison with previous TRUE or FALSE hazard configurations. This ability? predictive allows the system to work automatically with ever greater efficiency, minimizing the level of FALSE alerts and the need for of the manual intervention of the user to report the same.

[0018] Some basic danger configurations can be supplied together with the system, in the form of configurations written in the memory of the FPGA such as to generate, where necessary, an immediate request for help without manual confirmation by the user by pressing the button located on the key.

[0019] Once a dangerous situation has been recognised, the automatic system intervenes, alerting the forces of order.

[0020] The state of the art for which ? been made prioritization search? which turns out to be useful to the invention ? the following:

This application proposes to use pre-existing technologies in an innovative and coordinated way, for which the following references are reported, contextualised by comparison and/or integration with the proposed invention:

1. as regards existing technologies relating to alarm signaling devices for attacks on the person, we point out the Emergency call device patent WO2010088712A1, which concerns an emergency call device that can? be used to automatically detect and report incidents of a criminal nature. Said device comprises a digital camera and/or a microphone which interacts with it to optically and/or acoustically record events, an antenna having electronics which interacts with it to acquire dedicated geographical coordinates and to establish a radio connection to a distant station and to the automatic transmission of information thereto, and an optical and/or acoustic signaling device, optically and/or acoustically highlighting the function of the emergency call device at the site of the emergency call device, wherein all said means may be started with a push of a button. The emergency call device is part of a band to be worn on the body of a human being. The found indicated ? an example of a system based on the same logic of the invention proposed as a wearable alarm system capable of alerting an operations centre. The proposed invention differs from it for the management of a plurality? greater input (also including biometric parameters, for example) and for the presence of an artificial intelligence system capable of interpreting the inputs in the event of a disabling aggression suffered by the user.

with regard to existing technologies relating to motion detection devices incorporated in intelligent clothing: A kind of exercise data detection method and intelligent wearable device - CN110495892 (A) The invention describes methods of detecting data produced by an intelligent wearable device comprising a sensor movement and a magnetic sensor.

Apparatus and method for controlling smart wear KR20140120996 (A) The present invention ? relating to a control apparatus for the estimation of the trajectories of movement detected by a sensor and a unit? of data processing.

The proposed invention differs from the two aforementioned inventions, in that they only express a few possible methods of carrying out a necessary part of the operations, in this case the detection of a movement by the user who wears the invention itself.

.

as regards existing technologies relating to devices for recharging electronic devices incorporated in smart wear: System and Method for Charging Multiple Smart Wear KR20190062966 (A) which concerns a system for recharging smart clothes based on the use of a technology capable of making more simple to recharge them with only the actions normally performed with regard to clothes. The proposed invention proposes to benefit, optionally but not necessarily, from this recharging technology.

with regard to existing technologies related to smart clothing status monitoring devices: Smart wear for measuring bio-signal KR20170001032 (A): This is a wear status detection method for a smart wearable device, capable of accurately detecting the wear status of the smart wearable device. The proposed invention is connected to this invention with regard to the need to detect stress induced to the tissue induced by aggressions and violent attempts to remove it.

Reference with regard to existing technologies relating to devices for detecting biometric signals incorporated in smart clothing: Smart wear for measuring bio-signal KR20170001032 (A) This invention relates to smart clothing for measuring a biological signal. The connection with the proposed invention? in the type of operation performed, which represents a part of the signal processing necessary for the same.

with regard to existing technologies related to danger detection and alarm signaling devices incorporated in smart clothing: Portable data medium e.g. smart card, operating method for e.g. mobile telephone, involves detecting output of warning for user via terminal by data medium in case that rate of wear exceeds predetermined threshold DE102008056710 (A1) 436 / 5000 The invention concerns a method for operating a portable data medium, in particular a chip card , in a terminal. In the method according to the invention, if the degree of wear of a fabric exceeds a predetermined threshold, the data medium emits a warning message. The relationship with the found ? in the function of emitting a signal based on a determined degree of wear of the fabric: this function? crucial in the occurrence of an aggression as one of the criteria used to identify a dangerous situation? the sudden and violent wear of the fabric. Should this technology be considered as an indicator of the possibility? to perform that function, and not as an actual implementation of how that function is performed.

Reference with regard to existing technologies relating to biometric signal detection devices incorporated in smart clothing: Bio signal measurement system using woven textile sensor and smart wear having the same KR20190012823 (A) This invention refers to a biological signal measurement system which has a biosensor on a tissue and obtains biometric information of a user through the tissue sensor, and to a biosensor, a tissue electrode array and a biometric information display apparatus. The relationship with the proposed invention is identified in the intersection with the functions for detecting the biometric signals provided for it.

Reference with regard to existing technologies related to vocal sound detection devices incorporated in smart clothing: Wear speech recognition system of smart machine and wear smart machine CN205451769 (U). The connection with the proposed invention concerns the recognition function of shouts, sounds of a fight and ambush provided for the same. The invention analysed? considered as a demonstrative basis of the possibility? to perform this function.

Reference with regard to existing technologies relating to data collection devices connected to the Internet of Things: METHOD AND SYSTEM FOR AN EMERGENCY LOCATION INFORMATION SERVICE (E-LIS) FOR INTERNET OF THINGS (IoT) DEVICES US20170238129A1 The invention concerns a method and a system for determine and verify the location of network devices connected to the Internet of Things (IoT). The method and system provide a current physical geographic location for such IoT network devices and/or a user of said device in an emergency situation such as an accident, health issue, fitness, fire, terrorist attack, military incident, weather, flood event, and forward your physical geographic location to a legacy 911 network, NG-911 network, Emergency Services IP Networks (ESInet), or SMS networks (SMS) to alert emergency responders. The connection with the proposed invention ? relating to the connection function with the IoT in order to feed the flow of signals and data useful for identifying a dangerous situation.

Claims (8)

Claims 1. A system based on an intimate set consisting of a pair of panties and a bra equipped with audio, biometric, localization, tissue stress sensors and capable of interconnecting to the Internet of Things that produce an elaborate input complex by a controller regulated by a machine learning algorithm capable of processing the input configuration to determine if the user is in a dangerous situation and alert an operations center in contact with the authorities; an electronic key with button capable of manually activating the danger signal to the operations centre. The operating phases of the automatic system woven into the suit are: to. Automatic detection of sensor inputs on an ongoing basis, including: sounds from and around the user; biometric parameters such as sweating, muscle tension, heart rate; localization parameters, by interfacing with geolocation systems such as Google Maps or autonomous ones incorporated in the application; tissue stress parameters, based on the rate of change over time of the intensity? of tissue stress (to determine attempts at violent removal); other inputs from interconnection with the Internet of Things (such as video footage from a guarded parking lot); b. Processing of said inputs by a machine learning algorithm that programs a controller; based on a series of configurations that can be updated with experience, the system ? able to decide if the? user ? in danger and if it is necessary to activate the? alarm: in a first period, ? It is possible to hypothesize a training phase of the program to teach it to recognize the danger, through a continuous feedback cycle with the user, called to manually confirm to the program in correctly identifying the dangerous situations; c. Parallel to the operation of the automatic controller, availability? of manual activation of the alarm by pressing a button on an electronic key supplied together with the complete The basic operating scheme of the invention ? the following: If the button located on the electronic key (8) is pressed, the system alerts the operations centre, which warns the authorities giving the user's location to the authorities. In the presence of torsion of the fabric detected by the traction sensor (4) beyond a pre-set traction threshold, which will have to? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not; In the presence of anomalous sounds detected by the vocal sensor (1) beyond a pre-set threshold in dB, which should? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not; In the presence of altered biometric parameters detected by the biometric sensors (2) beyond a pre-set stress threshold (heart rate and perspiration), which should? be initialized in a customizable way, the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as danger configuration or not; Does the locator (3) have a list of types of place-time considered at risk, such as covered parking? after 21, configurable in a personalized way and of the necessary localization resolution to determine if the user? in these areas and at these times: the controller (5) simultaneously stores the data in question in the database (6) and transmits the data to the controller (5) which hosts the machine learning program (7) which acquires it as part of its input set to be interpreted as a danger configuration or not; The basic operating scheme of the automatic control algorithm? the following. there is an initial set of sensors given by: speech sensors (SV) from 1 to n biometric sensors (SB) from 1 to m location sensors (SL) 1 to p tissue stress sensors (SST) 1 to q functional forms and thresholds are given for each sensor, for example for vocal sensors the level in decibels of human voices including that of the user, fixed at SV1=80db; all the other specifications of the other sensors are given (e.g.: SL1, physical location with approximation of meters, SL2, type of location (covered garage, suburban road, etc.?) ? It is possible to have a danger configuration such as, for example: SV1 voice >80db SL2 covered garage Time: 02:00 AM SST: stress beyond threshold The general functional form ?: Input configuration at time t: it starts from the Speech Sensor Input 1 (ISV1) up to the Tissue Stress Sensor Input q initialized to an initial value of the time t at the start of operation ISV1 t value ? ISSTq t-value Current configuration: during operation the values change and the system records these changes continuously for any m greater than t SV1P t+m value ? SST1P t+m value Differential D-SV1 t-value - t+m value D-ST1 t-value - t+m-value If the values of one or more? sensors exceed the danger threshold defined for them, the system acts as follows: if the sensors that exceed the danger threshold are of three or more types, whatever they are, the alarm is triggered if the sensors that exceed the danger threshold are between two and three, the system sends a warning to the user, who must press the manual button on the key twice within 5 minutes to prevent activation of the alarm, otherwise the alarm will go off . if the sensors are of only one type, the automatic system does not start but sends a warning to the user who must press the manual key button twice within an hour, otherwise the alarm will go off. Every time a danger signal takes place, whether it is a false signal, deactivated by the user, or a real one, the record made up of the values detected by the sensors on that occasion ? stored in the form of a danger signal, respectively, FALSE or TRUE. The relative differentials associated with the sensors with respect to the initial configuration are also stored in the FPGA card used for storage and control. Over time, the system hones capabilities? predictive, based on continuous detection of sensor values and comparison with previous TRUE or FALSE hazard configurations. This ability? predictive allows the system to work automatically with ever greater efficiency, minimizing the level of FALSE alerts and the need for of the manual intervention of the user to report the same. 2. Method according to claim 1, reporting any type of algorithm and electronic control device capable of identifying sounds of suffocation, struggle, requests for help. 3. Method according to claim 1, reporting any type of algorithm and electronic control device able to identify alterations of heart rate, perspiration, muscle tone and other typical parameters subject to alteration in case of physical aggression and fight. 4. Method according to claim 1, reporting any type of algorithm and electronic control device capable of locating the position of an individual and providing qualitative information on the place under examination (road, parking, hotel, etc.) The method according to claim 1, setting out any type of algorithm and electronic control device capable of identifying mechanical stress inflicted on the garments by traction, twisting, and any movement which may indicate attempts at violent removal 6. Method according to claim 1, reporting any type of algorithm and electronic control device capable of integrating the system set forth from 1 to 6 with the Internet of Things. 7. Method according to claim 1, reporting any type of algorithm and electronic control device capable of making an intelligent coding of the danger situations as set out in point 1. 8. Method according to claim 1, reporting any type of algorithm and electronic control device capable of replacing or integrating the electronic key with button for manual activation.