CN107742399A - For sending the method and device of alarm signal - Google Patents

For sending the method and device of alarm signal Download PDF

Info

Publication number
CN107742399A
CN107742399A CN201711140773.XA CN201711140773A CN107742399A CN 107742399 A CN107742399 A CN 107742399A CN 201711140773 A CN201711140773 A CN 201711140773A CN 107742399 A CN107742399 A CN 107742399A
Authority
CN
China
Prior art keywords
tired
data
value
type
threshold
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201711140773.XA
Other languages
Chinese (zh)
Other versions
CN107742399B (en
Inventor
李宏言
拓冰
商兴奇
蔡浚宇
贾巍
李时聪
李鹏
贺诚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Baidu Online Network Technology Beijing Co Ltd
Beijing Baidu Netcom Science and Technology Co Ltd
Original Assignee
Beijing Baidu Netcom Science and Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Baidu Netcom Science and Technology Co Ltd filed Critical Beijing Baidu Netcom Science and Technology Co Ltd
Priority to CN201711140773.XA priority Critical patent/CN107742399B/en
Publication of CN107742399A publication Critical patent/CN107742399A/en
Priority to PCT/CN2018/099158 priority patent/WO2019095733A1/en
Application granted granted Critical
Publication of CN107742399B publication Critical patent/CN107742399B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/02—Alarms for ensuring the safety of persons
    • G08B21/06—Alarms for ensuring the safety of persons indicating a condition of sleep, e.g. anti-dozing alarms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1116—Determining posture transitions
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1118—Determining activity level

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Pathology (AREA)
  • Surgery (AREA)
  • Physiology (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Public Health (AREA)
  • Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Emergency Management (AREA)
  • Traffic Control Systems (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The embodiment of the present application discloses the method and device for sending alarm signal.One embodiment of this method includes：Transport condition data is obtained in real time；Above-mentioned transport condition data is imported to the tired identification model of training in advance, obtains the tired numerical value of tired type and corresponding tired type；Above-mentioned tired type and tired numerical value are imported to the attention force value computation model of training in advance, the power that gains attention numerical value；Judge the notice threshold range where notice numerical value, alarm signal is sent according to above-mentioned notice threshold range.This embodiment improves the accuracy rate of the degree of fatigue identification to driver.

Description

For sending the method and device of alarm signal
Technical field
The invention relates to field of computer technology, and in particular to technical field of vehicle control, more particularly, to Send the method and device of alarm signal.
Background technology
Automobile has expanded the scope of people's trip, and the trip to people brings facility, improves the quality of life of people. With the development and progress of science and technology, the automatic driving vehicle controlled by intelligence system can be obtained than manned automobile more More driving informations, possess higher security, turn into an important trend of future automobile development.
In vehicle travel process, easily there is traffic accident in fatigue driving in driver.In order to reduce traffic accident Occur, the methods of prior art to the eye motion of driver by being monitored judge driver whether fatigue driving.
The content of the invention
The purpose of the embodiment of the present application is to propose the method and device for sending alarm signal.
In a first aspect, the embodiment of the present application provides a kind of method for sending alarm signal, this method includes：In real time Transport condition data is obtained, above-mentioned transport condition data includes driver status data, driving environment data and vehicle-state number According to above-mentioned driver status data include hand motion data, mouth action data, headwork data and eye motion number Include lane line data, vehicle distances data according to, above-mentioned driving environment data, above-mentioned vehicle status data include speed data, Bearing data；Above-mentioned transport condition data is imported to the tired identification model of training in advance, obtains tired type and corresponding fatigue The tired numerical value of type, above-mentioned tired identification model are used to determine tired type and tired numerical value by transport condition data, on State the degree that tired numerical value is used to characterize corresponding tired type；Above-mentioned tired type and tired numerical value are imported to the note of training in advance Meaning force value computation model, the power that gains attention numerical value, above-mentioned attention force value computation model are used to pass through tired type and tired numerical value Notice numerical value is calculated, above-mentioned notice numerical value is used for the degree of fatigue for characterizing driver；Judge the note where notice numerical value Meaning force threshold scope, sends alarm signal, above-mentioned notice threshold range passes through default according to above-mentioned notice threshold range Notice that force threshold determines.
In certain embodiments, the step of above method includes building tired identification model, above-mentioned structure fatigue identification mould The step of type, includes：Carried respectively from driver status data acquisition system, driving environment data acquisition system and vehicle status data set Take time upper synchronous driver status data, driving environment data and vehicle status data；Respectively to driver status data, Driving environment data and vehicle status data carry out feature extraction, obtain driver status characteristic, driving environment characteristic According to vehicle-state characteristic；Tired type is determined according to above-mentioned driver status characteristic, above-mentioned tired type includes At least one of below：Eye closing type, type of yawning, sight offset type, type of making a phone call；Pass through driver status characteristic Tired numerical value is calculated according to, driving environment characteristic and vehicle-state characteristic；Using machine learning method, by driver's shape State characteristic, driving environment characteristic and vehicle-state characteristic are as input, by tired type and corresponding fatigue The tired numerical value of type obtains tired identification model as output, training.
In certain embodiments, above by driver status characteristic, driving environment characteristic and vehicle-state Characteristic, which calculates tired numerical value, to be included：Time threshold scope, distance threshold scope and data variable quantity threshold range are obtained, its In, time threshold scope includes the multiple time threshold subranges being made up of the time threshold pre-set, distance threshold scope The multiple distance threshold subranges formed including the distance threshold by pre-setting, data variation amount threshold range are included by advance Multiple data variation amount threshold value subranges that the data variation amount threshold value of setting is formed, each time threshold subrange are corresponding one Time weighting value, the corresponding distance weighting value of each distance threshold subrange, each data variation amount threshold value subrange are corresponding One data variation weighted value；The tired type for measuring every kind of tired type corresponding to driver status characteristic respectively continues Data variation value corresponding to distance change value corresponding to time, driving environment characteristic and vehicle-state characteristic, wherein, The tired type duration is by specifying the duration acted corresponding to tired type corresponding to driver status characteristic To characterize, distance change value is characterized by variable quantity of the driving environment characteristic within the unit interval, and data variation value is led to Variable quantity of the vehicle-state characteristic within the unit interval is crossed to characterize；It is determined that tired type duration, distance change value With data changing value respectively corresponding to the time weighting value of time threshold subrange of time threshold scope, distance threshold scope The distance weighting value of distance threshold subrange and the data of the data variation amount threshold value subrange of data variable quantity threshold range become Change weighted value；The tired of corresponding tired type is obtained to time weighting value, distance weighting value and data variation weighted value weighted sum Labor numerical value.
In certain embodiments, the above method includes the step of structure pays attention to force value computation model, above-mentioned structure notice The step of value computation model, includes：By the tired numerical division in tired numerical value set it is corresponding tired type according to tired type Tired numerical value subclass；Respectively every kind of tired type sets type weight value, by type weight value and the type weighted value The product of each tired numerical value in corresponding tired numerical value subclass is the tired numerical value in value as corresponding tired type When attention force component；Using machine learning method, using tired type and tired numerical value as input, will pay attention to force component as Output, train the force value computation model that gains attention.
In certain embodiments, the notice threshold range where above-mentioned judgement notice numerical value, according to above-mentioned notice Threshold range, which sends alarm signal, to be included：Notice threshold range is obtained, above-mentioned notice threshold range is included by pre-setting Multiple attention force threshold subranges for forming of attention force threshold, it is each to pay attention to the corresponding alarm signal of force threshold subrange, Above-mentioned alarm signal includes：Acoustic warning signal, image alarm signal；In response to paying attention to force threshold corresponding to notice numerical value Scope, send to should notice threshold value subrange alarm signal.
In certain embodiments, the notice threshold range where the above-mentioned notice numerical value of above-mentioned judgement, according to above-mentioned note Meaning force threshold scope, which sends alarm signal, also to be included：During in response to notice numerical value not in above-mentioned notice threshold range, open Dynamic automatic driving mode.
Second aspect, the embodiment of the present application provide a kind of device for being used to send alarm signal, and the above-mentioned device includes： Transport condition data acquiring unit, for obtaining transport condition data in real time, above-mentioned transport condition data includes driver status Data, driving environment data and vehicle status data, above-mentioned driver status data include hand motion data, mouth action number According to, headwork data and eye motion data, above-mentioned driving environment data include lane line data, vehicle distances data, on Stating vehicle status data includes speed data, bearing data；Tired numerical value acquiring unit, for above-mentioned transport condition data to be led Enter the tired identification model of training in advance, obtain the tired numerical value of tired type and corresponding tired type, above-mentioned fatigue identification mould Type is used to determine tired type and tired numerical value by transport condition data, and above-mentioned tired numerical value is used to characterize corresponding tired type Degree；Notice numerical value acquiring unit, for above-mentioned tired type and tired numerical value to be imported to the attention force value of training in advance Computation model, the power that gains attention numerical value, above-mentioned attention force value computation model are used to note by tired type and tired numerical computations Meaning power numerical value, above-mentioned notice numerical value are used for the degree of fatigue for characterizing driver；Alarm Unit, for judging notice numerical value institute Notice threshold range, alarm signal is sent according to above-mentioned notice threshold range, above-mentioned notice threshold range passes through Default attention force threshold determines.
In certain embodiments, said apparatus includes tired identification model construction unit, for building tired identification model, Above-mentioned tired identification model construction unit includes：Data extract subelement, for respectively from driver status data acquisition system, driving Extraction time upper synchronous driver status data in environmental data set and vehicle status data set, driving environment data and Vehicle status data；Sign extracts subelement, for respectively to driver status data, driving environment data and vehicle-state number According to feature extraction is carried out, driver status characteristic, driving environment characteristic and vehicle-state characteristic are obtained；Fatigue Type determination unit, for determining tired type according to above-mentioned driver status characteristic, above-mentioned tired type include with It is at least one of lower：Eye closing type, type of yawning, sight offset type, type of making a phone call；Tired numerical computations subelement, use Tired numerical value is calculated in passing through driver status characteristic, driving environment characteristic and vehicle-state characteristic；Fatigue Identification model builds subelement, for using machine learning method, by driver status characteristic, driving environment characteristic With vehicle-state characteristic as inputting, using the tired numerical value of tired type and corresponding tired type as output, training Obtain tired identification model.
In certain embodiments, above-mentioned tired numerical computations subelement includes：Threshold range acquisition module, during for obtaining Between threshold range, distance threshold scope and data variable quantity threshold range, wherein, time threshold scope is included by pre-setting Multiple time threshold subranges that time threshold is formed, distance threshold scope are more including being made up of the distance threshold pre-set Individual distance threshold subrange, data variation amount threshold range include the more numbers being made up of the data variation amount threshold value pre-set According to variable quantity threshold value subrange, the corresponding time weighting value of each time threshold subrange, each distance threshold subrange pair Answer a distance weighting value, the corresponding data variation weighted value of each data variation amount threshold value subrange；Data measurement module, Tired type duration, driving environment for measuring every kind of tired type corresponding to driver status characteristic respectively is special Data variation value corresponding to distance change value corresponding to data and vehicle-state characteristic is levied, wherein, when tired type continues Between by specifying duration for acting corresponding to tired type to characterize corresponding to driver status characteristic, distance change Value is characterized by variable quantity of the driving environment characteristic within the unit interval, and data variation value passes through vehicle-state characteristic Characterized according to the variable quantity within the unit interval；Weighted value determining module, for determining tired type duration, distance change Time weighting value, the distance threshold scope of the time threshold subrange of time threshold scope corresponding to value and data changing value difference Distance threshold subrange distance weighting value and data variable quantity threshold range data variation amount threshold value subrange data Change weighted value；Tired Numerical Simulation Module, for asking the weighting of time weighting value, distance weighting value and data variation weighted value With the tired numerical value for obtaining corresponding to tired type.
In certain embodiments, said apparatus includes paying attention to force value computation model construction unit, pays attention to force value for building Computation model, above-mentioned attention force value computation model construction unit include：Data divide subelement, for will be tired according to tired type Tired numerical division in labor numerical value set is the tired numerical value subclass of corresponding tired type；It is single to notice that force component calculates son Member, for being respectively that every kind of tired type sets type weight value, by type weight value fatigue corresponding with the type weighted value Notice of the product of each tired numerical value in numerical value subclass as corresponding tired type when value is the tired numerical value Component；Notice that force value computation model builds subelement, for using machine learning method, using tired type and tired numerical value as Input, it will notice that force component as output, trains the force value computation model that gains attention.
In certain embodiments, above-mentioned Alarm Unit includes：Notice threshold range obtains subelement, pays attention to for obtaining Force threshold scope, above-mentioned notice threshold range include multiple attention force thresholds being made up of the attention force threshold pre-set Scope, each to pay attention to the corresponding alarm signal of force threshold subrange, above-mentioned alarm signal includes：Acoustic warning signal, image Alarm signal；Subelement is alerted, in response to paying attention to force threshold subrange corresponding to notice numerical value, sending to that should pay attention to The alarm signal of force threshold subrange.
In certain embodiments, above-mentioned Alarm Unit also includes：In response to notice numerical value not in above-mentioned attention force threshold In the range of when, start automatic driving mode.
The third aspect, the embodiment of the present application provide a kind of server, including：One or more processors；Memory, use In storing one or more programs, when said one or multiple programs are by said one or multiple computing devices so that on State the method for sending alarm signal that one or more processors perform above-mentioned first aspect.
Fourth aspect, the embodiment of the present application provide a kind of computer-readable recording medium, are stored thereon with computer journey Sequence, it is characterised in that the program realizes the method for sending alarm signal of above-mentioned first aspect when being executed by processor.
The method and device for being used to send alarm signal that the embodiment of the present application provides, obtains transport condition data in real time, Above-mentioned transport condition data includes driver status data, driving environment data and vehicle status data, realize to driver, Then transport condition data is imported tired identification model and obtains fatigue by the data of running environment and vehicle data to considering The tired numerical value of type and corresponding tired type, and the attention force value of tired type and tired numerical value importing training in advance is calculated Model, the power that gains attention numerical value, corresponding alarm signal is finally sent according to the notice threshold range where notice numerical value, Improve the accuracy rate of the degree of fatigue identification to driver.
Brief description of the drawings
By reading the detailed description made to non-limiting example made with reference to the following drawings, the application's is other Feature, objects and advantages will become more apparent upon：
Fig. 1 is that the application can apply to exemplary system architecture figure therein；
Fig. 2 is the flow chart for being used to send one embodiment of the method for alarm signal according to the application；
Fig. 3 is the schematic diagram for being used to send an application scenarios of the method for alarm signal according to the application；
Fig. 4 is the structural representation for being used to send one embodiment of the device of alarm signal according to the application；
Fig. 5 is adapted for the structural representation of the computer system of the server for realizing the embodiment of the present application.
Embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to Be easy to describe, illustrate only in accompanying drawing to about the related part of invention.
It should be noted that in the case where not conflicting, the feature in embodiment and embodiment in the application can phase Mutually combination.Describe the application in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 shows the method for being used to send alarm signal that can apply the application or the dress for sending alarm signal The exemplary system architecture 100 for the embodiment put.
As shown in figure 1, system architecture 100 can include the first camera 101, second camera 102, car-mounted terminal 103, Network 104 and server 105.Network 104 between car-mounted terminal 103 and server 105 provide communication link medium. Network 104 can include various connection types, such as wired, wireless communication link or fiber optic cables etc..
First camera 101 can be arranged on outside vehicle, for gathering the image during Current vehicle travels, then Current vehicle and the distance between other vehicles on travel, pedestrian are calculated according to the image of collection, can also be calculated Go out the distance between Current vehicle and lane line；Second camera 102 can be arranged on vehicle interior, for monitoring driver's Various actions, including hand motion, mouth action, headwork and eye motion etc., and by these action be converted to corresponding to Data；Car-mounted terminal 103 can receive the status data (example of the first camera 101, second camera 102 and Current vehicle Such as can be speed data, bearing data), and the data of reception are handled, and then sent out in driver tired driving Go out corresponding alarm signal.
Server 105 can be to provide at the server of various services, such as the data sent to car-mounted terminal 103 The server of reason, the data that server 105 can be sent to car-mounted terminal 103 carry out data processing, and the notice that will be obtained The data such as numerical value are sent to car-mounted terminal 103.
It should be noted that the method for being used to send alarm signal that the embodiment of the present application is provided is typically by car-mounted terminal 103 are performed, and correspondingly, the device for sending alarm signal is generally positioned in car-mounted terminal 103.
It should be understood that the first camera 101, second camera 102, car-mounted terminal 103, network 104 and service in Fig. 1 The number of device 105 is only schematical.According to needs are realized, can have any number of first camera 101, second to take the photograph As first 102, car-mounted terminal 103, network 104 and server 105.
With continued reference to Fig. 2, the stream for being used to send one embodiment of the method for alarm signal according to the application is shown Journey 200.The method for being used to send alarm signal comprises the following steps：
Step 201, transport condition data is obtained in real time.
In the present embodiment, electronic equipment thereon is run (such as shown in Fig. 1 for sending the method for alarm signal Car-mounted terminal 103) transport condition data can be received by wired connection mode or radio connection.Wherein, above-mentioned row Driver status data, driving environment data and vehicle status data, above-mentioned driver status number can be included by sailing status data According to hand motion data, mouth action data, headwork data and eye motion data, above-mentioned driving environment number can be included According to that can include lane line data, vehicle distances data, above-mentioned vehicle status data can include speed data, bearing data. It is pointed out that above-mentioned radio connection can include but is not limited to 3G/4G connections, WiFi connections, bluetooth connection, WiMAX connections, Zigbee connections, UWB (ultra wideband) connections and other it is currently known or in the future exploitation it is wireless Connected mode.
When driver drives vehicle traveling, driver status data, driving environment data and vehicle can be obtained in real time Status data.Wherein, driver status data can be by setting second camera 102 in the car to gather；Driving environment data It can be gathered by the first camera 101 being arranged on outside car；Vehicle status data can be obtained directly from vehicle.
Step 202, above-mentioned transport condition data is imported to the tired identification model of training in advance, obtains tired type and right Answer the tired numerical value of tired type.
After getting transport condition data, transport condition data can be imported to tired identification model, obtain tired type (can be represented by the form of information) and the tired numerical value of corresponding tired type.Wherein, above-mentioned tired identification model is used to lead to Cross transport condition data and determine tired type and tired numerical value, above-mentioned tired numerical value is used for the degree for characterizing corresponding tired type. Tired numerical value can be represented by diversified forms.For example, the span of tired numerical value can be 0 to 1.When tired numerical value is 0 When, it is believed that driver is not tired；When tired numerical value is 1, it is believed that driver is in major fatigue.Tired numerical value is also It can be represented by the other forms such as 0 to 100, no longer repeated one by one herein.
In some optional implementations of the present embodiment, the present embodiment method can include building tired identification model The step of, it may comprise steps of the step of above-mentioned structure fatigue identification model：
The first step, respectively from driver status data acquisition system, driving environment data acquisition system and vehicle status data set Synchronous driver status data, driving environment data and vehicle status data on extraction time.
When driver is in fatigue driving, not only driver itself can show some fatigue characteristics, meanwhile, vehicle It is abnormal to occur traveling because of the fatigue driving of driver.Therefore, can be respectively from driving when judging whether driver is tired Extraction time upper synchronous driver status in member's status data set, driving environment data acquisition system and vehicle status data set Data, driving environment data and vehicle status data.It should be noted that driver status data, driving environment data and car Status data necessary synchronization (identical on the time) in time, got up with to will pass through a variety of aggregation of data is to driver Accurate judgement is made in no fatigue driving.
Second step, feature extraction is carried out to driver status data, driving environment data and vehicle status data respectively, obtained To driver status characteristic, driving environment characteristic and vehicle-state characteristic.
What driver status data, driving environment data and vehicle status data typically directly gathered.In order to driving Whether fatigue driving is judged, it is necessary to enter respectively to driver status data, driving environment data and vehicle status data member Row feature extraction, obtain driver status characteristic, driving environment characteristic and vehicle-state characteristic.That is, drive Member's state characteristic, driving environment characteristic and vehicle-state characteristic are the number related to driver tired driving According to.For example, driver status characteristic can be the time value of eye closure；Driving environment characteristic can be vehicle away from From changing value；Vehicle-state characteristic can be the changing value of speed.Driver status characteristic, driving environment feature Data and vehicle-state characteristic can also be other kinds of data, specifically depending on being actually needed.
3rd step, tired type is determined according to above-mentioned driver status characteristic.
Driver status data can include hand motion data, mouth action data, the headwork data of driver With the data such as eye motion data.When there are a kind of data therein or a few class data in driver, corresponding action data It can be showed by driver status characteristic.In the event of certain a kind of or a few class data, then illustrate driver Tired type corresponding to showing.Above-mentioned tired type includes at least one of following：Eye closing type, type of yawning, sight Offset type, type of making a phone call.
4th step, calculated by driver status characteristic, driving environment characteristic and vehicle-state characteristic Tired numerical value.
After getting driver status characteristic, driving environment characteristic and vehicle-state characteristic, pass through one Fixed data processing method can obtain the tired numerical value of corresponding tired type.For example, driver status spy can be calculated respectively Sign data, driving environment characteristic and vehicle-state characteristic correspond to the numerical value of a certain tired type, and then weighting obtains The tired numerical value of the tired type.
5th step, using machine learning method, by driver status characteristic, driving environment characteristic and vehicle shape State characteristic obtains fatigue as input using the tired numerical value of tired type and corresponding tired type as output, training Identification model.
Tired identification model can be that technical staff is based on to substantial amounts of driver status characteristic, driving environment feature Data, vehicle-state characteristic, tired type and tired numerical value are counted and pre-establish, are stored with driver status Characteristic, driving environment characteristic, vehicle-state characteristic, tired type and tired numerical value corresponding relation correspondence Relation table；Can also be that technical staff is pre-set and is stored in above-mentioned electronic equipment based on the statistics to mass data , numerical computations are carried out to driver status characteristic, driving environment characteristic and vehicle-state characteristic to obtain Calculation formula for the result of calculation of the tired numerical value that characterizes tired type.For example, the calculation formula can be to driver State characteristic, driving environment characteristic and vehicle-state characteristic are weighted the formula of summation, obtained result It can be used for the tired numerical value for characterizing corresponding tired type.
In some optional implementations of the present embodiment, above by driver status characteristic, driving environment Characteristic and vehicle-state characteristic calculate tired numerical value and may comprise steps of：
The first step, obtain time threshold scope, distance threshold scope and data variable quantity threshold range.
Generally, time threshold scope, distance threshold scope and data variable quantity threshold range can be pre-set. Wherein, time threshold scope can include the multiple time threshold subranges being made up of the time threshold pre-set；Apart from threshold Value scope can include the multiple distance threshold subranges being made up of the distance threshold pre-set；Data variation amount threshold range The multiple data variation amount threshold value subranges being made up of the data variation amount threshold value pre-set can be included.Each time threshold The corresponding time weighting value of subrange；Each corresponding distance weighting value of distance threshold subrange；Each data variation amount The corresponding data variation weighted value of threshold value subrange.
Second step, when the tired type of every kind of tired type continues corresponding to measurement driver status characteristic respectively Between, data variation value corresponding to distance change value corresponding to driving environment characteristic and vehicle-state characteristic.
In order to determine various data and the correlation of fatigue driving by way of quantization, driver's shape can be measured respectively Distance change corresponding to tired type duration of every kind of tired type corresponding to state characteristic, driving environment characteristic Data variation value corresponding to value and vehicle-state characteristic.Wherein, the tired type duration can pass through driver status The duration acted corresponding to tired type is specified to characterize corresponding to characteristic, distance change value can be by driving ring Variable quantity of the border characteristic within the unit interval characterizes, and data variation value can be by vehicle-state characteristic in unit Variable quantity in time characterizes.
3rd step, it is determined that tired type duration, distance change value and time threshold corresponding to data changing value difference The time weighting value of the time threshold subrange of scope, the distance weighting value sum of the distance threshold subrange of distance threshold scope According to the data variation weighted value of the data variation amount threshold value subrange of variable quantity threshold range.
After obtaining tired type duration, distance change value and data changing value, by data comparison, it can obtain respectively Time weighting value, distance change value to the tired type duration in time threshold subrange corresponding to time threshold scope exist The distance weighting value and data changing value of distance threshold subrange corresponding to distance threshold scope are in data variation amount threshold range The data variation weighted value of corresponding data variation amount threshold value subrange.
4th step, corresponding tired class is obtained to time weighting value, distance weighting value and data variation weighted value weighted sum The tired numerical value of type.
Time weighting value, distance weighting value and data variation weighted value can not from the correlation of different tired types Together.Therefore, can be that time weighting value, distance weighting value and data variation weighted value set difference according to specific tired type Weights, sum, obtain again after the product of time weighting value, distance weighting value and data variation weighted value with respective weights is obtained To the tired numerical value of corresponding tired type.For example, time weighting value, distance weighting value and data variation weighted value can be respectively 0.1st, 0.3 and 0.5, the weights of corresponding a certain tired type can be 0.2,0.1 and 0.4 respectively, then to should tired type Tired numerical value can be 0.1 × 0.2+0.3 × 0.1+0.5 × 0.4=0.25.
Step 203, above-mentioned tired type and tired numerical value are imported to the attention force value computation model of training in advance, noted Meaning power numerical value.
After obtaining tired type and tired numerical value, tired type and tired numerical value can be imported and pay attention to force value computation model The power that gains attention numerical value.Wherein, above-mentioned attention force value computation model is used to pass through tired type and tired numerical computations notice Numerical value, above-mentioned notice numerical value are used for the degree of fatigue for characterizing driver.It is similar with tired numerical value, the value model of notice numerical value Enclosing can also be 0 to 1.When notice numerical value is 0, it is believed that driver's major fatigue；, can when notice numerical value is 1 To think that driver is not tired.Notice numerical value can also be represented by other forms, no longer be repeated one by one herein.
In some optional implementations of the present embodiment, the present embodiment method can include structure and notice that force value calculates The step of model, the step of above-mentioned structure pays attention to force value computation model, may comprise steps of：
The first step, the fatigue according to tired type by the tired numerical division in tired numerical value set for corresponding tired type Numerical value subclass.
Seen from the above description, different tired types corresponds to respective tired numerical value.Therefore, can be according to tired type Tired numerical value subclass by the tired numerical division in tired numerical value set for corresponding tired type.
Second step, respectively every kind of tired type sets type weight value, by type weight value and the type weighted value pair The product of each tired numerical value in the tired numerical value subclass answered is as corresponding tired type when value is the tired numerical value Attention force component.
Can be that every kind of tired type sets type weight according to the driving habit of different drivers and driving behavior Value, so that pointedly whether fatigue driving judges to driver according to the feature of driver.It is a certain tired when needing to obtain , can be by the type weight value of the tired type tired numerical value corresponding with the type weighted value during the attention class component of labor type Attention force component of the product of each tired numerical value in subclass as corresponding tired type when value is the tired numerical value. That is, how many tired numerical value is with regard to how many corresponding attention force component., can when a variety of tired types occurs in driver So that the attention force component of corresponding each tired type to be summed to obtain the notice numerical value of driver.For example, driver occurs Tired type A, B, C, there is tired numerical value 0.1,0.2 and 0.3, corresponding type power respectively in corresponding tired numerical value subclass Weight values can be respectively 0.2,0.2 and 0.6, then the notice numerical value of the driver is respectively 0.1,0.2 and 0.3 in tired numerical value When, can be 0.1 × 0.2+0.2 × 0.2+0.3 × 0.6=0.24.
3rd step, using machine learning method, using tired type and tired numerical value as input, will pay attention to force component as Output, train the force value computation model that gains attention.
Pay attention to force value computation model be technical staff be based on counting substantial amounts of tired type and tired numerical value and Mapping table pre-establishing, being stored with tired type, tired numerical value and the corresponding relation for paying attention to force component；Can also be Technical staff is pre-set based on the statistics to mass data and is stored in above-mentioned electronic equipment, and tired numerical value is carried out Numerical computations with obtain for characterize tired type attention force component result of calculation calculation formula.For example, the calculating is public Formula can be that the formula of summation is weighted to tired numerical value, and obtained result can be used for the attention for characterizing corresponding tired type Force component.Afterwards, the summation of the attention force component of each tired type is gained attention power numerical value.
It should be noted that above-mentioned tired identification model and attention force value computation model can be beforehand through servers 105 training obtain.
Step 204, judge the notice threshold range where notice numerical value, sent according to above-mentioned notice threshold range Alarm signal.
Multiple attention force thresholds can be previously provided with car-mounted terminal 103, is made up of and pays attention to multiple attention force thresholds Force threshold scope, i.e. above-mentioned notice threshold range is determined by default attention force threshold.It is then possible to pay attention to be each Force threshold scope sets different alarm signals, and different alarm signals can embody the current notice of driver.Work as attention When power numerical value belongs to some notice threshold range, alarm signal corresponding to the notice threshold range is sent.
In some optional implementations of the present embodiment, the attention force threshold model where above-mentioned judgement notice numerical value Enclose, sending alarm signal according to above-mentioned notice threshold range may comprise steps of：
The first step, obtain notice threshold range.
, it is necessary to get notice threshold range first before alarm signal is sent.Wherein, above-mentioned notice threshold range The multiple attention force threshold subranges formed including the attention force threshold by pre-setting, it is each to notice that force threshold subrange is corresponding One alarm signal, above-mentioned alarm signal can include：Acoustic warning signal, image alarm signal etc..
Second step, in response to corresponding to notice numerical value pay attention to force threshold subrange, send to should notice threshold value son The alarm signal of scope.
For example it is to be noted that force threshold scope can be by noticing that force threshold 0,0.5 and 1 is formed.Where notice numerical value Notice threshold range be 0.5 to 1 between when, it is believed that the notice of driver is concentrated, and corresponding, driver is not on Fatigue driving state, alarm signal can be that voice message driver pays attention to driving safety；Attention where notice numerical value Force threshold scope be 0 to 0.5 between when, it is believed that driver's is absent minded, corresponding, and driver is not on gently Micro- fatigue driving state, alarm signal can prompt driver to pay attention to driving safety for louder volume voice, picture signal.
In some optional implementations of the present embodiment, the notice threshold where the above-mentioned above-mentioned notice numerical value of judgement It is worth scope, sending alarm signal according to above-mentioned notice threshold range can also include：When notice numerical value is not in above-mentioned attention When in the range of force threshold, start automatic driving mode.
There is Function for Automatic Pilot in response to vehicle, and when notice numerical value is not in above-mentioned notice threshold range, car Mounted terminal 103 can be with automatic start automatic driving mode.For example, when the notice threshold range where notice numerical value is 0 When, it is believed that driver scatters and disappears notice, and now, the can of car-mounted terminal 103 directly initiates the automatic driving mode of vehicle, To improve the security of traveling.
With continued reference to Fig. 3, Fig. 3 is one that is used to send the application scenarios of the method for alarm signal according to the present embodiment Schematic diagram.In Fig. 3 application scenarios, car-mounted terminal 103 gets the vehicle status data of vehicle, the first camera 101 is adopted After the transport condition datas such as the driver status data of driving environment data and second camera 102 collection of collection, shape will be travelled State data import tired identification model, obtain the tired numerical value of the current tired type of driver and corresponding tired type；Then, Tired type and tired numerical value are imported and pay attention to force value computation model, the power that gains attention numerical value；Finally, notice numerical value institute is judged Notice threshold range, alarm signal " please noting driving safety " is sent according to above-mentioned notice threshold range.
The method that above-described embodiment of the application provides can obtain transport condition data, above-mentioned transport condition data in real time Including driver status data, driving environment data and vehicle status data, realize to driver, running environment and vehicle number According to data to considering, transport condition data is then imported into tired identification model and obtains tired type and corresponding tired class The tired numerical value of type, and tired type and tired numerical value are imported to the attention force value computation model of training in advance, gain attention power Numerical value, corresponding alarm signal is finally sent according to the notice threshold range where notice numerical value, improved to driver Degree of fatigue identification accuracy rate.
With further reference to Fig. 4, as the realization to method shown in above-mentioned each figure, it is used to send announcement this application provides one kind One embodiment of the device of alert signal, the device embodiment is corresponding with the embodiment of the method shown in Fig. 4, and the device specifically may be used With applied in various electronic equipments.
As shown in figure 4, the device 400 for being used to send alarm signal of the present embodiment can include：Transport condition data obtains Take unit 401, tired numerical value acquiring unit 402, notice numerical value acquiring unit 403 and Alarm Unit 404.Wherein, shape is travelled State data capture unit 401 be used in real time obtain transport condition data, above-mentioned transport condition data include driver status data, Driving environment data and vehicle status data, above-mentioned driver status data include hand motion data, mouth action data, head Portion's action data and eye motion data, above-mentioned driving environment data include lane line data, vehicle distances data, above-mentioned vehicle Status data includes speed data, bearing data；Tired numerical value acquiring unit 402 is used to import above-mentioned transport condition data in advance The tired identification model first trained, obtains the tired numerical value of tired type and corresponding tired type, and above-mentioned tired identification model is used In determining tired type and tired numerical value by transport condition data, above-mentioned tired numerical value is used for the journey for characterizing corresponding tired type Degree；Notice numerical value acquiring unit 403 is by above-mentioned tired type and tired numerical value to be imported based on the attention force value of training in advance Model, the power that gains attention numerical value are calculated, above-mentioned attention force value computation model is used to pay attention to by tired type and tired numerical computations Power numerical value, above-mentioned notice numerical value are used for the degree of fatigue for characterizing driver；Alarm Unit 404 is used to judge notice numerical value institute Notice threshold range, alarm signal is sent according to above-mentioned notice threshold range, above-mentioned notice threshold range passes through Default attention force threshold determines.
In some optional implementations of the present embodiment, the device 400 for being used to send alarm signal of the present embodiment Tired identification model construction unit (not shown) can be included, for building tired identification model, above-mentioned fatigue identification mould Type construction unit can include：Data extraction subelement (not shown), feature extraction subelement (not shown), fatigue Type determination unit (not shown), tired numerical computations subelement (not shown) and tired identification model structure Unit (not shown).Wherein, data extraction subelement is used for respectively from driver status data acquisition system, driving environment data Extraction time upper synchronous driver status data, driving environment data and vehicle-state in set and vehicle status data set Data；Feature extraction subelement is used to carry out spy to driver status data, driving environment data and vehicle status data respectively Sign extraction, obtains driver status characteristic, driving environment characteristic and vehicle-state characteristic；Tired type determines Subelement is used to determine tired type according to above-mentioned driver status characteristic, and above-mentioned tired type includes following at least one ：Eye closing type, type of yawning, sight offset type, type of making a phone call；Tired numerical computations subelement is used to pass through driving Member's state characteristic, driving environment characteristic and vehicle-state characteristic calculate tired numerical value；Tired identification model structure Build subelement to be used to utilize machine learning method, by driver status characteristic, driving environment characteristic and vehicle-state Characteristic obtains tired knowledge as input using the tired numerical value of tired type and corresponding tired type as output, training Other model.
In some optional implementations of the present embodiment, above-mentioned tired numerical computations subelement can include：Threshold value Scope acquisition module (not shown), data measurement module (not shown), weighted value determining module (not shown) With tired Numerical Simulation Module (not shown).Wherein, threshold range acquisition module is used to obtain time threshold scope, distance Threshold range and data variable quantity threshold range, wherein, time threshold scope includes what is be made up of the time threshold pre-set Multiple time threshold subranges, distance threshold scope include the sub- model of multiple distance thresholds being made up of the distance threshold pre-set Enclose, data variation amount threshold range includes multiple data variation amount threshold values being made up of the data variation amount threshold value pre-set Scope, the corresponding time weighting value of each time threshold subrange, the corresponding distance weighting of each distance threshold subrange Value, the corresponding data variation weighted value of each data variation amount threshold value subrange；Data measurement module is used for measurement respectively and driven Corresponding to tired type duration of every kind of tired type corresponding to the person's of sailing state characteristic, driving environment characteristic away from From data variation value corresponding to changing value and vehicle-state characteristic, wherein, the tired type duration passes through driver's shape The duration acted corresponding to tired type is specified to characterize corresponding to state characteristic, distance change value passes through driving environment Variable quantity of the characteristic within the unit interval characterizes, and data variation value is by vehicle-state characteristic within the unit interval Variable quantity characterize；Weighted value determining module is used to determine tired type duration, distance change value and data changing value The time weighting value of the time threshold subrange of time threshold scope, the sub- model of the distance threshold of distance threshold scope corresponding to respectively The distance weighting value and the data variation weighted value of the data variation amount threshold value subrange of data variable quantity threshold range enclosed；Fatigue Numerical Simulation Module is used to obtain time weighting value, distance weighting value and data variation weighted value weighted sum corresponding tired class The tired numerical value of type.
In some optional implementations of the present embodiment, the device 400 for being used to send alarm signal of the present embodiment It can include paying attention to force value computation model construction unit (not shown), pay attention to force value computation model, above-mentioned note for building Meaning force value computation model construction unit can include：Data division subelement (not shown), notice that force component calculates son list First (not shown) and attention force value computation model structure subelement (not shown).Wherein, data division subelement is used In the tired numerical value subclass according to tired type by the tired numerical division in tired numerical value set for corresponding tired type；Note Force component computation subunit of anticipating is used for respectively every kind of tired type and sets type weight value, and type weight value and the type are weighed The product of each tired numerical value in tired numerical value subclass corresponding to weight values is the fatigue in value as corresponding tired type Attention force component during numerical value；Notice that force value computation model structure subelement is used to utilize machine learning method, by tired type With tired numerical value as inputting, it will notice that force component as output, trains the force value computation model that gains attention.
In some optional implementations of the present embodiment, above-mentioned Alarm Unit 404 can include：Pay attention to force threshold model Enclose and obtain subelement (not shown) and alarm subelement (not shown).Wherein, it is single to obtain son for notice threshold range Member is used to obtain notice threshold range, and above-mentioned notice threshold range is more including being made up of the attention force threshold pre-set Individual attention force threshold subrange, each to pay attention to the corresponding alarm signal of force threshold subrange, above-mentioned alarm signal includes：Sound Alarm signal, image alarm signal；Subelement is alerted to be used in response to paying attention to force threshold subrange, hair corresponding to notice numerical value Go out to should notice threshold value subrange alarm signal.
In some optional implementations of the present embodiment, above-mentioned Alarm Unit 404 can also include：In response to paying attention to When power numerical value is not in above-mentioned notice threshold range, start automatic driving mode.
The present embodiment additionally provides a kind of server, including：One or more processors；Memory, for storing one Or multiple programs, when said one or multiple programs are by said one or multiple computing devices so that said one is more The above-mentioned method for sending alarm signal of individual computing device.
The present embodiment additionally provides a kind of computer-readable recording medium, is stored thereon with computer program, the program quilt The above-mentioned method for sending alarm signal is realized during computing device.
Below with reference to Fig. 5, it illustrates suitable for for realizing the computer system 500 of the server of the embodiment of the present application Structural representation.Server shown in Fig. 5 is only an example, should not be to the function and use range band of the embodiment of the present application Carry out any restrictions.
As shown in figure 5, computer system 500 includes CPU (CPU) 501, it can be read-only according to being stored in Program in memory (ROM) 502 or be loaded into program in random access storage device (RAM) 503 from storage part 508 and Perform various appropriate actions and processing.In RAM 503, also it is stored with system 500 and operates required various programs and data. CPU 501, ROM 502 and RAM 503 are connected with each other by bus 504.Input/output (I/O) interface 505 is also connected to always Line 504.
I/O interfaces 505 are connected to lower component：Importation 506 including keyboard, mouse etc.；Penetrated including such as negative electrode The output par, c 507 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage part 508 including hard disk etc.； And the communications portion 509 of the NIC including LAN card, modem etc..Communications portion 509 via such as because The network of spy's net performs communication process.Driver 510 is also according to needing to be connected to I/O interfaces 505.Detachable media 511, such as Disk, CD, magneto-optic disk, semiconductor memory etc., it is arranged on as needed on driver 510, in order to read from it Computer program be mounted into as needed storage part 508.
Especially, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description Software program.For example, embodiment of the disclosure includes a kind of computer program product, it includes being carried on computer-readable medium On computer program, the computer program include be used for execution flow chart shown in method program code.In such reality To apply in example, the computer program can be downloaded and installed by communications portion 509 from network, and/or from detachable media 511 are mounted.When the computer program is performed by CPU (CPU) 501, perform what is limited in the present processes Above-mentioned function.
It should be noted that the above-mentioned computer-readable medium of the application can be computer-readable signal media or meter Calculation machine readable storage medium storing program for executing either the two any combination.Computer-readable recording medium for example can be --- but not Be limited to --- electricity, magnetic, optical, electromagnetic, system, device or the device of infrared ray or semiconductor, or it is any more than combination.Meter The more specifically example of calculation machine readable storage medium storing program for executing can include but is not limited to：Electrical connection with one or more wires, just Take formula computer disk, hard disk, random access storage device (RAM), read-only storage (ROM), erasable type and may be programmed read-only storage Device (EPROM or flash memory), optical fiber, portable compact disc read-only storage (CD-ROM), light storage device, magnetic memory device, Or above-mentioned any appropriate combination.In this application, computer-readable recording medium can any include or store journey The tangible medium of sequence, the program can be commanded the either device use or in connection of execution system, device.And at this In application, computer-readable signal media can include in a base band or as carrier wave a part propagation data-signal, Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can Any computer-readable medium beyond storage medium is read, the computer-readable medium, which can send, propagates or transmit, to be used for By instruction execution system, device either device use or program in connection.Included on computer-readable medium Program code can be transmitted with any appropriate medium, be included but is not limited to：Wirelessly, electric wire, optical cable, RF etc., or it is above-mentioned Any appropriate combination.
Flow chart and block diagram in accompanying drawing, it is illustrated that according to the system of the various embodiments of the application, method and computer journey Architectural framework in the cards, function and the operation of sequence product.At this point, each square frame in flow chart or block diagram can generation The part of one module of table, program segment or code, the part of the module, program segment or code include one or more use In the executable instruction of logic function as defined in realization.It should also be noted that marked at some as in the realization replaced in square frame The function of note can also be with different from the order marked in accompanying drawing generation.For example, two square frames succeedingly represented are actually It can perform substantially in parallel, they can also be performed in the opposite order sometimes, and this is depending on involved function.Also to note Meaning, the combination of each square frame and block diagram in block diagram and/or flow chart and/or the square frame in flow chart can be with holding Function as defined in row or the special hardware based system of operation are realized, or can use specialized hardware and computer instruction Combination realize.
Being described in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard The mode of part is realized.Described unit can also be set within a processor, for example, can be described as：A kind of processor bag Include transport condition data acquiring unit, tired numerical value acquiring unit, notice numerical value acquiring unit and Alarm Unit.Wherein, this The title of a little units does not form the restriction to the unit in itself under certain conditions, for example, Alarm Unit can also be described For " being used for the unit for sending alarm signal ".
As on the other hand, present invention also provides a kind of computer-readable medium, the computer-readable medium can be Included in device described in above-described embodiment；Can also be individualism, and without be incorporated the device in.Above-mentioned calculating Machine computer-readable recording medium carries one or more program, when said one or multiple programs are performed by the device so that should Device：Transport condition data is obtained in real time, and above-mentioned transport condition data includes driver status data, driving environment data and car Status data, above-mentioned driver status data include hand motion data, mouth action data, headwork data and eye Action data, above-mentioned driving environment data include lane line data, vehicle distances data, and above-mentioned vehicle status data includes speed Data, bearing data；Above-mentioned transport condition data is imported to the tired identification model of training in advance, obtains tired type and correspondingly The tired numerical value of tired type, above-mentioned tired identification model are used to determine tired type and tired number by transport condition data Value, above-mentioned tired numerical value are used for the degree for characterizing corresponding tired type；Above-mentioned tired type and tired numerical value are imported into instruction in advance Experienced attention force value computation model, the power that gains attention numerical value, above-mentioned attention force value computation model are used for by tired type and tired Labor numerical computations notice numerical value, above-mentioned notice numerical value are used for the degree of fatigue for characterizing driver；Judge notice numerical value institute Notice threshold range, alarm signal is sent according to above-mentioned notice threshold range, above-mentioned notice threshold range passes through Default attention force threshold determines.
Above description is only the preferred embodiment of the application and the explanation to institute's application technology principle.People in the art Member should be appreciated that invention scope involved in the application, however it is not limited to the technology that the particular combination of above-mentioned technical characteristic forms Scheme, while should also cover in the case where not departing from foregoing invention design, carried out by above-mentioned technical characteristic or its equivalent feature The other technical schemes for being combined and being formed.Such as features described above has similar work(with (but not limited to) disclosed herein The technical scheme that the technical characteristic of energy is replaced mutually and formed.

Claims (14)

  1. A kind of 1. method for sending alarm signal, it is characterised in that methods described includes：
    Transport condition data is obtained in real time, and the transport condition data includes driver status data, driving environment data and car Status data, the driver status data include hand motion data, mouth action data, headwork data and eye Action data, the driving environment data include lane line data, vehicle distances data, and the vehicle status data includes speed Data, bearing data；
    The transport condition data is imported to the tired identification model of training in advance, tired type is obtained and corresponds to tired type Tired numerical value, the tired identification model are used to determine tired type and tired numerical value, the fatigue by transport condition data Numerical value is used for the degree for characterizing corresponding tired type；
    The tired type and tired numerical value are imported to the attention force value computation model of training in advance, the power that gains attention numerical value, institute State and notice that force value computation model is used to be used for by tired type and tired numerical computations notice numerical value, the notice numerical value Characterize the degree of fatigue of driver；
    Judge the notice threshold range where notice numerical value, alarm signal is sent according to the notice threshold range, institute Notice threshold range is stated to determine by default attention force threshold.
  2. 2. according to the method for claim 1, it is characterised in that the step of methods described includes building tired identification model, The step of structure tired identification model, includes：
    Extraction time is upper same from driver status data acquisition system, driving environment data acquisition system and vehicle status data set respectively Driver status data, driving environment data and the vehicle status data of step；
    Feature extraction is carried out to driver status data, driving environment data and vehicle status data respectively, obtains driver's shape State characteristic, driving environment characteristic and vehicle-state characteristic；
    Tired type is determined according to the driver status characteristic, the tired type includes at least one of following：Close one's eyes Type, type of yawning, sight offset type, type of making a phone call；
    Tired numerical value is calculated by driver status characteristic, driving environment characteristic and vehicle-state characteristic；
    Using machine learning method, by driver status characteristic, driving environment characteristic and vehicle-state characteristic As input, tired identification model is obtained using the tired numerical value of tired type and corresponding tired type as output, training.
  3. 3. according to the method for claim 2, it is characterised in that described to pass through driver status characteristic, driving environment Characteristic and vehicle-state characteristic, which calculate tired numerical value, to be included：
    Obtain time threshold scope, distance threshold scope and data variable quantity threshold range, wherein, time threshold scope include by Multiple time threshold subranges that the time threshold pre-set is formed, distance threshold scope are included by pre-setting apart from threshold It is worth the multiple distance threshold subranges formed, data variation amount threshold range includes the data variation amount threshold value structure by pre-setting Into multiple data variation amount threshold value subranges, the corresponding time weighting value of each time threshold subrange, each apart from threshold It is worth the corresponding distance weighting value of subrange, the corresponding data variation weighted value of each data variation amount threshold value subrange；
    The tired type duration of every kind of tired type, driving environment spy are measured corresponding to driver status characteristic respectively Data variation value corresponding to distance change value corresponding to data and vehicle-state characteristic is levied, wherein, when tired type continues Between by specifying duration for acting corresponding to tired type to characterize corresponding to driver status characteristic, distance change Value is characterized by variable quantity of the driving environment characteristic within the unit interval, and data variation value passes through vehicle-state characteristic Characterized according to the variable quantity within the unit interval；
    It is determined that the time threshold of tired type duration, distance change value and time threshold scope corresponding to data changing value difference The distance weighting value and data variable quantity threshold value of the time weighting value of value subrange, the distance threshold subrange of distance threshold scope The data variation weighted value of the data variation amount threshold value subrange of scope；
    The tired number of corresponding tired type is obtained to time weighting value, distance weighting value and data variation weighted value weighted sum Value.
  4. 4. according to the method for claim 1, it is characterised in that methods described includes the step that structure pays attention to force value computation model Suddenly, the step of structure pays attention to force value computation model includes：
    Tired numerical value subclass according to tired type by the tired numerical division in tired numerical value set for corresponding tired type；
    Respectively every kind of tired type sets type weight value, by type weight value tired numerical value corresponding with the type weighted value Attention force component of the product of each tired numerical value in subclass as corresponding tired type when value is the tired numerical value；
    Using machine learning method, using tired type and tired numerical value as input, it will notice that force component as output, is trained To attention force value computation model.
  5. 5. according to the method for claim 1, it is characterised in that the attention force threshold model judged where notice numerical value Enclose, sending alarm signal according to the notice threshold range includes：
    Notice threshold range is obtained, the notice threshold range is multiple including being made up of the attention force threshold pre-set Pay attention to force threshold subrange, each to pay attention to the corresponding alarm signal of force threshold subrange, the alarm signal includes：Sound is accused Alert signal, image alarm signal；
    In response to corresponding to notice numerical value pay attention to force threshold subrange, send to should notice threshold value subrange alarm believe Number.
  6. 6. according to the method for claim 5, it is characterised in that the notice threshold judged where the notice numerical value It is worth scope, sending alarm signal according to the notice threshold range also includes：
    During in response to notice numerical value not in the notice threshold range, start automatic driving mode.
  7. 7. a kind of device for being used to send alarm signal, it is characterised in that described device includes：
    Transport condition data acquiring unit, for obtaining transport condition data in real time, the transport condition data includes driver Status data, driving environment data and vehicle status data, the driver status data include hand motion data, mouth moves Make data, headwork data and eye motion data, the driving environment data include lane line data, vehicle distances number According to the vehicle status data includes speed data, bearing data；
    Tired numerical value acquiring unit, for the transport condition data to be imported to the tired identification model of training in advance, obtain tired The tired numerical value of labor type and corresponding tired type, the tired identification model are used to determine tired class by transport condition data Type and tired numerical value, the tired numerical value are used for the degree for characterizing corresponding tired type；
    Notice numerical value acquiring unit, the attention force value for the tired type and tired numerical value to be imported to training in advance calculate Model, the power that gains attention numerical value, the attention force value computation model are used to pass through tired type and tired numerical computations notice Numerical value, the notice numerical value are used for the degree of fatigue for characterizing driver；
    Alarm Unit, for judging the notice threshold range where notice numerical value, sent out according to the notice threshold range Go out alarm signal, the notice threshold range is determined by default attention force threshold.
  8. 8. device according to claim 7, it is characterised in that described device includes tired identification model construction unit, uses In building tired identification model, the tired identification model construction unit includes：
    Data extract subelement, for respectively from driver status data acquisition system, driving environment data acquisition system and vehicle-state number According to extraction time in set upper synchronous driver status data, driving environment data and vehicle status data；
    Feature extraction subelement, driver status data, driving environment data and vehicle status data are carried out for respectively special Sign extraction, obtains driver status characteristic, driving environment characteristic and vehicle-state characteristic；
    Tired type determination unit, for determining tired type, the tired class according to the driver status characteristic Type includes at least one of following：Eye closing type, type of yawning, sight offset type, type of making a phone call；
    Tired numerical computations subelement, for passing through driver status characteristic, driving environment characteristic and vehicle-state Characteristic calculates tired numerical value；
    Tired identification model builds subelement, for using machine learning method, by driver status characteristic, driving environment Characteristic and vehicle-state characteristic are as input, using the tired numerical value of tired type and corresponding tired type as defeated Go out, training obtains tired identification model.
  9. 9. device according to claim 8, it is characterised in that the tired numerical computations subelement includes：
    Threshold range acquisition module, for obtaining time threshold scope, distance threshold scope and data variable quantity threshold range, its In, time threshold scope includes the multiple time threshold subranges being made up of the time threshold pre-set, distance threshold scope The multiple distance threshold subranges formed including the distance threshold by pre-setting, data variation amount threshold range are included by advance Multiple data variation amount threshold value subranges that the data variation amount threshold value of setting is formed, each time threshold subrange are corresponding one Time weighting value, the corresponding distance weighting value of each distance threshold subrange, each data variation amount threshold value subrange are corresponding One data variation weighted value；
    Data measurement module, the tired type for measuring every kind of tired type corresponding to driver status characteristic respectively are held Data variation value corresponding to distance change value corresponding to continuous time, driving environment characteristic and vehicle-state characteristic, its In, the tired type duration by corresponding to driver status characteristic specify tired type corresponding to act it is lasting when Between characterize, distance change value is characterized by variable quantity of the driving environment characteristic within the unit interval, data variation value Characterized by variable quantity of the vehicle-state characteristic within the unit interval；
    Weighted value determining module, corresponding to determining tired type duration, distance change value and data changing value respectively The time weighting value of the time threshold subrange of time threshold scope, the distance of the distance threshold subrange of distance threshold scope power The data variation weighted value of the data variation amount threshold value subrange of weight values and data variable quantity threshold range；
    Tired Numerical Simulation Module, for being obtained to time weighting value, distance weighting value and data variation weighted value weighted sum The tired numerical value of corresponding tired type.
  10. 10. device according to claim 7, it is characterised in that described device includes noticing that force value computation model structure is single Member, pay attention to force value computation model for building, the attention force value computation model construction unit includes：
    Data divide subelement, for the tired numerical division in tired numerical value set to be corresponded into tired class according to tired type The tired numerical value subclass of type；
    Pay attention to force component computation subunit, for being respectively that every kind of tired type sets type weight value, by type weight value with The product of each tired numerical value in tired numerical value subclass corresponding to the type weighted value is as corresponding tired type in value For the tired numerical value when attention force component；
    Notice that force value computation model builds subelement, for using machine learning method, using tired type and tired numerical value as Input, it will notice that force component as output, trains the force value computation model that gains attention.
  11. 11. device according to claim 7, it is characterised in that the Alarm Unit includes：
    Notice threshold range obtains subelement, for obtaining notice threshold range, the notice threshold range include by Multiple attention force threshold subranges that the attention force threshold pre-set is formed, it is each to pay attention to the corresponding announcement of force threshold subrange Alert signal, the alarm signal include：Acoustic warning signal, image alarm signal；
    Subelement is alerted, in response to paying attention to force threshold subrange corresponding to notice numerical value, sending to should notice threshold It is worth the alarm signal of subrange.
  12. 12. device according to claim 11, it is characterised in that the Alarm Unit also includes：
    During in response to notice numerical value not in the notice threshold range, start automatic driving mode.
  13. 13. a kind of server, including：
    One or more processors；
    Memory, for storing one or more programs,
    When one or more of programs are by one or more of computing devices so that one or more of processors Any described method in perform claim requirement 1 to 6.
  14. 14. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the program is by processor The method as described in any in claim 1 to 6 is realized during execution.
CN201711140773.XA 2017-11-16 2017-11-16 Method and device for sending alarm signal Active CN107742399B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201711140773.XA CN107742399B (en) 2017-11-16 2017-11-16 Method and device for sending alarm signal
PCT/CN2018/099158 WO2019095733A1 (en) 2017-11-16 2018-08-07 Method and device used to emit alarm signals

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711140773.XA CN107742399B (en) 2017-11-16 2017-11-16 Method and device for sending alarm signal

Publications (2)

Publication Number Publication Date
CN107742399A true CN107742399A (en) 2018-02-27
CN107742399B CN107742399B (en) 2022-02-22

Family

ID=61234884

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711140773.XA Active CN107742399B (en) 2017-11-16 2017-11-16 Method and device for sending alarm signal

Country Status (2)

Country Link
CN (1) CN107742399B (en)
WO (1) WO2019095733A1 (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109191789A (en) * 2018-10-18 2019-01-11 斑马网络技术有限公司 Method for detecting fatigue driving, device, terminal and storage medium
CN109582529A (en) * 2018-09-29 2019-04-05 阿里巴巴集团控股有限公司 A kind of setting method and device of alarm threshold value
WO2019095733A1 (en) * 2017-11-16 2019-05-23 百度在线网络技术（北京）有限公司 Method and device used to emit alarm signals
CN110191011A (en) * 2019-04-15 2019-08-30 厦门科灿信息技术有限公司 Smart machine monitoring method, device and equipment based on data center's monitoring system
CN110458191A (en) * 2019-07-05 2019-11-15 中国平安财产保险股份有限公司 Fatigue state judgment method, device, computer equipment and storage medium
CN111062300A (en) * 2019-12-11 2020-04-24 深圳市赛梅斯凯科技有限公司 Driving state detection method, device, equipment and computer readable storage medium
CN112365680A (en) * 2020-10-29 2021-02-12 福信富通科技股份有限公司 Active safety early warning method and system based on AI identification
CN113034895A (en) * 2021-02-04 2021-06-25 招商局公路网络科技控股股份有限公司 ETC portal system, and expressway fatigue driving early warning method and device
CN113460060A (en) * 2021-08-20 2021-10-01 武汉霖汐科技有限公司 Driver fatigue degree evaluation system, control method, and storage medium
CN113744499A (en) * 2021-08-12 2021-12-03 科大讯飞股份有限公司 Fatigue early warning method, glasses, system and computer readable storage medium

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110705502B (en) * 2019-10-14 2023-07-28 首约科技(北京)有限公司 Optimization method of driver monitoring equipment
CN114312800B (en) * 2022-02-14 2022-11-15 深圳市发掘科技有限公司 Vehicle safe driving method and device, computer equipment and storage medium

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6927694B1 (en) * 2001-08-20 2005-08-09 Research Foundation Of The University Of Central Florida Algorithm for monitoring head/eye motion for driver alertness with one camera
CN102509418A (en) * 2011-10-11 2012-06-20 东华大学 Fatigue driving estimation and early-warning method and device of multi-sensor information fusion
CN103606245A (en) * 2013-11-08 2014-02-26 北京工业大学 Fatigue driving detection early warning system based on Bluetooth brain wave earphone and Android mobile phone
CN104112334A (en) * 2013-04-16 2014-10-22 百度在线网络技术（北京）有限公司 Fatigue driving early warning method and fatigue driving early warning system
CN104408879A (en) * 2014-11-19 2015-03-11 湖南工学院 Method, device and system for processing fatigue driving early warning
CN104794855A (en) * 2014-01-22 2015-07-22 径卫视觉科技（上海）有限公司 Driver's attention comprehensive assessment system
CN105922993A (en) * 2015-02-27 2016-09-07 本田技研工业株式会社 Attention calling system for vehicle
WO2016200924A1 (en) * 2015-06-08 2016-12-15 REM SAFE Technologies, Inc. Situational awareness analysis and fatigue management system

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107742399B (en) * 2017-11-16 2022-02-22 百度在线网络技术（北京）有限公司 Method and device for sending alarm signal

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6927694B1 (en) * 2001-08-20 2005-08-09 Research Foundation Of The University Of Central Florida Algorithm for monitoring head/eye motion for driver alertness with one camera
CN102509418A (en) * 2011-10-11 2012-06-20 东华大学 Fatigue driving estimation and early-warning method and device of multi-sensor information fusion
CN104112334A (en) * 2013-04-16 2014-10-22 百度在线网络技术（北京）有限公司 Fatigue driving early warning method and fatigue driving early warning system
CN103606245A (en) * 2013-11-08 2014-02-26 北京工业大学 Fatigue driving detection early warning system based on Bluetooth brain wave earphone and Android mobile phone
CN104794855A (en) * 2014-01-22 2015-07-22 径卫视觉科技（上海）有限公司 Driver's attention comprehensive assessment system
CN104408879A (en) * 2014-11-19 2015-03-11 湖南工学院 Method, device and system for processing fatigue driving early warning
CN105922993A (en) * 2015-02-27 2016-09-07 本田技研工业株式会社 Attention calling system for vehicle
WO2016200924A1 (en) * 2015-06-08 2016-12-15 REM SAFE Technologies, Inc. Situational awareness analysis and fatigue management system

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019095733A1 (en) * 2017-11-16 2019-05-23 百度在线网络技术（北京）有限公司 Method and device used to emit alarm signals
CN109582529A (en) * 2018-09-29 2019-04-05 阿里巴巴集团控股有限公司 A kind of setting method and device of alarm threshold value
CN109191789A (en) * 2018-10-18 2019-01-11 斑马网络技术有限公司 Method for detecting fatigue driving, device, terminal and storage medium
CN110191011A (en) * 2019-04-15 2019-08-30 厦门科灿信息技术有限公司 Smart machine monitoring method, device and equipment based on data center's monitoring system
CN110458191A (en) * 2019-07-05 2019-11-15 中国平安财产保险股份有限公司 Fatigue state judgment method, device, computer equipment and storage medium
CN110458191B (en) * 2019-07-05 2024-04-12 中国平安财产保险股份有限公司 Fatigue state judging method, device, computer equipment and storage medium
CN111062300A (en) * 2019-12-11 2020-04-24 深圳市赛梅斯凯科技有限公司 Driving state detection method, device, equipment and computer readable storage medium
CN112365680A (en) * 2020-10-29 2021-02-12 福信富通科技股份有限公司 Active safety early warning method and system based on AI identification
CN113034895A (en) * 2021-02-04 2021-06-25 招商局公路网络科技控股股份有限公司 ETC portal system, and expressway fatigue driving early warning method and device
CN113744499A (en) * 2021-08-12 2021-12-03 科大讯飞股份有限公司 Fatigue early warning method, glasses, system and computer readable storage medium
CN113460060A (en) * 2021-08-20 2021-10-01 武汉霖汐科技有限公司 Driver fatigue degree evaluation system, control method, and storage medium

Also Published As

Publication number Publication date
WO2019095733A1 (en) 2019-05-23
CN107742399B (en) 2022-02-22

Similar Documents

Publication Publication Date Title
CN107742399A (en) For sending the method and device of alarm signal
CN104882001B (en) Driving behavior monitoring method based on drive recorder, apparatus and system
CN109389068A (en) The method and apparatus of driving behavior for identification
CN110782111B (en) Risk assessment method and system
CN109190487A (en) Face Emotion identification method, apparatus, computer equipment and storage medium
US11003926B2 (en) Method and apparatus for recognizing boundary of traffic sign
JP2018508090A (en) System and method for adapting user interface to user attention and driving conditions
CN109584567A (en) Traffic management method based on bus or train route collaboration
CN111846046B (en) System, method and device for detecting safety of bicycle
CN106652645A (en) Vehicle driving training device, as well as operation method and device of vehicle driving training device
CN109285238A (en) Information processing method, device, terminal and computer readable storage medium
CN108052920A (en) For the method and apparatus of output information
CN112885112B (en) Vehicle driving detection method, vehicle driving early warning method and device
CN109215160A (en) Data transmission method for uplink and device for black box
CN106686341A (en) Interactive vehicle-mounted intelligent terminal system and monitoring method
CN109902624A (en) The method and apparatus of information for rendering
CN107563455A (en) For obtaining the method and device of information
CN205232353U (en) Interactive intelligent terminal system
CN110248335A (en) The method and computer readable storage medium of ITS Information are collected using Internet of Things
CN106781510A (en) A kind of method of the event that collided in use mobile phone automatic detection driving procedure
CN107295062A (en) Article uses control method, equipment, system and article
CN116343493A (en) Method and device for identifying violation of non-motor vehicle, electronic equipment and storage medium
JP7092958B1 (en) Information processing methods, information processing devices, and programs
CN108960432A (en) Decision rule method, apparatus and computer readable storage medium
CN113538895B (en) Vehicle regulation and control method, device and system and computer storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant