WO2020106212A1

WO2020106212A1 - Intelligent impact sensor and uses

Info

Publication number: WO2020106212A1
Application number: PCT/SG2018/050574
Authority: WO
Inventors: Siddharth Mazumdar
Original assignee: Optimax Management Services Pte. Ltd.
Priority date: 2018-11-22
Filing date: 2018-11-22
Publication date: 2020-05-28
Also published as: SG11202107912PA

Abstract

The present disclosure describes a wearable impact sensor (100) and uses as a personal fall detector, personal alarm, as a monitor for one's driving or riding behavior, as a monitor for glass safety; a monitor for person in need of life support, and so on. The impact sensor works as a system with an app (210) in a smartphone (200) using an artificial intelligence algorithm (300) for impact determination. Simulated impact tests data are stored in the impact sensor as propositional logic data, whilst actual use/user motion profile data defining normal impact signatures are stored in the smartphone as first order logic data. When real time impact signals are logically compared in the AI algorithm, a positive determination would result in sending SOS alerts together with GPS information to all the first responders. A user can select from a choice of three levels of sensor accuracy.

Description

Intelligent Impact Sensor and Uses

Field of Invention

[001] The present invention relates to an intelligent portable sensor configured to detect and differentiate various types of impacts or jerks using user’s motion profiles or simulated test profiles. The uses ranges from personal fall detection, personal alarm, capturing driving/riding behavior, glass window safety, motion monitor for patients or military personnel, and so on.

Background

[002] Portal devices such as smartphones with applications for sports or health fitness are known. Also known are wearable devices that monitors a user’s pulse rate, blood pressure, walking distance and calorie balance, and so on. Also known are portable personal security alarms. The limitations of some of these wearable devices is battery life, integration of separate types of electronic sensors such as accelerometers, gyroscope, magnetometer, and so on. These known electronic sensors use preset or static threshold magnitudes of sensor signals to differentiate between sensed conditions, instead of determining from a user’s characteristic movements or behavior patterns, or from those established from simulated profiles obtained in a calibration laboratory.

[003] It can thus be seen that there exists a need to provide an intelligent impact sensor that can overcome the limitations of known devices. Preferably, the intelligent impact sensor can be put to various uses, such as, being worn by a user to detect a fall or accident, to monitor a driver/rider behavior pattern and linking such behavior pattern to road safety and insurance, to detect glass window safety, motion monitor for patients or military personnel, and so on. Preferably, the sensed conditions are compared to the user’s movement profiles and/or simulated calibration profiles, instead of static threshold values. Summary

[004] The following presents a simplified summary to provide a basic understanding of the present invention. This summary is not an extensive overview of the invention, and is not intended to identify key features of the invention. Rather, it is to present some of the inventive concepts of this invention in a generalised form as a prelude to the detailed description that is to follow.

[005] The present invention seeks to provide an intelligent impact sensor that is portable or wearable. Preferably, the impact sensor is configured to capture a user’s motion profile or driving/riding habit, or to detect glass window safety, and so on. In use, the impact sensor detects instances of unusual impact or jerks on the basis of user’s motion profiles and/or simulated calibration profiles; an unusual impact or jerk may be life threatening and may trigger some automated SOS alerts. Preferably, the impact sensor may link with an app installed in a smartphone or may link with scanners disposed in a local wide area network (LAN) for central monitoring.

[006] In one embodiment, the present invention provides a method for determining impact experienced by an impact sensor. The method comprises: obtaining simulated impact test data, storing the simulated impact test data in the impact sensor and using the simulated impact test data for propositional logic determination in an artificial intelligence (AI) algorithm; obtaining actual use/user motion profile data during initial predetermined hours of usage and using the actual use/user motion data for a first order logic determination in the AI algorithm; capturing real time user data from the impact sensor and logically comparing the real time user data to both the simulated impact test data and actual use/user motion data according to the respective propositional and first order logic determinations; and if the logic determination is positive, issuing SOS alerts to all predefined first responders.

[007] Preferably, the method comprises installing an app in a smartphone and pairing the app with the impact sensor via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology, and acquiring GPS information from the smartphone and sending the SOS alerts with the GPS information. Alternatively, the method comprises installing a local area network (LAN) in an area or building; and installing a scanner in each designated zone in the area or each level in the building, so that each impact sensor present in the area or building is traceable at a central monitor station. Preferably, the scanner is operable via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology.

[008] In another embodiment, the present invention provides a system for sensing impact and differentiating abnormal impact; the system comprises: a wearable impact sensor device comprising, at least, a low-g accelerometer, a high-g accelerometer, a gyroscope, a microprocessor, a communication module and inbuilt antenna, and a memory unit; wherein the memory unit contains simulated impact tests data which constitute propositional logic for artificial intelligence (AI) impact determination; a smartphone with an app for pairing with the communication module, with the smartphone containing actual use/user motion profile data, which constitute first order logic for AI impact determination; and an AI algorithm operable in the smartphone, wherein real time sensor signals are logically compared with both the propositional logic and the first order logic data, and if a logical determination is positive, predefined SOS alerts together with GPS information from the smartphone are sent out to all the first responders.

[009] Preferably, the impact sensor device is used as a personal fall detector; a panic alarm; a sensor for monitoring one’s driving or riding motion; a sensor for monitoring glass window safety; a sensor for monitoring patient or military person in need of critical life support, and so on.

[0010] In yet another embodiment, the present invention provides a system for monitoring movements of a person wearing an impact sensor in an area or building equipped with a local area network (LAN); the system comprises: a wearable impact sensor device comprising, at least, a low-g accelerometer, a high-g accelerometer, a gyroscope, a microprocessor, a communication module and inbuilt antenna, and a memory unit; and a scanner located in a designated zone in the area or floor of the building for connecting with each communication module that is within communication range, so that movements of the person is traceable on a central monitor station. Preferably, communication is via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology. Brief Description of the Drawings

[0011] This invention will be described by way of non-limiting embodiments of the present invention, with reference to the accompanying drawings, in which:

[0012] FIG. 1 illustrates components of an intelligent impact sensor according to an embodiment of the present invention;

[0013] FIG. 2A illustrates operation of the intelligent impact sensor in a cellular network, whilst FIG. 2B illustrates another operation within a local area network; and

[0014] FIG. 3 illustrates an artificial intelligence logic tree impact or jerk determination in an algorithm operable for impact/jerk sensing.

Detailed Description

[0015] One or more specific and alternative embodiments of the present invention will now be described with reference to the attached drawings. It shall be apparent to one skilled in the art, however, that this invention may be practised without such specific details. Some of the details may not be described at length so as not to obscure the present invention.

[0016] The present invention involves impact sensing and impact determination; very often, impacts also involve jerks, sudden twists or sudden turns; some impact, jerks, sudden twist or sudden turn experienced by a person may precede a fall, an assault, a hit-and-run accident, and so on. In the unfortunate even that a victim is injured or becomes unconscious, an impact sensor equipped with SOS alerting and worn on the body can help save a life. Such impact sensing application can be extended to promote safe driving/riding behaviour, glass window security, motion monitor in hospitals, and so on.

[0017] FIG. 1 shows a schematic of an intelligent impact sensor 100 according to an embodiment of the present invention. As shown in FIG. 1, the impact sensor 100 includes a microprocessor 110. Connected to the microprocessor 110 are: a memory unit 120, a communication module 130 and built-in antenna 135, a low-g accelerometer 140, a high-g accelerometer 150, a gyroscope 160, a battery 170 and a panic button 190. The memory unit 120 may be partitioned to store ROM data, such as, impact sensor identity, initialization data, such as, user identity, and a RAM for storing pre-defmed“pitch and roll” data, which are characteristic signatures obtained by simulated or calibrated impact tests in a laboratory; these simulated calibration profdes constitute a default propositional logic in an artificial intelligence (AI) algorithm 300 for impact determination, which will be described with FIG. 3. Preferably, the date and time information are obtained from the microprocessor 110. Preferably, the impact sensor 100 also has a port to recharge the battery 170 and/or initialise/export data from the memory unit 120. A default mode of use relies on the AI impact sensing, whilst the panic button 190 provides a manual trigger for sending SOS messages to all pre-defmed first responders (by-passing AI impact sensing mode). A hook 195 feature is provided on a body of the impact sensor 100 as a point for connecting the impact sensor to a lanyard or to an attachment ring.

[0018] In one embodiment, the communication module uses Bluetooth or narrow band IoT (NBIoT) technology, which has low power consumption and is suitable for this battery operated impact sensor 100.

[0019] After a user is issued with the above impact sensor 100, assuming the battery 170 has sufficient electric power, is initialized with the user identity and an app 210 in installed in an associated smartphone 200, the low-g accelerometer 140, the high-g accelerometer 150 and gyroscope 160 capture respective sensor signals in real time and send the real time acceleration signals (ie. a_x, a _y and a_z respectively in the X-, Y- and Z-axes) and rate of change of angle signals (ie. Ox, 0y and 0z respectively about the X-, Y- and Z-axes) to the microprocessor 110. Preferably, each sensor real time signal is filtered to remove electric noises. Depending on the application of this impact sensor, the user or usage predefined motion profiles pertain to initial motions captured by the impact sensor 100 worn on a part of the body, mounted on a dashboard of a vehicle, mounted on a glass window, and so on. Preferably, the user predefined motion profile covers an initial 150 hours of usage and constitutes a first order logic in the AI algorithm 300 for impact determination. Preferably, the smartphone 200 is paired with the impact sensor 100 and the impact sensor sends sensor data to the smartphone. A display screen on the smartphone shows the user identity, sensor identity, a current history of impact sensing, a choice of 3 levels of sensor accuracy (namely, normal, normal+ and normal++), settings for first responder contacts, and so on. If the smartphone 200 is linked to a cellular network 220, as seen in FIG. 2A, the user usage history may be transmitted to a server 230 and stored in a database 240. When SOS alerts are triggered and sent to the first responders, the SOS alerts are tagged with GPS information from the smartphone. In another embodiment, it is possible that the smartphone 200 is connected to an internet 220a, a server 230a and a database 240a belonging to an insurance institution or a security service provider, and/or is linked to a national emergency hotline. The information sent to a database of an insurance institution contains useful big data, which may be used for actuarial research, for eg. to develop motor insurance policies; real time analytics from this big data may be used to enhance road safety, develop road safety programs, help analyse road safety crash tests, etc.

[0020] Preferably, the low-g accelerometer 140 is selected to sense an acceleration value of up to substantially 6g, whilst the high-g accelerometer 150 may sense an acceleration value of substantially lOOg or more. With this combination, the impact sensor 100 becomes highly sensitive and can be put to a wide range of uses. The gyroscope 160 may be selected with 6 degrees-of-freedom (DoF) for 2D motion sensing or 9 DoF for 3D motion sensing. With a combination of the low-g accelerometer 140, high-g accelerometer 150 and gyroscope 160, the impact sensor 100 becomes highly sensitive for detecting impacts and sudden jerks or turns when worn by a user or attached onto a vehicle/motorcycle or glass window.

[0021] When there is no cellular or satellite communication, it is possible that a local area network (LAN) is provided, as shown in FIG. 2B. In an eg., on an oil drilling platform, cruise ship, a factory, etc, a LAN 250 may be provided, where each work zone or floor level is served with a designated Bluetooth (BT) scanner or a plurality of designated BT scanners. Each impact sensor 100 worn by a person is located within range of a neighbouring BT scanner and, as the person moves around, presence or location of the person can be traced by identifying the designated zone or floor level on a central monitor station 260. This application may also be useful in a security area, such as, a prison where the occupants are located by BT tracking which is enabled by the impact sensor 100. Instead of using BT, it is possible to use narrow band IoT or other short-range communication technology. [0022] FIG. 3 shows a logic tree structure employed in the artificial intelligence (AI) algorithm 300 for impact determination. In FIG. 3, the propositional logic data (containing predefined pitch and roll data obtained by simulation or calibrated impact tests conducted in a laboratory and stored in the associated impact sensor 100) is shown in function block 310, whilst initial user or usage profile captured during the first 150 hours of use and constituting the first order logic data is stored in the smartphone 200. The first order logic data are shown in function block 320. The impact sensor 100 may be configured for wearing by a person, mounted on a vehicle for capturing a driving/riding behavior, mounted on a glass window, as a device for tracking movement in a monitored environment, and so on.

[0023] Personal fall sensing: The impact sensor 100 may be worn on a wrist, hung on a lanyard, pinned onto an apparel of the user, disposed in a pocket of the user, disposed in a carrying pouch, and so on. Before putting the impact sensor into use, the user may choose to select one level of sensor accuracy (namely, normal, normal+ or normal++) and connect the impact sensor with the app 210 by BT pairing; as the user starts to move about (as shown in block 330), the impact sensor 100 becomes activated. In an event of a fall, the low-g accelerometer 140 or high-g accelerometer 150 may be triggered by a sharp acceleration. It is also possible that the gyroscope 160 may also be triggered by a sudden twist or turn about any one axis. Real time accelerometer and gyroscope signals in the X-, Y- and Z-axes (defined as second order logic data) are captured in function block 340 and then compared logically in block 350 with both the propositional logic data and first order logic data. A logical decision is then made in block 355 whether the fall involves an impact and/or jerk. If the logical decision in block 355 is negative, the display screen on the smartphone would indicate the sensor data falling with a normal range of movement, preferably, in a graphic mode or displayed in colour code in block 360. Whilst the AI algorithm is shown to end in block 370, in this application, it is preferred that the algorithm 300 repeats continuously in cyclic loops or until the smartphone connection with the impact sensor is terminated by the user. If the logical decision in block 355 is positive, real time impact signals from the impact sensor are determined to be abnormal and, as a result, SOS alerts are generated in block 380 and sent to all the predefined first responders. [0024] Sensing driver/rider behavior patern: Preferably, an impact sensor 100 is mounted on a dashboard of a vehicle or near a meter board of a motorcycle. As in the above description, the user begins monitoring a driving/riding journey by selecting a level of sensor accuracy and connecting the impact sensor 100 with the app 210 via BT pairing. In use, the low-g accelerometer 140, high-g accelerometer 150 and gyroscope 160 capture real time sensor signals in function blocks 340. In block 350, the AI algorithm 300 compares the real time accelerometer and gyroscope signals with both the propositional logic data and first order logic data, and a logical decision is made in block 355. If an unusual impact is sensed and/or an abnormal twist or turn is detected and resulting in a crash, the AI algorithm 300 process would proceed to block 380 and SOS alerts would be sent out from the smartphone with GPS information to all the predefined first responders. When such unusual impact or abnormal twist or turn are detected, it is possible that such incidents be captured in the app 210. When no unusual impact and twists/tums are detected, the AI algorithm continues to monitor the driver/rider behavior until the journey has ended or the BT pairing is disconnected. In another embodiment, it is possible that the impact sensor 100 is worn on a body or an apparel of a driver/rider.

[0025] Glass window safety monitor: The impact sensor 100 is mounted on a glass panel of a retail store as a safety monitor. For eg., in an event of vandalism or an atempted break-in, vibrations on the glass panel would be picked up by the low-g accelerometer 140, or even by the high-g accelerometer 150. The propositional logic data may contain vibration signatures of a diamond tip cuting a glass panel, whilst the first order logic data contain vibration signatures that the glass panel is subjected to after installation, for eg., as typically caused by shoppers or vehicle traffic in the vicinity. In the event of vandalism or an atempted break-in, any vibrations that are unusual from those stored in the propositional logic and/or first order logic data would trigger a positive detection and SOS alert would be sent to the first responder and/or an associated security provider.

[0026] The impact sensor of the present invention can also be put to other uses. For eg., if the impact sensors 100 are tracked at the monitor centre 260, say in, a hospital or a military institution, movement of patients/trainees wearing these impact sensors may help to identify cases requiring critical life support. [0027] There are many advantages when using the impact sensor 100 of the present invention. Each impact sensor 100 is compact, light and portable, and can thus be worn on a wrist, hung on a lanyard, pinned onto an apparel of the user, disposed in a pocket of the user, disposed in a carrying pouch, attached to a surface on a vehicle, motorcycle, glass window, and so on. The impact sensor is powered by a small battery and with use of BT, it has low power consumption. Power consumption is even lower when pairing with the app 210 in a smartphone 200 is only enabled when the impact sensor is in use. Further, each impact sensor employs a combination of a low-g accelerometer 140, a high-g accelerometer 150 and a gyroscope 160, which devices are made up of electronic chips-on-boards, thus making the impact sensor 100 highly sensitive yet small in form factor.

[0028] The impact sensor is differentiated from known devices by the use of artificial intelligence (AI) algorithm 300 impact determination. Instead of relying on sensor threshold magnitudes, pitch and roll data obtained by simulation or calibrated impact tests conducted in a laboratory constitute the propositional logic data; in addition, actual use or user characteristic motion profiles that contain sensor signatures that are detected during an initial use period (preferably, of 150 hours) constitute the first order logic data; these propositional and first order logic data are used to determine whether an impact, jerk, sudden twist or sudden turn experienced by the low-g accelerometer 140, high-g accelerometer 150 or gyroscope 160 is logically usual or abnormal. If an impact is abnormal, SOS alerts would be generated and sent to the first responders; by basing on calibration data, actual use/user motion data and AI logic, impact determination is precise and accurate, making false positives a rare occurrence.

[0029] With use of the impact sensor during driving or riding, a driver/rider becomes aware of one’s driving/riding habits and this would lead to improve driving/riding behavior and general road safety. It is also possible that incentives, such as, lower insurance premiums would encourage drivers/riders to change some undesirable driving/riding habits.

[0030] While specific embodiments have been described and illustrated, it is understood that many changes, modifications, variations and combinations thereof could be made to the present invention without departing from the scope of the present invention. For example, a mounting bracket/holder (not shown in the figures) may be provided to give a snug support for an impact sensor; the mounting bracket/holder may be attachable onto a glass window or a dashboard of a vehicle or motorcycle, so that attenuation of the real time sensors data are substantially matching the attenuation when the pitch and roll calibration data are obtained with the impact sensor being fitted in the mounting bracket/holder.

Claims

CLAIMS:

1. A method for determining impact experienced by an impact sensor, the method comprises:

obtaining simulated impact test data, storing the simulated impact test data in the impact sensor and using the simulated impact test data for propositional logic

determination in an artificial intelligence (AI) algorithm;

obtaining actual use/user motion profile data during initial predetermined hours of usage and using the actual use/user motion data for a first order logic determination in the AI algorithm;

capturing real time user data from the impact sensor and logically comparing the real time user data to both the simulated impact test data and actual use/user motion data according to the respective propositional and first order logic determinations; and

if the logic determination is positive, issuing SOS alerts to all predefined first responders.

2. The method according to claim 1, further comprises:

installing an app in a smartphone and pairing the app with the impact sensor via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology.

3. The method according to item 2, further comprises:

acquiring GPS information from the smartphone and sending the SOS alerts with the GPS information.

4. The method according to claim 1, further comprises:

installing a local area network (LAN) in an area or building; and

installing a scanner in each designated zone in the area or each level in the building, so that each impact sensor present in the area or building is traceable at a central monitor station.

5. The method according to claim 4, wherein the scanner is operable via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology.

6. A system for sensing impact and differentiating abnormal impact comprises:

a wearable or portable impact sensor device comprising, at least, a low-g accelerometer, a high-g accelerometer, a gyroscope, a microprocessor, a communication module and inbuilt antenna, and a memory unit; wherein the memory unit contains simulated impact tests data which constitute propositional logic for artificial intelligence (AI) impact determination;

a smartphone with an app for pairing with the communication module, with the smartphone containing actual use/user motion profile data, which constitute first order logic for AI impact determination; and

an AI algorithm operable in the smartphone, wherein real time sensor signals are logically compared with both the propositional logic and the first order logic data, and if a logical determination is positive, predefined SOS alerts together with GPS information from the smartphone are sent out to all the first responders.

7. The system according to claim 6, wherein the impact sensor device is used as a personal fall detector; a panic alarm; a sensor for monitoring one’s driving or riding motion; a sensor for monitoring glass window safety; a sensor for monitoring patient or military person in need of critical life support, and so on.

8. The system according to claim 7, wherein, when the impact sensor is used for monitoring one’s driving or riding motion, the smartphone is linked to a server of an associated insurance institution, which collects big data useful for actuarial research, to enhance road safety and to better analyse road safety crash tests.

9. The system according to claim 7, wherein, when the impact sensor is used for monitoring glass window safety, the impact sensor is linked to a security provider.

10. The system according to claim 7, wherein, when the impact sensor is not worn on a user, the impact sensor is supported or held in a mounting bracket/holder.

11. A system for monitoring movements of a person wearing an impact sensor in an area or building equipped with a local area network (LAN), the system comprises: a wearable impact sensor device comprising, at least, a low-g accelerometer, a high- g accelerometer, a gyroscope, a microprocessor, a communication module and inbuilt antenna, and a memory unit; and

a scanner located in a designated zone in the area or floor of the building for connecting with the communication module that is within communication range, so that movements of the person is traceable on a central monitor station.

12. The system according to claim 11, wherein communication with the communication module is via Bluetooth, narrow band IoT (NBIoT) or similar wireless technology.