CN105632101A - Human body anti-tumbling early warning method and system - Google Patents

Abstract

The invention provides a human body anti-tumbling early warning method and a system. The early warning method comprises steps of acquiring triaxial acceleration information, triaxial angular velocity information and triaxial geomagnetic field information of to-be-monitored objects of multiple sampling points in real time; calculating combined accelerations at all sampling points according to the triaxial acceleration information; generating Euler angle difference value sequences according to the triaxial acceleration information, the triaxial angular velocity information and the triaxial geomagnetic field information of all the sampling points; judging whether moving states of the to-be-monitored objects are suspected tumbling states according to the combined acceleration and preset combined acceleration threshold values of all the sampling points; if yes, judging whether the to-be-monitored objects have tumbling tendencies according to the combined acceleration average values obtained via the combined acceleration of all sampling points and the preset acceleration average value threshold values by combining the Euler angle difference value sequences and preset Euler angle difference value threshold values; and when the to-be-monitored object has the tumbling tendencies, generating early warning information so as to carrying out early warning before the to-be-monitored objects tumble.

Description

A kind of human body anti-fall method for early warning and system

Technical field

The present invention relates to wearable technology field, particularly relate to a kind of human body anti-fall method for early warning and system.

Background technology

Fall and refer to that burst, Body Position Change involuntary, unintentional fall on the ground or in less plane. According to statistics, there are about 424000 examples dead directly related with the behavior of falling in the world every year, the second largest reason (being only second to vehicle accident) being to cause Accidents death in the world of falling. Research shows, in the old people of over-65s, has the ratio fallen 1 time or repeatedly fall experience every year up to 1/3, and wherein the old man of 20%��30% can cause scratch, Hip Fracture, injury of head etc. in the event of falling, and increases along with the increase at age. In the U.S., the annual medical total cost for falling is more than 20,000,000,000 dollars. China has the old people of about 1.3 hundred million at present, there are about 20,000,000 old peoples every year and 25,000,000 accidents of falling altogether at least occur, and direct medical cost is more than 5,000,000,000 RMB. As can be seen here, fall become threaten old people's life security and increase burden on society key factor.

At present judge to fall mainly have following 3 kinds of methods: 1) based on the analysis of video image, the real time kinematics of object is monitored by photographic head, and its deficiency is it cannot be guaranteed that the personal secrets of user, and uses scene limited, cost intensive; 2) based on the analysis of sound signal, being caused that the frequency-portions of vibration judges to fall event by analyzing to impact, but this kind equipment is installed more complicated, fund input is also relatively larger; 3) detect based on the device of Wearable. Considering the right of privacy of user and reduce interference user's life style as far as possible, the device of Wearable is best suited for. Fall detection/the monitoring of Wearable and alarm device, be typically all based on inertia sensing unit. Sensing module common " fall alarm " at present is generally adopted single three axis accelerometer, certainty of measurement and data modality are limited, or the thresholding algorithm adopted is excellent not, maximum haveing the drawback that can only realize warning of falling, namely warning after falling, seek help, and can not accomplish that anticipation drops to risk and reminds user.

Chinese invention patent 201210472813.1 discloses a kind of fall detection automatic alarm belt, including belt and detecting device, its detecting device adopts acceleration detection module, by detecting the 3-axis acceleration information of user, (module adopted is Freescale MMA7455 3-axis acceleration sensor, processor is single-chip microcomputer) to judge whether user is in the state of falling, customer position information can be obtained automatically by GPS module and seek advice note to guardian's mobile phone by gsm module transmission, it is simple to after human body is fallen, carry out timely treatment.

Chinese invention patent 201410134724.5 discloses a kind of fall monitoring warning system, including earphone portion and mobile communication terminal portion, earphone portion can continue to obtain the information such as the acceleration information of human body, body angle data, mobile communication terminal portion receives earphone portion data, judge whether to meet condition of falling, it is judged that then first can send, to server end, fall warning and location information for the state of falling. This mode communication terminal design comparison is flexible, but earphone end needs to be worn on ear position for a long time, easily causes human body discomfort, and real-time data collection power consumption is high, and earphone volume is little, and battery capacity is low, is unsuitable for wearing for a long time use.

Utility model patent 201520186743.2 discloses one and falls alarm waistcoat, including emergency alarm switch, acceleration sensor, signal projector and a few part of locating module, falling when detecting, can pass through signal projector and send relief information to signal receiver, request is succoured. This vest is dressed convenient, has the functions such as communication and location, but is a lack of preventing mechanism and the warning function of necessity.

Above-mentioned technology has the disadvantage that 1) unrealized fall before early warning. Sensing module common " fall alarm " at present is generally adopted single three axis accelerometer, certainty of measurement and data modality are limited, or the thresholding algorithm adopted is excellent not, therefore it is only capable of realizing the warning function after human body is fallen and the early warning before falling can not being realized. Although having related to " early warning " at Chinese invention patent 201210561342.1, but its implementation is based on video image analysis, as previously described, it is unfavorable for ensureing privacy of user, and only mainly for house Monitoring Data, cause its application scenarios significantly limited, thus unlikely and then realize instant anti-intervening measure and the real-time protection of falling.

2) function singleness, interface are simple: published precaution device only realizes simple early warning function mostly, and precaution device is as user's thing worn next to the skin, are actually a critically important data acquisition platform. If configuring certain interface and hardware device, the function that network is connected with the guardian of user, with the mobile devices interconnect of oneself, even can also can also realize social activity namely can be passed through.

3) use is worn convenient not: a part of tumble alarm device have employed the multimode compound modes such as accelerometer module, attitude transducer module and pressure sensor module, although being conducive in theory improving the accuracy judged of falling, but inevitable volume and the power consumption that also can increase device of multiple sensing modules, have impact on the convenience worn and use. It addition, microprocessor and thresholding algorithm are had higher requirement by the Data Fusion of multiple sensing modules in precision especially speed. It addition, sensor assembly is closer to human body in theory, the sensing data recorded is more accurate, and in actual applications, can the combination of multimode really be effectively improved the accuracy judged of falling and need further checking.

4) accuracy of falling judges undesirable: mainly only rely on acceleration judges the generation of the event of falling to current judgment means of falling mostly, but the motion of human body schedule is extremely complex, and the accident behavior (1��2 second) usually occurred in a flash of falling, relying on less or single human motion attitude information to predict the reliability fallen, not high (analysis of multisensor block combiner mode is shown in upper one, repeat no more) herein, when human body fall tendency and when failing to judge in time, just do not have early warning effect, when human body do not fall tendency and when making erroneous judgement, necessarily make troubles to wearer.

5) preventer is portable not, dress inconvenience: disclosed in utility model patent 201220626477.7, protective garment can provide large-area protection, but dress particularly troublesome, be unfavorable for that the handicapped people such as old people dresses, and there is no corresponding fall detection algorithm, simply simple physical protection; Remotely falling disclosed in Chinese invention patents 201410141280.8,201510038329.1 etc. and guard Intelligent crutch, anti-seat etc. of falling, can only partly realize falling safeguard function, uses scene limited, it is impossible to spread to whole old people colony on a large scale.

Summary of the invention

For solving above-mentioned technical problem, the invention provides a kind of human body anti-fall method for early warning and system.

One aspect of the present invention provides the anti-method for early warning of falling of a kind of human body, and the anti-method for early warning of falling of described human body includes:

The 3-axis acceleration information of the monitored target of the multiple sampled point of Real-time Collection, three axis angular rate information and three axle geomagnetic field information;

The monitored target resultant acceleration at each sampled point is calculated according to described 3-axis acceleration information;

Eulerian angles sequence of differences is generated according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information;

Whether resultant acceleration and the kinestate of default resultant acceleration threshold decision monitored target according to each sampled point are doubtful state of falling;

If it is, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold;

Warning message is generated, to report to the police before monitored target is fallen when monitored target falls tendency.

In one embodiment, the anti-method for early warning of falling of described human body also includes: air bag actuator punctures gas cylinder therein so that air bag therein to be inflated according to described warning message, to provide buffering when human body is fallen.

In one embodiment, the anti-method for early warning of falling of described human body also includes: start timing after generating described warning message, judge whether described monitored target cancels described warning message within the preset alarm time, if it is not, then described warning message is sent to guardian after the meter full described preset alarm time.

In one embodiment, the anti-method for early warning of falling of described human body also includes: the described 3-axis acceleration information of each sampled point collected, three axis angular rate information and three axle geomagnetic field information are carried out Kalman filtering.

In one embodiment, calculate the monitored target resultant acceleration at each sampled point according to described 3-axis acceleration information, including:

According to described 3-axis acceleration information, acceleration modulus algorithm is utilized to calculate the monitored target resultant acceleration A at each sampled point_SVM:

$A_{SVM}=\sqrt{A_x^2+A_y^2+A_z^2}$

Wherein, A_x��A_y��A_zRespectively monitored target is at three axial acceleration of same sample point.

In one embodiment, Eulerian angles sequence of differences is generated according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information:

Described 3-axis acceleration information, three axis angular rate information and three axle geomagnetic field information are fused into Eulerian angles, and described Eulerian angles include course angle, the angle of pitch and roll angle;

Described course angle, the angle of pitch and roll angle are carried out respectively difference, obtain comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference:

$\left\{\begin{array}{c}{d}_{roll}=roll\left(k\right)-roll(k-1)\\ d_{pitch}=pitch\left(k\right)-pitch(k-1)\\ d_{yaw}=yaw\left(k\right)-yaw(k-1)\end{array}\right.;$

Wherein, yaw is course angle, and pitch is the angle of pitch, and roll is roll angle, and k is current sample time, and k-1 is a upper sampling instant.

In one embodiment, according to whether the kinestate of the resultant acceleration of each sampled point and default resultant acceleration threshold decision monitored target is doubtful state of falling, including:

When the resultant acceleration of at least one sampled point is more than described resultant acceleration threshold value, it is determined that the kinestate of described monitored target is doubtful state of falling.

In one embodiment, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold, including:

Described resultant acceleration according to each sampled point, average is accelerated in the conjunction calculating monitored target;

When described resultant acceleration average is more than default resultant acceleration average threshold value, and the absolute value of at least one described Eulerian angles difference more than described Eulerian angles difference threshold time, it is determined that described monitored target has tendency of falling.

Another aspect of the present invention additionally provides the anti-early warning system of falling of a kind of human body, and the anti-early warning system of falling of described human body includes:

Sensor unit, for the 3-axis acceleration information of monitored target of the multiple sampled point of Real-time Collection, three axis angular rate information and three axle geomagnetic field information;

Resultant acceleration computing unit, for calculating the monitored target resultant acceleration at each sampled point according to described 3-axis acceleration information;

Eulerian angles sequence of differences generates unit, for generating Eulerian angles sequence of differences according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information;

First judging unit, for according to whether the resultant acceleration of each sampled point and the kinestate of default resultant acceleration threshold decision monitored target are doubtful state of falling;

Second judging unit, for judge at described first judging unit the kinestate of monitored target be doubtful fall state time, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold;

Warning message generates unit, for generating warning message when monitored target falls tendency, to report to the police before monitored target is fallen.

In one embodiment, the anti-early warning system of falling of described human body also includes:

Timing unit, for starting timing after generating described warning message;

3rd judging unit, is used for judging whether described monitored target cancels described warning message within the preset alarm time.

In one embodiment, the anti-early warning system of falling of described human body also includes: positioning unit, for gathering the geographic position data of described monitored target.

In one embodiment, the anti-system of falling of described human body also includes: communication unit, and for described warning message being sent to guardian after the meter full described preset alarm time, described warning message includes described geographic position data.

In one embodiment, the anti-early warning system of falling of described human body also includes:

Kalman filtering unit, for carrying out Kalman filtering to the described 3-axis acceleration information of each sampled point collected, three axis angular rate information and three axle geomagnetic field information.

In one embodiment, described resultant acceleration computing unit specifically for:

According to described 3-axis acceleration information, acceleration modulus algorithm is utilized to calculate the monitored target resultant acceleration A at each sampled point_SVM:

$A_{SVM}=\sqrt{A_x^2+A_y^2+A_z^2}$

Wherein, A_x��A_y��A_zRespectively monitored target is at three axial acceleration of same sample point.

In one embodiment, described Eulerian angles sequence of differences generation unit includes:

Digital moving processor, for described 3-axis acceleration information, three axis angular rate information and three axle geomagnetic field information are fused into Eulerian angles, described Eulerian angles include course angle, the angle of pitch and roll angle;

Difference calculating module, for described course angle, the angle of pitch and roll angle are carried out difference respectively, obtains comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference:

$\left\{\begin{array}{c}{d}_{roll}=roll\left(k\right)-roll(k-1)\\ d_{pitch}=pitch\left(k\right)-pitch(k-1)\\ d_{yaw}=yaw\left(k\right)-yaw(k-1)\end{array}\right.;$

Wherein, yaw is course angle, and pitch is the angle of pitch, and roll is roll angle, and k is current sample time, and k-1 is a upper sampling instant.

In one embodiment, described first judging unit specifically for:

When the resultant acceleration of at least one sampled point is more than described resultant acceleration threshold value, it is determined that the kinestate of described monitored target is doubtful state of falling.

In one embodiment, described second judging unit includes:

Resultant acceleration mean value computation module, for the described resultant acceleration according to each sampled point, average is accelerated in the conjunction calculating monitored target;

Judge submodule, for when described resultant acceleration average is more than default resultant acceleration average threshold value, and the absolute value of at least one described Eulerian angles difference more than described Eulerian angles difference threshold time, it is determined that described monitored target has tendency of falling.

In one embodiment, the anti-early warning system of falling of described human body also includes an air bag actuator, described air bag actuator is generated unit with described warning message and is connected by steering wheel interface, for puncturing gas cylinder therein so that air bag therein to be inflated according to described warning message, thus provide buffering when human body is fallen.

In one embodiment, the anti-early warning system of falling of described human body also includes: alarm unit, including voice alarm module, buzzer, vibrator and LED, for sending sound and light alarm according to described warning message.

In one embodiment, the anti-early warning system of falling of described human body also includes: memory element, generates, for storing the intermediate data generated in the data of described sensor unit collection, described information process and described warning message, the warning message that unit generates.

In one embodiment, the anti-early warning system of falling of described human body also includes: power subsystem, for the anti-early warning system of falling of described human body is powered.

Utilizing the present invention, it is possible to carry out early warning before human body is fallen, remind monitored target and people about, monitored target has fall risk, should be noted strick precaution.

Accompanying drawing explanation

In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, the accompanying drawing used required in embodiment or description of the prior art will be briefly described below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the premise not paying creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.

Fig. 1 is that embodiment of the present invention human body is prevented falling the schematic flow sheet of method for early warning;

Fig. 2 is the resultant acceleration change curve before and after embodiment of the present invention monitored target is fallen;

Fig. 3 is that monitored target is sat down the resultant acceleration change curve before and after action;

Fig. 4 is that monitored target is hurried up/jog the curve of the resultant acceleration change before and after action;

Fig. 5 is that embodiment of the present invention human body is prevented falling the structural representation of early warning system;

Fig. 6 is that embodiment of the present invention human body is prevented falling the frame structure of early warning system and connection diagram;

Fig. 7 is embodiment of the present invention alarm unit hardware composition frame chart;

Fig. 8 is the anti-early warning system workflow schematic diagram of falling of embodiment of the present invention human body;

Fig. 9 is that embodiment of the present invention human body prevent the falling monitoring pattern of early warning system converts schematic diagram;

Figure 10 is embodiment of the present invention gesture recognition algorithms schematic flow sheet;

Figure 11 is that the embodiment of the present invention is fallen pre-warning time result schematic diagram.

Detailed description of the invention

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments. Based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain under not making creative work premise, broadly fall into the scope of protection of the invention.

Fig. 1 is that embodiment of the present invention human body is prevented falling the schematic flow sheet of method for early warning. As it is shown in figure 1, this method for early warning mainly comprises the steps that

The 3-axis acceleration information of the monitored target of the multiple sampled point of step S1, Real-time Collection, three axis angular rate information and three axle geomagnetic field information.

Step S2, calculate monitored target according to described 3-axis acceleration information in the resultant acceleration of each sampled point.

Step S3, according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information generate Eulerian angles sequence of differences.

Step S4, according to whether the resultant acceleration of each sampled point and the kinestate of default resultant acceleration threshold decision monitored target are doubtful state of falling.

Step S5, when the kinestate of monitored target be doubtful fall state time, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold.

Step S6, monitored target fall tendency time generate warning message, to report to the police before monitored target is fallen.

Utilizing the present invention, it is possible to carry out early warning before human body is fallen, remind monitored target and people about, monitored target has fall risk, should be noted strick precaution.

If according to step S4 judges that the kinestate of monitored target is non-doubtful state of falling, namely return step S1 and continue to gather the exercise data of monitored target.

In one embodiment, after generating warning message in step s 6, namely when monitored target falls tendency, air bag actuator punctures gas cylinder therein so that air bag therein to be inflated according to this warning message, to provide buffering when human body is fallen.

In one embodiment, the anti-method for early warning of falling of human body provided by the invention also includes a step S7, is used for judging whether above-mentioned warning message is issued guardian. When being embodied as, generally preset a time of fire alarming and store, to start timing after generating warning message in step s 6, if monitored target is thought without reporting to the police, then within the above-mentioned preset alarm time, cancel this warning message, if monitored target does not cancel this warning message in predetermined time of fire alarming, then warning message is sent to guardian (step S8) after the meter full above-mentioned preset alarm time. The above-mentioned preset alarm time can be 20 seconds or 1 minute.

When utilizing the 3-axis acceleration information of sensor acquisition monitored target, three axis angular rate information and three axle geomagnetic field information, shake due to sensor, certain noise can be produced, therefore will to sensor acquisition to 3-axis acceleration information, three axis angular rate information and three axle geomagnetic field information be filtered carrying out again data process. Kalman filtering is suitable to real time signal processing, and estimated value and the observation of now of available previous moment update the estimation to state variable, obtain the estimated value of current time, it is possible to for processing in real time.

The embodiment of the present invention adopt Kalman filtering algorithm to sensor acquisition to data be filtered time, the discrete state equations of linear time varying system and observational equation be:

X (k)=F (k, k-1) X (k-1)+T (k, k-1) U (k-1)

(1)

Y (k)=H (k) X (k)+N (k)

In formula (1), X (k) and Y (k) is state vector and the measurement vector in k moment respectively, F (k, k-1) for state-transition matrix, U (k) is k moment dynamic noise, and T (k, k-1) controls matrix for system, H (k) is k moment observing matrix, and N (k) is the observation noise in k moment.

Before filtering starts, correlated variables is initialized, then utilized the state vector in a upper moment to estimate the state vector of current time:

$\widehat{X\left(k\right)}=F(k,k-1)\·X(k-1)---\left(2\right)$

Being the estimated value of current time (i.e. k moment) state vector, X (k-1) is the state vector of a upper moment (i.e. k-1 moment).

Secondly, pre-estimation covariance matrix is calculated according to the following formula

$\widehat{C\left(k\right)}=F(k,k-1)\·C\left(k\right)\·F^T(k,k-1)+T(k,k-1)\·Q\left(k\right)\·T^T(k,k-1)---\left(3\right)$

Wherein, Q (k) is intermediate variable, Q (k)=U (k) U^TK () is intermediate variable, the covariance matrix that C (k) is current time, and subscript T represents the transposition of matrix.

After trying to achieve pre-estimation covariance matrix, substitute into and formula (4) calculate kalman gain matrix:

$K\left(k\right)=\frac{\widehat{C\left(k\right)}\·H^T\left(k\right)}{H\left(k\right)\·\widehat{C\left(k\right)}\·H^T\left(k\right)+R\left(k\right)})---\left(4\right)$

Wherein, R (k) is intermediate variable, R (k)=N (k) N^T(k)��

Secondly, the estimated value of state vector is updated:

$\widehat{X\left(k\right)}=K\left(k\right)\·\[Y\left(k\right)-H\left(k\right)\·\widehat{X\left(k\right)}\]+\widehat{X\left(k\right)}---\left(5\right)$

Again, the estimated value of the covariance matrix after updating is calculated:

$\widehat{C\left(k\right)}=\[I-K\left(k\right)\·H\left(k\right)\]\·\widehat{C\left(k\right)}\·{\[I-K\left(k\right)\·H\left(k\right)\]}^T+K\left(k\right)\·H\left(k\right)\·\widehat{K\left(k\right)}---\left(6\right)$

Wherein, I is unit matrix.

Finally, state vector and measurement vector are updated:

$\begin{array}{c}X\left(k\right)=\widehat{X\left(k\right)}\\ C\left(k\right)=\widehat{C\left(k\right)}\end{array}---\left(7\right)$

According to formula (2) to (7) cycle calculations, until completing filtering.

Ask for monitored target in step s 2 when the resultant acceleration of each sampled point, it is possible to use acceleration modulus algorithm calculates the monitored target resultant acceleration A at each sampled point according to formula (8)_SVM:

$A_{SVM}=\sqrt{A_x^2+A_y^2+A_z^2}---\left(8\right)$

Wherein, A_x��A_y��A_zRespectively monitored target is at three axial acceleration of a certain sample point.

The advantage of resultant acceleration (acceleration mould vector) mode is in that to be combined into three accekerations axially the change of one accekeration, no matter whichaway is fallen to make human body, can both reflect that acceleration is worth change, rational threshold value is set and just can complete judgement of simply falling.

Falling moment at human body, and can produce a very big acceleration between ground, therefore the present invention adopts acceleration information to detect to fall, compare individual axis acceleration information, and resultant acceleration more can reflect the impact process size that human body is fallen. Fig. 2 is the resultant acceleration change curve before and after embodiment of the present invention monitored target is fallen, and Fig. 3 is that monitored target is sat down the resultant acceleration change curve before and after action, and Fig. 4 is that monitored target is hurried up/jog the curve of the resultant acceleration change before and after action. The present invention is through carrying out correlation analysis to measurement data, it is possible to find the kinestate using resultant acceleration instantaneous value to judge monitored target, very easily obscures " running " and occurs with " falling ", produces false alarm information. Namely utilize merely resultant acceleration to differentiate to fall, false positive and false negative can be caused to increase. Reason is in that, the monitored target instantaneous resultant acceleration value under " running " and " falling " two kinds of kinestates, it is possible that equal situation, causes that by " running " condition discrimination be " falling " state, thus producing false-alarm. Contrast finds, the action excessive velocities if run, instantaneous resultant acceleration will be produced and exceed the situation of resultant acceleration when falling, and the situation that resultant acceleration when falling exceedes a certain setting threshold value only occurs in the very short time, and the resultant acceleration of action of running can periodically exceed threshold value.

Therefore, in step s 4, resultant acceleration according to each sampled point judge the kinestate of monitored target be whether doubtful fall state time, a resultant acceleration threshold value need to be preset and store, when the resultant acceleration of at least one sampled point obtained in step S2 is more than above-mentioned resultant acceleration threshold value, then judge that the kinestate of monitored target is as doubtful state of falling. The kinestates such as the doubtful state of falling herein includes running, jogs, hurries up, stands, sits down, jumps, rotation, also include falling. But now it is only capable of judging whether monitored target is doubtful state of falling only by resultant acceleration, will determine whether monitored target falls tendency, in addition it is also necessary to the size in conjunction with Eulerian angles determines whether.

In order to reduce human body due to erroneous judgements that significantly action causes such as race, jumpings with for whether starting auto alarm signal offer foundation after falling, the embodiment of the present invention introduces Eulerian angles as the subsidiary conditions judged of falling, the attitude angle of Eulerian angles and human body. First Real-time Collection three axis angular rate, 3-axis acceleration, three axle geomagnetic field information, then quaternary number mode is adopted to calculate three Eulerian angles (angle of pitch, roll angle, course angle) of human body, then utilize three Eulerian angles information and resultant acceleration, identify the kinestate (sit down, walking, lie down, fall) of monitored target according to recognizer.

When being embodied as, in step s3 the 3-axis acceleration information collected, three axis angular rate information and three axle geomagnetic field information are fused into Eulerian angles, then Eulerian angles are being carried out difference.

For referential in three dimensions, the orientation of any coordinate system can represent by three Eulerian angles. Eulerian angles use course angle yaw, angle of pitch pitch and roll angle roll to represent the rotational value on three axial components of x, y, z, it is possible to the difference changed according to the angle of above three Eulerian angles judges the change of monitored target attitude. When being embodied as, described course angle, the angle of pitch and roll angle are carried out respectively difference, obtain comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference:

$\left\{\begin{array}{c}{d}_{roll}=roll\left(k\right)-roll(k-1)\\ d_{pitch}=pitch\left(k\right)-pitch(k-1)\\ d_{yaw}=yaw\left(k\right)-yaw(k-1)\end{array}\right.---\left(9\right)$

In formula, yaw is course angle, and pitch is the angle of pitch, and roll is roll angle, and k is current sample time, and k-1 is a upper sampling instant. Wherein d_roll(0)=d_pitch(0)=d_yaw(0)=0,1 < k < K_max, K_maxWhether, for the length of above-mentioned sequence of differences, identical with sampled point quantity, when the sampling period is T, the present invention generally samples within 0.3��0.6 sampling period, fall tendency detecting monitored target, to realize early warning.

The deficiency sensitive for the threshold value in instantaneous resultant acceleration discriminant approach, rate of false alarm is high, in step s 5, the attitude of monitored target is identified the present invention by sequence of differences further that obtain in integrating step S3.

When being embodied as, when the kinestate of the monitored target being in doubtful state of falling is determined whether, the present invention has preset a resultant acceleration average threshold value and locally stored, according to the resultant acceleration average being calculated monitored target by the resultant acceleration of all sampled points, when the resultant acceleration average tried to achieve is more than default resultant acceleration average threshold value, and when the absolute value of at least one above-mentioned Eulerian angles difference (difference of such as current time and a upper moment angle of pitch) is more than described Eulerian angles difference threshold, then judge that monitored target has tendency of falling.

In sum, the mode being based on inertia sensing technology that the embodiment of the present invention adopts detects the behavior of falling, monitoring principle based on " resultant acceleration is judged as main; Eulerian angles are judged as auxiliary ", do not need external auxiliary equipment, internal algorithm is complicated, but has real-time, reliability high.

Based on the inventive concept that method for early warning of preventing with the human body shown in Fig. 1 falling is identical, the embodiment of the present invention additionally provides the anti-early warning system of falling of a kind of human body, as described in example below. Owing to this human body is prevented falling, to solve the principle of problem similar with the anti-method for early warning of falling of human body in Fig. 1 for early warning system, and therefore prevent the falling enforcement of early warning system of this human body may refer to human body in Fig. 1 and prevents falling the enforcement of method for early warning, and repetition part repeats no more.

Fig. 5 is that embodiment of the present invention human body is prevented falling the structural representation of early warning system. As it is shown in figure 5, the anti-early warning system of falling of human body includes: sensor unit 1, resultant acceleration computing unit 2, Eulerian angles sequence of differences generate unit 3, judging unit 4, judging unit 5, warning message generation unit 6.

Sensor unit 1 is for the 3-axis acceleration information of monitored target of the multiple sampled point of Real-time Collection, three axis angular rate information and three axle geomagnetic field information.

Resultant acceleration computing unit 2 is for calculating the monitored target resultant acceleration at each sampled point according to described 3-axis acceleration information. When being embodied as, the 3-axis acceleration information that resultant acceleration computing unit 2 can collect according to sensor unit 1, utilize acceleration modulus algorithm to calculate the monitored target resultant acceleration at each sampled point.

Eulerian angles sequence of differences generates unit 3 for generating Eulerian angles sequence of differences according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information.

Judging unit 4 is for according to whether the resultant acceleration of each sampled point and the kinestate of default resultant acceleration threshold decision monitored target are doubtful state of falling.

Judging unit 5 for judge at judging unit 4 kinestate of monitored target be doubtful fall state time, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold.

Warning message generates unit 6 for generating warning message when monitored target falls tendency, to report to the police before monitored target is fallen.

In one embodiment, the anti-early warning system of falling of above-mentioned human body also includes timing unit 7 and a judging unit 8. Wherein, timing unit 7 starts timing for generating after unit 6 generates warning message at warning message, it is judged that unit 8 is used for judging whether monitored target cancels above-mentioned warning message within the preset alarm time.

In one embodiment, the anti-early warning system of falling of above-mentioned human body also includes locating unit 9, for gathering the geographic position data of monitored target. Such as, when monitored target is in outdoor environment, positioning unit 9 can adopt the Big Dipper/GPS dual-mode location to obtain the geographic position data of monitored target; When monitored target is in indoor environment, positioning unit 9 can adopt architecture to obtain the geographic position data of monitored target.

Common, the anti-early warning system of falling of human body also includes a communication unit 10, for warning message being sent to guardian after the timing unit 7 full above-mentioned preset alarm time, and this warning message includes the geographic position data that positioning unit 9 gathers, in order to monitoring launches rescue after knowing the place that monitored target is fallen.

Due to sensor unit 1 gather monitored target exercise data time it may happen that shake; the exercise data so collected has noise; for filtering noise; the anti-early warning system of falling of above-mentioned human body generally also includes a Kalman filtering unit 11; for the 3-axis acceleration information of each sampled point collected, three axis angular rate information and three axle geomagnetic field information are carried out Kalman filtering, so that the follow-up precision carried out when data process is higher.

Alarm unit 12 includes voice alarm module, buzzer, vibrator and LED, and the warning message for generating unit 6 generation according to warning message sends sound and light alarm.

Usually, Eulerian angles sequence of differences generation unit 3 includes a digital moving processor and a difference calculating module. Digital moving processor is fused into Eulerian angles for 3-axis acceleration information, three axis angular rate information and the three axle geomagnetic field information collected by sensor unit 1, and wherein above-mentioned Eulerian angles include course angle, the angle of pitch and roll angle. Difference calculating module, for above-mentioned course angle, the angle of pitch and roll angle are carried out difference respectively, obtains comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference.

Judging unit 4 is when being identified the kinestate of monitored target with resultant acceleration threshold value according to resultant acceleration, when having only to the resultant acceleration of a sampled point more than default resultant acceleration threshold value, namely can determine that the kinestate of monitored target is doubtful state of falling. If the resultant acceleration having multiple sampled point is all higher than above-mentioned resultant acceleration threshold value, then the kinestate one of monitored target is decided to be doubtful state of falling.

Judging unit 5 generally includes a resultant acceleration mean value computation module and and judges submodule. Wherein, average is accelerated in the conjunction that resultant acceleration mean value computation module calculates monitored target for the resultant acceleration according to each sampled point. Judge that submodule is for when above-mentioned resultant acceleration average is more than default resultant acceleration average threshold value, and the absolute value of at least one above-mentioned Eulerian angles difference more than default Eulerian angles difference threshold time, it is determined that monitored target has tendency of falling.

When monitored target falls tendency, the probability injured when falling for reducing monitored target, the anti-system of falling of above-mentioned human body also includes an air bag actuator 13. Air bag actuator 13 generally includes a compressed gas cylinder and an air bag, air bag actuator 13 is generated unit 6 with warning message and can be connected by steering wheel interface, for puncturing the air port of compressed gas cylinder according to above-mentioned warning message so that air bag to be inflated, thus provide buffering when human body is fallen, reduce personal injury's probability. Above-mentioned air bag can make the form of belt or vest, but the present invention is not limited thereto.

Air bag actuator 13 in the embodiment of the present invention can be designed to the purse form being made up of dark lattice and storage space, dark lattice are used for fixing control circuit plate and gas cylinder, actuator, purse bandage includes air bag, there is purse, early warning of falling, protection of falling, lamp several functions of seeking help of falling. It addition, air bag actuator 13 is also designed to other forms such as belt, vest, flexible design is various.

Monitored target fall tendency time trigger air bag actuator for body weight for humans want joint provide protect, communication unit 10 is timely alert after monitored target is fallen, guardian can receive, by network, note, message of falling, can hear, user people at one's side, the voice help information that alarm unit 12 sends, offer help in time.

The anti-early warning system of falling of human body also includes a memory element 14, and for the intermediate data generated in the data of storage sensor list 1 yuan collection, information processing process process, and warning message generates the warning message etc. that unit 6 generates.

At present, on market, most of similar products do not have memory module 14 or use fixing storage chip on circuit boards. Human body in embodiment of the present invention memory module 14 in early warning system of preventing falling can select built-in TF card, storage inertia sensing data, the parameters such as pre-warning time of falling, conveniently carry out off-line analysis, and convenient plug, contribute to user is carried out personality analysis, modeling, it is provided that personalized service more accurately. In another embodiment, memory module can also adopt SDHC card.

The anti-early warning system of falling of above-mentioned human body also includes power subsystem 15, for human body being prevented, each unit falling in early warning system is powered.

Human body anti-fall early warning system and beneficial effect thereof in order to be more fully understood that the present invention, illustrates below in conjunction with specific example.

Fig. 6 is that embodiment of the present invention human body is prevented falling the frame structure of early warning system and connection diagram. The human body that the present embodiment provides prevent the falling controller mainboard of early warning system is made up of sensor unit, micro controller module, supply module, communication module, alarm unit, rudder control interface module and other interfaces, and its connected mode is shown in Fig. 6.

When the anti-early warning system of falling of human body shown in Fig. 6 is carried out PCB layout, can by power line and holding wire layered arrangement, avoid parallel long distance cabling, to improve Electro Magnetic Compatibility, sensor unit and micro controller module are integrated on same circuit board simultaneously, make the anti-early warning system of falling of human body of more high integration and specificity.

As shown in Figure 6, sensor unit in the present embodiment adopts the combination of MPU6050 chip and AK8975 chip, MPU6050 built-in chip type three axis accelerometer and three-axis gyroscope, AK8975 built-in chip type three axle magnetometers, it is respectively used to gather the 3-axis acceleration information of monitored target, three axis angular rate information and three axle geomagnetic field information of monitored target.

The main control chip of micro controller module adopts the STM32F407VGTE microcontrol processor of ST company, it has 32 ARMContexM4 kernels, and there is 1MBFlash, 192KBRAM, it is possible to provide the clock of up to 168MHz, it is ensured that algorithm can the execution of fast and reliable. Micro controller module is by spi bus with three-axis gyroscope, three axis accelerometer, three axle magnetometers (for sake of convenience, the present invention by this by three sensors referred to as nine axle inertial sensors) communication, the 3-axis acceleration of human body, three axis angular rates, three axle geomagnetic field information can be received, thus the identification for human motion and attitude provides the original motion data to originate. The exercise data of nine axle inertial sensor collections is carried out data fusion and the gesture recognition of complexity by main control chip, thus judging the athletic posture of monitored target. Attitude according to monitored target realizes the dangerous play to people and reminds and seek help when falling. It addition, the abundant serial port resource that main control chip additionally provides coordinates communication module (such as Simcom company releases SIM900) can realize the data communication with server, it is also possible to realized and tutorial short message interacting by short message mode.

In another embodiment, the communication between sensor unit and micro controller module can also adopt I²C mode.

In the present embodiment, alarm unit group to be made up of audio alert, buzzer, vibrator and LED, wherein, audio alert part is made up of a piece of speech chip and loudspeaker, can also being sought help by voice crowd towards periphery when monitored target is fallen, its hardware composition frame chart is as shown in Figure 7.

Speech chip receives the alarm signal from microcontrol processor, selects the voice segments order preset, and exports corresponding voice signal to loudspeaker, drives loudspeaker to send sound. It addition, vibrator receives signal, producing vibration, danger the action of prompting monitored target, in case falling. The mode of vibration old people to being hard of hearing and the scene needed peace and quiet have better effect. After alarm unit receives warning message, LED therein can constantly be glimmered, and points out dangerous play behavior.

Locating module can adopt the UM220-III-N and antenna and peripheral circuit composition thereof that produce with Xin Xingtong company, it is the third generation BDS/GPS bimodulus locating module newly released with Xin Xingtong company, it it is the smallest size of BDS/GPS module domesticized completely in the market, there is the features such as low-power consumption, low cost, high performance-price ratio and high integration, support single system location-independent and multisystem combined location, under bimodulus positioning scenarios, outdoor positioning precision can reach 3-5 rice, being much higher than the GPS precision 10-15 rice individually positioned, positioning precision is outstanding.

In the present embodiment, the anti-early warning system of falling of human body can select battery power supply mode, it is also possible to selects USB power supply mode. When being elected to by battery power supply mode, it is possible to adopting the 3.7V lithium battery power supply of standard, final product form can be non-dismountable formula battery. If selecting USB power supply mode, then can using the MiniUSB interface of standard, the charging management chip TP4056 in combined charge circuit completes the charging to lithium battery. USB power source can directly be powered for whole circuit board after linear voltage stabilization chip PAM3310 (not shown) voltage stabilizing rectification. It addition, be generally also configured with two charging indicator lights in the anti-early warning system of falling of human body, red colored lamp is used for indicating power supply to connect, and blue lamp is used for glimmering continuously when battery underfill, being full of rear Chang Liang.

Why the anti-early warning system of falling of human body that the embodiment of the present invention provides can realize human body is fallen protection, be by detect fall after control steering wheel puncture compressed gas cylinder, quickly realize airbag aeration wanting joint to protect body weight for humans, it is to avoid human body hard landing. It is " control-perform " separation design due to what adopt, therefore after being completed exercise data acquisition, human body attitude analysis identification work by controller mainboard, completed, by air bag actuator, protection action of falling again, realized by steering wheel interface module reserved on controller mainboard mutual between the two. Steering wheel interface module is made up of power supply and holding wire, and holding wire can connect the control signal wire of steering wheel, and the control completing steering wheel rotates; Power line provides the voltage of 7.8V, thinks that steering wheel power supply is powered, and additionally power line also provides charging inlet for steering wheel battery. In the present invention, controlling, the separate design of actuator, facilitate the design of product and wearing of user, what add product manifests form.

Controller mainboard also leaves jtag interface, STM32F407 serial port debugger can be passed through download from host computer to main control chip, 20 reserved extension mouths are to be connected with other system, such as connection can adjust the anti-intervening measures of falling such as " the anti-rehabilitative shoe of falling " of plantar pressure branch automatically, realizes more function later.

This human body main applicable object of early warning system of preventing falling is elderly population or easily falls risk population; rely on and be worn on the user's body needing monitoring; realize user's attitude Real-time Collection; thus making corresponding early warning and protection act when occurring of falling; its main workflow includes initialization, monitoring pattern adjustment, fall early warning and the steps such as process of falling, as shown in Figure 8.

Initialization step: on user wears after the anti-system of falling of human body, open the power supply of this system, enter init state, all will carry out initialization operation including micro controller module, communication module, locating module and sensor unit. It should be noted that the initialization time of nine axle data pick-ups is longer, it is necessary to about 30 seconds, user needs to wait the corresponding time, and period need to keep standing state, in order to pick up calibration.

Monitoring pattern adjusts: after having initialized, automatically into real-time monitoring pattern, the present embodiment human body prevents that the communication module in the system of falling is in running order, fall once be monitored to monitored target, communication module will send warning message from trend server, comprising the location information of monitored target in this warning message, communication module also can SMS notification guardian. The shortcoming of real-time monitoring pattern is exactly that human body prevents that in system of falling, all of hardware device is in state of activation, and power consumption is higher. For this, the present embodiment also sets the mode of mode adjustment, when nine axle sensors differentiate that human body remains static for a long time, can automatically switch to remote tracing pattern, in this mode, micro controller module reduces the purpose of power consumption by opening and closing communication module discontinuously. Can also manually arranging conventional tracking and monitoring pattern, communication module is closed in this mode, greatly reduces power consumption, and shortcoming is cannot to notify guardian after monitored target is fallen in time, is adapted under the scene of indoor environment, household's company and uses. It addition, can mutually switch between above-mentioned Three models, it is possible to select intelligent mode, namely micro controller module is according to the automatic identification switching of the data gathered, it is also possible to switched over by switching, send the mode of command set by user.

The hardware resource used under different monitoring patterns is different, causes that hardware power consumption is different. In order to adapt to different scenes, reduce power consumption, extend the purpose of use time, the power consumption of early warning system of falling also for making human body prevent can realize Intelligent adjustment, and the embodiment of the present invention additionally provides a kind of human body and prevents falling the monitoring pattern method for transformation of early warning system, as shown in Figure 9.

After system initialization, acquiescence enters real-time monitoring pattern, and the kinestate of monitored target is monitored in real time. Monitored target can Non-follow control, send corresponding instruction and be switched to other patterns, to adapt to specifically used scene. Otherwise, under default situations, when monitored target remains stationary the regular hour, the anti-fall monitoring system of human body then thinks that monitored target is in non-athletic state, remote tracing monitoring pattern will be automatically switched to, communication module by separated in time dormancy once, to reduce power consumption. If fallen during dormancy, micro controller module will actively wake communication module up, initializes rapidly and alert. The embodiment of the present invention provides note, GPRS data communication and three kinds of command interfaces of serial ports, it is achieved mutual with human body anti-fall monitoring system.

Multiple-working mode expands the use scene of the present invention, and the intelligence switching of monitoring pattern greatly reduces product power consumption, reduces the operation of user as far as possible, improves its ease for use, it is achieved that wearer can the function of intelligent selection pattern without intervening.

Falling early warning: no matter under which kind of monitoring pattern, once monitored target has tendency of falling, micro controller module all can remind monitored target note taking precautions against fall risk by alarm unit, and concrete dangerous play may refer to table 1. After detecting that monitored target is fallen, it is probably monitored target maloperation or to monitored target injury relatively light (air bag can want joint to provide protection for body weight for humans when falling), therefore the present embodiment additionally provides a kind of mechanism cancelling warning, if monitored target independently cancels warning within the time of regulation, then warning message is not transmitted to guardian; If monitored target cannot independently cancel warning within the time of regulation, namely injury situation ratio is more serious, then, after timing reaches the time of regulation, warning message will be sent to server and guardian by communication module.

Table 1 dangerous play list

It addition, by the distinguishing rule in table 1, in conjunction with human body attitude, the present embodiment also distinguished lie down, stand up, stand up, sit down, several actions such as fall, and can accurately distinguish and give corresponding information, refer to table 2 and table 3.

Table 2 common action list

Table 3 doubtful fall state and judgement of falling

Wherein, " inclination angle " in table 1 above��table 3 refers to the absolute value of Eulerian angles difference.

Fall process: according to the distinguishing rule in table 2 and table 3; when human body prevent the micro controller module falling in early warning system detect monitored target have fall tendency time; now human body does not land; compressed gas cylinder can be triggered by rudder control interface; moment is airbag aeration; air bag launches before human body lands, thus realizing the protection to health key position. If not pressing cancellation alarm keys at monitored target within the preset alarm time (such as 20 seconds), micro controller module will send warning message by wireless communication module to server and guardian. Meanwhile, alarm unit, according to warning message, opens voice alarm module, and relief information is reported in circulation, reminds surrounding population to give the wounded relief. And communication module provides three kinds of interface operation modes: note, GPRS data and USB (serial ports), the effect that these three interface realizes is just as, can realizing human body is prevented the fall configuration of early warning system, operation and data interactive function, different scenes can select operating and controlling interface flexibly. The warning message sent to server is the data with specific format, mainly comprise: synchronous head " BSN ", telephone number, base station location information, GPS/BD position information (longitude and latitude), human body anti-fall early warning system voltage and verification and etc., the size of Frame is fixed, and facilitates packet to resolve. Wherein, BSN represents BodySensorNetwork, means human body sensor network, for instance a complete data content following " BSN130119511******** ... ". The short message content sent to guardian includes relief information, longitude and latitude, and content is customizable.

Figure 10 is embodiment of the present invention gesture recognition algorithms schematic flow sheet. According to the reference standard provided in table 1, table 2, in conjunction with the algorithm flow shown in Figure 10, the attitude that human body is current can be identified by the exercise data of sensor acquisition, in order to represent the attitude of human body more accurately, classification situation in table 1 and table 2 can also be segmented further, with the more complete real-time attitude describing human body. It should be noted that, low-power consumption monitoring is in order to the monitoring pattern before coordinating adds, it is defined as: when acceleration maintains near 0 within a certain period of time, and sensor acquisition to exercise data slowly fluctuate in prescribed limit time, then it is assumed that user is in non-athletic state.

Through experimental verification, result proves that human body provided by the invention prevents falling method for early warning and system is feasible. The embodiment of the present invention chooses adult human (10 male 5 female of 15 young healthy altogether, age: 24 �� 3.5 years old, body weight: 62 �� 14.5kg, height: 170 �� 12cm) take part in experiment, experimental arrangement includes calibrating, sits down, stands up, lies down, a left side is stood up, the right side is stood up, front fall, right fall and after fall, everyone respectively does 3 times at each action, altogether do 360 times (without calibration actions) to fall experiment, simultaneously with the pre-warning time of falling (data that can also pass through to store in memory module draw) that a high-speed camera record is actual.

The results show, when being set to 4.5m/s in order to the acceleration rate threshold of the early warning that judges to fall²Time, its detection sensitivity is 98.61%, and specificity is 98.61, and results contrast is desirable, and pre-warning time of on average falling is 300ms, if suitably reducing sensitivity (such as 80%��90%), pre-warning time can reach 400ms��500ms. Figure 11 is the figure that the result according to above-mentioned 360 experiment samples of falling is made. In fig. 11, + number expression outlier, piston-like figure has 5 horizontal lines from top to bottom, represents maximum, 75% quantile, median, 25% quantile, minima respectively, so the box indicating in the middle part of piston-like figure concentrates on the experiment sample of 25%��75%. The present invention only illustrates for Right deviation, volunteer is when falling experiment as Right deviation, the pre-warning time of major part experiment sample concentrates between 0.25s��0.31s, the pre-warning time only having an outlier is about 0.45s, illustrates that human body provided by the invention prevent the falling early warning sensitivity of method for early warning and system is higher.

The present invention to fall the real-time judged and accuracy more remarkable, it is achieved that stand up, sit down, lie down, the various motion detection such as stand up, significant increase discrimination of falling.

Find according to documents and materials and actual test, the hands arm of human body, the position such as wrist and leg is because action randomness is very big in daily life behavior, activity is also relatively frequently, and below cervical region, the position of waist area above in physical activity process movable infrequently and in the process of falling acceleration and angular velocity can find violent change, and the position of centre of gravity of human body also concentrates on this subregion, nine axle sensor devices are worn on the position in the middle of waist and can reflect the kinestate of whole trunk and the change of attitude better, it is more conducive to fall detection.

In the embodiment shown in fig. 6, sensor unit directly can also be replaced the combination of MPU6050 chip and AK8975 chip by MPU9150 chip and peripheral circuit thereof, MPU9150 chip is INVENSENSE company new generation product, accelerometer, gyroscope and magnetometer it is integrated with in one single chip, and built-in DMP (DigitalMotionProcessor, digital moving processor) merge for attitude, built-in chip type 16 bit A/D converter, 16 bit data output (magnetic field data is 13 outputs), have self-calibration function.

Sensing unit in the embodiment of the present invention separates independent design with microcontroller mainboard, it is simple to upgrading or changing other sensor units is conducive to improving the Electro Magnetic Compatibility of whole circuit board; Reserved multiple I/O mouths (Input/Output, input/output) ensure that maintainability and the extensibility of product, is also beneficial to later stage HardwareUpgring.

The present invention is directed to old people etc. easily to fall crowd and design, there is early warning of falling, protection and the function such as communicate, before human body is fallen, namely have the function of " early warning ", breach similar products on Vehicles Collected from Market only have fall after the function of Alarm or only have early warning and without falling safeguard function. Human body provided by the invention prevent the falling overall construction design of early warning system is ingenious, can individually wear, use as warning function; Air bag actuator can also be coordinated to use, it is achieved fall rear protecting and relief; It addition, reserved serial ports can conveniently realize IAP (InApplicationPrograming, in-service units), it is beneficial to firmware upgrade.

Acceleration mode is generally only had for similar products on market, or the sensing data of multiple mode is provided by different circuit modules, it is unfavorable for the problem merged in real time and accurately of multi-modal sensing data, the present invention adopts the sensor chip of high accuracy and multi-modal data, and nine axle inertia sensing design data on same circuit module, the precision and the mode quantity that collect data are relatively more, can much sooner and accurately to the judgement fallen.

Present invention achieves " early warning " before human body is fallen, trigger air bag actuator when falling wants joint to provide protection for body weight for humans simultaneously, timely alert after falling, guardian can receive, by network, note, message of falling, can hear, user people at one's side, the voice help information that precaution device sends, offer help in time.

The present invention is the anti-precaution device of falling of a kind of intelligent human-body. For the problem that current social aging is serious, old people's coordination of body ability progressively declined with the age, falls in daily life inevitable, and falls and would generally cause very big life security and economic problems. And prior art mostly has shortcoming or the deficiencies such as above-mentioned " reporting to the police after falling ", " wearing use convenient not ", " accuracy of falling judges undesirable ", therefore the present invention devises the anti-precaution device of falling of a kind of intelligent human-body. Whether this precaution device inevitably can occur before human body is fallen in 300ms��500ms anticipation behavior of falling, if can avoid, can be avoided risk by voice, buzzer and electromagnetic shaker call user's attention; If inevitable, then can provide protection by trigger protection device, notify that guardian, voice reminder surrounding population provide relief by wireless network simultaneously.

Applying specific embodiment in the present invention principles of the invention and embodiment are set forth, the explanation of above example is only intended to help to understand method and the core concept thereof of the present invention; Simultaneously for one of ordinary skill in the art, according to the thought of the present invention, all will change in specific embodiments and applications, in sum, this specification content should not be construed as limitation of the present invention.

Claims (21)

1. the anti-method for early warning of falling of human body, it is characterised in that the anti-method for early warning of falling of described human body includes:

The 3-axis acceleration information of the monitored target of the multiple sampled point of Real-time Collection, three axis angular rate information and three axle geomagnetic field information;

The monitored target resultant acceleration at each sampled point is calculated according to described 3-axis acceleration information;

Eulerian angles sequence of differences is generated according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information;

Whether resultant acceleration and the kinestate of default resultant acceleration threshold decision monitored target according to each sampled point are doubtful state of falling;

If it is, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold;

Warning message is generated, to report to the police before monitored target is fallen when monitored target falls tendency.

2. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that the anti-method for early warning of falling of described human body also includes:

Air bag actuator punctures gas cylinder therein so that air bag therein to be inflated according to described warning message, to provide buffering when human body is fallen.

3. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that the anti-method for early warning of falling of described human body also includes:

Timing is started, it is judged that whether described monitored target cancels described warning message within the preset alarm time, if it is not, then described warning message is sent to guardian after the meter full described preset alarm time after generating described warning message.

4. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that the anti-method for early warning of falling of described human body also includes:

The described 3-axis acceleration information of each sampled point collected, three axis angular rate information and three axle geomagnetic field information are carried out Kalman filtering.

5. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that calculate the monitored target resultant acceleration at each sampled point according to described 3-axis acceleration information, including:

According to described 3-axis acceleration information, acceleration modulus algorithm is utilized to calculate the monitored target resultant acceleration A at each sampled point_SVM:

$A_{SVM}=\sqrt{A_x^2+A_y^2+A_z^2}$

Wherein, A_x��A_y��A_zRespectively monitored target is at three axial acceleration of same sample point.

6. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that generate Eulerian angles sequence of differences according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information:

Described 3-axis acceleration information, three axis angular rate information and three axle geomagnetic field information are fused into Eulerian angles, and described Eulerian angles include course angle, the angle of pitch and roll angle;

Described course angle, the angle of pitch and roll angle are carried out respectively difference, obtain comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference:

$\left\{\begin{array}{c}{d}_{roll}=roll\left(k\right)-roll(k-1)\\ d_{pitch}=pitch\left(k\right)-pitch(k-1)\\ d_{yaw}=yaw\left(k\right)-yaw(k-1)\end{array}\right.;$

Wherein, yaw is course angle, and pitch is the angle of pitch, and roll is roll angle, and k is current sample time, and k-1 is a upper sampling instant.

7. the anti-method for early warning of falling of human body according to claim 1, it is characterised in that according to whether the kinestate of the resultant acceleration of each sampled point and default resultant acceleration threshold decision monitored target is doubtful state of falling, including:

When the resultant acceleration of at least one sampled point is more than described resultant acceleration threshold value, it is determined that the kinestate of described monitored target is doubtful state of falling.

8. the anti-method for early warning of falling of human body according to claim 1, it is characterized in that, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold, including:

Described resultant acceleration according to each sampled point, average is accelerated in the conjunction calculating monitored target;

When described resultant acceleration average is more than default resultant acceleration average threshold value, and the absolute value of at least one described Eulerian angles difference more than described Eulerian angles difference threshold time, it is determined that described monitored target has tendency of falling.

9. the anti-early warning system of falling of human body, it is characterised in that the anti-early warning system of falling of described human body includes:

Sensor unit, for the 3-axis acceleration information of monitored target of the multiple sampled point of Real-time Collection, three axis angular rate information and three axle geomagnetic field information;

Resultant acceleration computing unit, for calculating the monitored target resultant acceleration at each sampled point according to described 3-axis acceleration information;

Eulerian angles sequence of differences generates unit, for generating Eulerian angles sequence of differences according to the described 3-axis acceleration information of each sampled point, three axis angular rate information and three axle geomagnetic field information;

First judging unit, for according to whether the resultant acceleration of each sampled point and the kinestate of default resultant acceleration threshold decision monitored target are doubtful state of falling;

Second judging unit, for judge at described first judging unit the kinestate of monitored target be doubtful fall state time, according to the resultant acceleration average obtained by the resultant acceleration of all sampled points and default resultant acceleration average threshold value, judge whether monitored target falls tendency in conjunction with described Eulerian angles sequence of differences and default Eulerian angles difference threshold;

Warning message generates unit, for generating warning message when monitored target falls tendency, to report to the police before monitored target is fallen.

10. the anti-early warning system of falling of human body according to claim 9, it is characterised in that the anti-early warning system of falling of described human body also includes:

Timing unit, for starting timing after generating described warning message;

3rd judging unit, is used for judging whether described monitored target cancels described warning message within the preset alarm time.

11. the anti-early warning system of falling of human body according to claim 10, it is characterised in that the anti-early warning system of falling of described human body also includes: positioning unit, for gathering the geographic position data of described monitored target.

12. the anti-early warning system of falling of human body according to claim 11, it is characterized in that, the anti-system of falling of described human body also includes: communication unit, and for described warning message being sent to guardian after the meter full described preset alarm time, described warning message includes described geographic position data.

13. the anti-early warning system of falling of human body according to claim 9, it is characterised in that the anti-early warning system of falling of described human body also includes:

Kalman filtering unit, for carrying out Kalman filtering to the described 3-axis acceleration information of each sampled point collected, three axis angular rate information and three axle geomagnetic field information.

14. the anti-early warning system of falling of human body according to claim 9, it is characterised in that described resultant acceleration computing unit specifically for:

According to described 3-axis acceleration information, acceleration modulus algorithm is utilized to calculate the monitored target resultant acceleration A at each sampled point_SVM:

$A_{SVM}=\sqrt{A_x^2+A_y^2+A_z^2}$

Wherein, A_x��A_y��A_zRespectively monitored target is at three axial acceleration of same sample point.

15. the anti-early warning system of falling of human body according to claim 9, it is characterised in that described Eulerian angles sequence of differences generates unit and includes:

Digital moving processor, for described 3-axis acceleration information, three axis angular rate information and three axle geomagnetic field information are fused into Eulerian angles, described Eulerian angles include course angle, the angle of pitch and roll angle;

Difference calculating module, for described course angle, the angle of pitch and roll angle are carried out difference respectively, obtains comprising the sequence of described course angle difference, angle of pitch difference and roll angle difference:

$\left\{\begin{array}{c}{d}_{roll}=roll\left(k\right)-roll(k-1)\\ d_{pitch}=pitch\left(k\right)-pitch(k-1)\\ d_{yaw}=yaw\left(k\right)-yaw(k-1)\end{array}\right.;$

Wherein, yaw is course angle, and pitch is the angle of pitch, and roll is roll angle, and k is current sample time, and k-1 is a upper sampling instant.

16. the anti-early warning system of falling of human body according to claim 9, it is characterised in that described first judging unit specifically for:

When the resultant acceleration of at least one sampled point is more than described resultant acceleration threshold value, it is determined that the kinestate of described monitored target is doubtful state of falling.

17. the anti-early warning system of falling of human body according to claim 9, it is characterised in that described second judging unit includes:

Resultant acceleration mean value computation module, for the described resultant acceleration according to each sampled point, average is accelerated in the conjunction calculating monitored target;

Judge submodule, for when described resultant acceleration average is more than default resultant acceleration average threshold value, and the absolute value of at least one described Eulerian angles difference more than described Eulerian angles difference threshold time, it is determined that described monitored target has tendency of falling.

18. the anti-early warning system of falling of human body according to claim 9, it is characterized in that, the anti-early warning system of falling of described human body also includes an air bag actuator, described air bag actuator is generated unit with described warning message and is connected by steering wheel interface, for puncturing gas cylinder therein so that air bag therein to be inflated according to described warning message, thus provide buffering when human body is fallen.

19. the anti-early warning system of falling of human body according to claim 9, it is characterized in that, the anti-early warning system of falling of described human body also includes: alarm unit, including voice alarm module, buzzer, vibrator and LED, for sending sound and light alarm according to described warning message.

20. the anti-early warning system of falling of the human body according to any one of claim 9-19, it is characterized in that, the anti-early warning system of falling of described human body also includes: memory element, generates, for storing the intermediate data generated in the data of described sensor unit collection, described information process and described warning message, the warning message that unit generates.

21. the anti-early warning system of falling of the human body according to any one of claim 9-19, it is characterised in that the anti-early warning system of falling of described human body also includes: power subsystem, for powering to the anti-early warning system of falling of described human body.