WO2018186625A1

WO2018186625A1 - Electronic device, warning message providing method therefor, and non-transitory computer-readable recording medium

Info

Publication number: WO2018186625A1
Application number: PCT/KR2018/003734
Authority: WO
Inventors: 박찬종; 김지만; 양도준; 이현우
Original assignee: 삼성전자주식회사
Priority date: 2017-04-06
Filing date: 2018-03-29
Publication date: 2018-10-11

Abstract

An electronic device, a warning message providing method therefor, and a non-transitory computer-readable recording medium are provided. Disclosed is an artificial intelligence (AI) system using a machine learning algorithm such as deep learning and an application thereof. Disclosed, according to one embodiment, is an electronic device which can comprise: a position determination unit for determining a current position of the electronic device; a communication unit for receiving accident data and a driving situation; an output unit for outputting a warning message; and a processor for learning the received accident data to establish a plurality of accident prediction models, selecting an accident prediction model to be applied from among the plurality of accident prediction models on the basis of the determined current position, determining possibility of accident occurrence by using the selected accident prediction model, and controlling the output unit such that the output unit provides a warning message when the determined possibility of accident occurrence is greater than or equal to a preset value.

Description

Electronic device, method of providing alert message thereof and non-transitory computer readable recording medium

The present disclosure relates to an electronic device, a method of providing a warning message thereof, and a non-transitory computer readable recording medium. More particularly, the electronic device, a method of providing a warning message thereof, and a non-transitory computer capable of preventing a similar accident by learning a traffic accident pattern A readable recording medium.

In addition, the present disclosure relates to an artificial intelligence (AI) system that simulates functions such as cognition and judgment of the human brain by using a machine learning algorithm such as deep learning, and its application.

Artificial Intelligence (AI) system is a computer system that implements human-level intelligence, and unlike conventional rule-based smart systems, the machine learns, judges, and becomes smart. The more artificial intelligence systems are used, the better the recognition rate and the more accurate the user's taste will be. Therefore, the existing rule-based smart system is gradually being replaced by artificial intelligence system based on deep learning.

Artificial intelligence technology consists of elementary technologies that utilize machine learning (ex. Deep learning) and machine learning. Machine learning is an algorithmic technique for classifying and learning features of input data. Element technology is a technology that utilizes machine learning algorithms such as deep learning, and may be composed of technical fields such as linguistic understanding, visual understanding, inference / prediction, knowledge expression, and motion control.

The various fields in which artificial intelligence technology is applied are as follows. Linguistic understanding is a technique for recognizing and applying / processing human language / characters, including natural language processing, machine translation, dialogue system, question and answer, speech recognition / synthesis, and the like. Visual understanding is a technology that recognizes and processes objects as human vision, and includes object recognition, object tracking, image retrieval, person recognition, scene understanding, spatial understanding, and image enhancement. Inference / prediction is a technique for determining, logically inferring, and predicting information. It includes knowledge / probability-based inference, optimization prediction, preference-based planning, and recommendation. Knowledge representation is a technology that automates human experience information into knowledge data, and includes knowledge construction (data generation / classification) and knowledge management (data utilization). Motion control is a technology for controlling autonomous driving of a vehicle, movement of a robot, and the like, and includes motion control (navigation, collision, driving), operation control (action control), and the like.

On the other hand, in the past, a person directly performed the task of analyzing and classifying the cause of an accident. In addition, the existing machine learning method was only capable of determining the cause according to the criteria classified by humans.

In addition, there are accident occurrence points of various causes according to driving habits or surrounding factors, but there is a problem in that it is difficult for a user to recognize an accurate accident risk factor because an accident occurrence point is displayed only by a sign.

The present disclosure is to solve the above-described problems, it is possible to prevent similar accidents by learning the accident pattern using information obtained from the vehicle in the event of an accident, and providing a warning message by comparing the learned accident pattern with the current driving situation. An electronic device, a method of providing a warning message thereof, and a non-transitory computer readable recording medium are provided.

An electronic device according to an embodiment of the present disclosure for achieving the above object is, a positioning unit for determining the current position of the electronic device, a communication unit for receiving accident data and driving conditions, an output unit for outputting a warning message and the Constructs a plurality of accident prediction models by learning the received accident data, selects an accident prediction model to be applied among the plurality of accident prediction models based on the current position determined by the position determining unit, and selects the selected accident prediction model. The processor may include a processor configured to determine a possibility of an accident and to provide a warning message when the determined probability of occurrence of the accident is equal to or greater than a predetermined value.

If it is determined that an accident has occurred more than a predetermined number of times in an area having a predetermined area, the processor sets the area as an accident occurrence area, and the plurality of accident prediction models are general accidents applicable to all areas. It may include a prediction model and at least one special accident prediction model applicable to the at least one accident area.

The processor may build the general accident prediction model based on the entire accident data, and build the special accident prediction model based only on the accident data generated in the accident occurrence area.

The processor may determine the frequency by dividing the accident data by accident type, and may learn the accident data by weighting the accident data.

The processor may select the special accident prediction model when the determined current location is the accident occurrence area, and select the general accident prediction model in other cases.

The processor may determine the possibility of an accident by controlling the communication unit to collect a current driving situation and calculating a similarity between the accident occurrence situation learned by the accident prediction model and the collected current driving condition.

In addition, the processor may control the output unit to provide another warning message in stages as the probability of an accident increases.

The processor may reinforce the general accident prediction model and the special accident prediction model with different weights when an accident occurs in the accident occurrence area.

The accident data may include at least one of vehicle information, vehicle operation information, location information, road shape information, weather information, distance between vehicles, image information, acceleration information, and steering information at the time of the accident.

In addition, the accident prediction model may be an artificial neural network model.

On the other hand, the method for providing a warning message of the electronic device according to an embodiment of the present disclosure for achieving the above object, the step of learning the accident data to build a plurality of accident prediction model, based on the current location of the electronic device Selecting an accident prediction model to be applied among a plurality of accident prediction models, determining an occurrence probability of the accident using the selected accident prediction model, and providing a warning message if the determined occurrence probability is equal to or greater than a preset value. It may include.

And if it is determined that an accident occurs more than a predetermined number of times in an area having a predetermined area, setting the area as an accident-prone area, wherein the plurality of accident prediction models are general applicable to all areas. It may include an accident prediction model and at least one special accident prediction model applicable to at least one accident area.

The building may include building the general accident prediction model based on the entire accident data, and building the special accident prediction model based only on the accident data generated in the accident occurrence area.

The building may include classifying the accident data by accident type to determine a frequency and learning the accident data by weighting the accident data.

The selecting may include determining a current location of the electronic device, selecting the special accident prediction model when the determined current location is the accident area, and selecting the general accident prediction model in other cases. It may include the step of selecting.

The determining of the likelihood of occurrence of an accident may include collecting a current driving situation and calculating similarity between the accident occurrence situation learned by the accident prediction model and the collected current driving situation to determine the likelihood of an accident occurrence. .

In addition, the providing of the warning message may provide another warning message in stages as the possibility of an accident increases.

When the accident occurs in the accident occurrence area, the method may further include strengthening the general accident prediction model and the special accident prediction model with different weights.

The accident prediction model may be an artificial neural network model. Meanwhile, the accident prediction model may include a program for executing a method of providing an alert message of an electronic device according to an embodiment of the present disclosure. The non-transitory computer readable recording medium may include: constructing a plurality of accident prediction models by learning accident data; selecting an accident prediction model to be applied among the plurality of accident prediction models based on a current position of the electronic device; The method may include determining a possibility of an accident by using the selected accident prediction model, and providing a warning message when the determined probability of occurrence of an accident is equal to or greater than a preset value.

According to various embodiments of the present disclosure as described above, a warning message may be provided when a situation similar to a situation in which an accident occurs in the past may occur to prevent a similar accident.

1 is a schematic block diagram illustrating a configuration of an electronic device according to an embodiment of the present disclosure;

2 is a block diagram illustrating a configuration of an electronic device according to an embodiment of the present disclosure;

3 is a block diagram of a processor in accordance with some embodiments of the present disclosure;

4A is a block diagram of a data learning unit according to some embodiments of the present disclosure;

4B is a block diagram of a data recognizer according to some embodiments of the present disclosure;

5 is a view for explaining a method for establishing an accident prediction model of an electronic device according to an embodiment of the present disclosure;

6 is a view for explaining a method for setting an accident-prone area according to an embodiment of the present disclosure;

FIG. 7 is a diagram for explaining changing an accident prediction model applied according to a current position of an electronic device; FIG.

8 is a view for explaining the difference between applying a general accident prediction model and a special accident prediction model, respectively;

9 is a diagram for describing a method of providing, by an electronic device, a different warning message for each type of accident, according to an embodiment of the present disclosure;

10 is a diagram for explaining reinforcement learning with different weights for a general accident prediction model and a special accident prediction model when an accident occurs;

11 is a flowchart illustrating a warning message providing method of an electronic device according to an embodiment of the present disclosure;

12 is a diagram for describing an electronic device interoperating with a server according to another embodiment of the present disclosure;

13 is a sequence diagram illustrating an operation of an electronic device and a server according to another exemplary embodiment of the present disclosure.

Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In describing the present disclosure, when it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present disclosure, the detailed description thereof will be omitted. The terms to be described below are terms defined in consideration of functions in the present disclosure, and may vary according to a user, an operator, or a custom. Therefore, the definition should be made based on the contents throughout the specification.

Terms including ordinal numbers such as first and second may be used to describe various components, but the components are not limited by the terms. The terms are only used to distinguish one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component. The term and / or includes any one of a plurality of related items or a combination of a plurality of related items.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting and / or limiting of the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, the terms including or having are intended to indicate that there is a feature, number, operation, component, part, or a combination thereof described in the specification, but one or more other features or numbers, operation, configuration It should be understood that it does not preclude the presence or possibility of addition of elements, parts or combinations thereof.

In an embodiment, the module or unit performs at least one function or operation, and may be implemented by hardware or software or a combination of hardware or software. In addition, a plurality of 'modules' or a plurality of 'units' may be integrated into at least one module except for 'modules' or 'units' that need to be implemented by specific hardware, and may be implemented as at least one processor.

Meanwhile, in the present disclosure, the term user may refer to a person who uses an electronic device or a device (eg, an artificial intelligence electronic device) that uses an electronic device.

Hereinafter, the present disclosure will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically illustrating a configuration of an electronic device 100 according to an embodiment of the present disclosure. The electronic device 100 may be implemented as a mobile device such as a smartphone, a tablet PC, or a laptop. In addition, the electronic device 100 may be implemented as an In-Vehicle Infotainment (IVI) mounted in a vehicle. IVI refers to a device installed in a vehicle and refers to a device that provides audio and visual entertainment. For example, the IVI may be implemented as a navigation, a black box, a car audio system, a head up display (HUD), or the like.

The electronic device 100 may independently build an accident prediction model and prevent similar accidents based on the constructed accident prediction model. In addition, the electronic device 100 may interwork with an external device such as the server 200. The accident prediction model may be built in the server 200, and the electronic device 100 may receive and use the accident prediction model from the server 200. First, in describing the embodiment of FIG. 1, it is assumed that the electronic device 100 operates alone. An embodiment of interworking with the server 200 will be described later.

Referring to FIG. 1, the electronic device 100 may include a location determiner 110, a communicator 120, an outputter 130, and a processor 140.

The location determiner 110 may determine a current location of the electronic device 100. The location determiner 110 may be implemented as a global navigation satellite system (GNSS). Examples of satellite navigation systems include GPS (Global Positioning System), Galileo positioning system (Galileo positioning system), and GLONASS (GLObal NAvigation Satellite System). The electronic device 100 may determine an accident prediction model to apply based on the current location determined by the location determiner 110.

The communication unit 120 may transmit and receive data and / or control signals with an external device. For example, the communicator 120 may receive accident data, a current driving situation, an accident prediction model, and the like from an external device. In addition, the communication unit 120 may transmit a control signal to output a warning message to the external device.

The output unit 130 may output a warning message. For example, the output unit 130 may output at least one of a visual and audio signal. As another example, the output unit 130 may provide a warning message to an external device without directly providing a message to a user.

The processor 140 may build an accident prediction model based on the accident data. The processor 140 may determine whether the current driving situation is similar to the situation at the time of the occurrence of the accident, based on the constructed accident prediction model. If the similarity is greater than or equal to a preset value, the processor 140 may provide a warning message to the user. Through this, the electronic device 100 may prevent an accident similar to the existing accident case.

And also by various embodiments according to the present disclosure. The electronic device 100 may obtain the accident prediction data by using the received accident data as input data of the recognition model. The recognition model learned in the present disclosure may be constructed in consideration of application fields of the recognition model or computer performance of the apparatus. The learned object recognition model may be, for example, a model based on a neural network. The object recognition model may be designed to simulate a human brain structure on a computer and may include a plurality of weighted network nodes that simulate neurons in a human neural network. The plurality of network nodes may form a connection relationship so that neurons simulate the synaptic activity of neurons that send and receive signals through synapses. In addition, the object recognition model may include, for example, a neural network model or a deep learning model developed from the neural network model. In the deep learning model, a plurality of network nodes may be located at different depths (or layers) and exchange data according to a convolutional connection relationship. Examples of the object recognition model may include, but are not limited to, a Deep Neural Network (DNN), a Recurrent Neural Network (RNN), and a Bidirectional Recurrent Deep Neural Network (BRDNN).

In addition, the electronic device 100 may use an artificial intelligence agent to obtain a warning message for the received accident data as described above. At this time, the artificial intelligence agent is a dedicated program for providing an AI (Artificial Intelligence) based service (for example, a voice recognition service, a secretary service, a translation service, a search service, etc.), and an existing general purpose processor (for example, CPU) or a separate AI dedicated processor (eg, GPU, etc.).

2 is a block diagram illustrating a configuration of an electronic device 100 according to an embodiment of the present disclosure in detail. Referring to FIG. 2, the electronic device 100 includes a positioning unit 110, a communication unit 120, an output unit 130, a processor 140, a camera 150, a sensor unit 160, and an operation control unit 170. ) And memory 180.

In addition to the components illustrated in the embodiment of FIG. 2, the electronic device 100 may include various components such as an image processor (not shown), an image analyzer (not shown), a power source (not shown), and the like. In addition, the electronic device 100 is not necessarily limited to being implemented by including all the configurations shown in FIG. 2. For example, the electronic device 100 implemented without the camera 150 may receive image data through the communication unit 120 from the external device.

The location determiner 110 may include a circuit, software, and the like for implementing the satellite navigation system. The type of satellite navigation system to be applied may vary according to the specifications of each of the implemented electronic devices 100. For example, the applied satellite navigation system may be one of a Global Positioning System (GPS), a Galileo positioning system, and a Global NAvigation Satellite System (GLONASS). In addition, the location determiner 110 may use a combination of Assisted GPS (A-GPS), Differential GPS (D-GPS), and the like. The location determiner 110 may determine the current location of the electronic device 100 by further using location information of an AP, a base station, etc. connected through the communication unit 120.

The communication unit 120 communicates with an external device. For example, the external device may be implemented as a server, cloud storage, a network, or the like. Receiving accident data from an external device, the electronic device 100 may directly build, learn, and update an accident prediction model. The electronic device 100 may also receive an accident prediction model constructed by an external device. For example, when the electronic device 100 enters a specific accident occurrence area, the communication unit 120 may request an external server to transmit an accident prediction model corresponding to the accident occurrence area.

To this end, the communication unit 120 may include various communication modules such as a short range wireless communication module (not shown), a wireless communication module (not shown), and the like. Here, the short range wireless communication module is a module for communicating with an external device located in a short range according to a short range wireless communication scheme such as Bluetooth, Zigbee, or the like. In addition, the wireless communication module is a module that is connected to an external network and performs communication according to a wireless communication protocol such as WiFi, WiFi direct, or IEEE. In addition, the wireless communication module performs communication by connecting to a mobile communication network according to various mobile communication standards such as 3G (3rd Generation), 3GPP (3rd Generation Partnership Project), Long Term Evoloution (LTE), LTE Advanced (LTE-A), etc. It may further include a mobile communication module.

The output unit 130 may output the warning message in various ways. For example, the output unit 130 may be implemented as the display 131 to display a visual warning message. As another example, the output unit 130 may be implemented as a speaker 133 to output an auditory warning message. As another example, the output unit 130 may be implemented as an LED (not shown) to output a flashing signal. Of course, the warning message may be output through the display 131 and the speaker 133 at the same time.

The display 131 may display a warning message. For example, the display 131 may be implemented in the form of a head up display (HUD) on the vehicle front glass. The display 131 may be implemented as a liquid crystal display (LCD), an organic light emitting diode (OLED), a plasma display panel (PDP), or the like, thereby forming the electronic device 100. Various screens that can be provided can be displayed.

The speaker 133 may output voice. For example, the speaker 133 may output a warning message in the form of a notification sound or a voice message as well as various audio data. The speaker 133 may be built in the electronic device 100 or may be implemented in the form of an output port such as a jack.

The camera 150 may capture a still image or a video. For example, the camera 150 may photograph the front area of the vehicle. Using the difference between the captured image frames, the processor 140 may obtain acceleration information, steering information, and the like of the vehicle.

The camera 150 may be implemented as an image sensor such as a charge coupled device (CCD) or a complementary metal oxide semiconductor (CMOS). A CCD is a device in which charge carriers are stored and transported in a capacitor while each metal-oxide-silicon (MOS) capacitor is in close proximity to each other. The CMOS image sensor adopts a switching method that uses a CMOS technology that uses a control circuit and a signal processing circuit as peripheral circuits to make MOS transistors by the number of pixels, and sequentially detects the output using the same. It is an element to make.

When the electronic device 100 is implemented without the camera 150, the processor 140 may control the communication unit 120 to receive image information from a black box device installed in the vehicle.

The sensor unit 160 may measure a distance from the surrounding environment of the vehicle. The sensor unit 160 may collect information about a distance from another vehicle and a distance from a traffic facility such as a central separator. When the electronic device 100 is not one component of the vehicle (for example, when the electronic device 100 is implemented as a mobile device), the processor 140 may be connected to the surrounding vehicle from the sensor unit 160 of the vehicle. The communicator 120 may be controlled to receive distance information.

The manipulation controller 170 may control manipulation of the vehicle. For example, the manipulation controller 170 may be implemented as an electronic control unit (ECU), which is a system that electronically manages all operations of the vehicle. The processor 140 may be provided with steering wheel manipulation information, accelerator / brake information, component state information such as an engine, etc. from the manipulation controller 170. When the electronic device 100 is not one component of the vehicle, the processor 140 may control the communication unit 120 to receive operation information from the operation control unit 170 of the vehicle.

The memory 180 may store various modules, software, and data for driving the electronic device 100. For example, the memory 180 may store accident data, parasitic warning messages, and collected driving information. As another example, an accident prediction model may be stored in the memory 180 that may be used to determine a possibility of an accident.

The memory 180 is a storage medium that stores various programs necessary for operating the electronic device 100. The memory 180 may be implemented in the form of a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or the like. For example, the memory 180 may include a ROM for storing a program for performing an operation of the electronic device 100 and a RAM for temporarily storing data for performing an operation of the electronic device 100.

The processor 140 may control the above-described components of the electronic device 100. For example, the processor 140 may control the output unit 130 to output a warning message.

The processor 140 may be implemented as a single CPU to construct and update an accident prediction model, determine an accident prediction model to be applied, collect and analyze driving information, generate a warning message, or perform a plurality of processors and IPs. It may be implemented as.

The processor 140 may build at least one accident prediction model by learning the accident data received through the communication unit 120. For example, the processor 140 may analyze the accident data in a deep learning manner to determine a criterion for classifying the accident type. In addition, the processor 140 may classify accident cases based on the divided criteria.

The processor 140 may separately build a general accident prediction model that can be used in all regions and a special accident prediction model that can be used only in a specific region. For example, the processor 140 may set a accident area in a specific area, and build a special accident prediction model for each set accident area.

When it is determined that an accident has occurred more than a predetermined number of times in an area having a predetermined area based on the accident data, the processor 140 may set the area as an accident occurrence area.

The processor 140 may build a general accident prediction model based on the entire accident data. In addition, the processor 140 may build a special accident prediction model based only on accident data generated in the set accident area.

The processor 140 may analyze the types of accidents that frequently occur in the accident occurrence area. According to the analyzed frequency of accidents by type, the processor 140 may learn the accident data by giving a weight different from that of the general accident prediction model.

The processor 140 may determine an accident prediction model to be applied among the plurality of accident prediction models based on the current location of the electronic device 100. For example, based on the current location of the electronic device 100 determined by the location determiner 110, the processor 140 may determine whether the accident has entered an accident-prone area. As another example, the electronic device 100 may further determine whether to enter the accident-prone region by further using map data.

If it is determined that the incident has entered the accident area, the processor 140 may determine the possibility of the accident by using a special accident prediction model corresponding to the incident accident area that has entered. If it is determined that the accident is out of the accident area, the processor 140 may determine the possibility of the accident using the general accident prediction model.

The processor 140 may collect a current driving situation. The processor 140 may calculate the similarity with the collected current driving situation to the accident occurrence situation learned by the accident prediction model. For example, the current driving situation collected may include vehicle information, steering wheel operation information, excel / brake operation information, gear information, location information, road shape information, weather information, distance information with surrounding vehicles, image information, acceleration information and steering. It may include at least one of the information.

If the similarity is greater than or equal to a predetermined value, the processor 140 may provide a warning message indicating that there is a possibility of an accident. For example, the processor 140 may set a plurality of preset reference values. In addition, the processor 140 may provide a different kind of warning message whenever the similarity is increased and the predetermined reference value is exceeded.

As another example, the processor 140 may change the content of the warning message based on the type of accident determined based on the similarity. If it is determined that a high speed accident is likely to occur, the processor 140 may provide a warning message for slowing down the speed. If the user enters an area in which load kill occurs frequently, the processor 140 may provide a warning message to warn the wild animal that suddenly jumps.

Meanwhile, as described above, the processor 140 may perform the above-described operations by an existing general purpose processor (for example, a CPU or an application processor), but dedicated hardware for artificial intelligence (AI) for specific operations. The chip can perform the operation.

More specific operation of the processor 140 will be described later with reference to the accompanying drawings.

3 is a block diagram of a processor 140 in accordance with some embodiments of the present disclosure. Referring to FIG. 3, the processor 140 according to some embodiments may include a data learner 141 and a data recognizer 142.

The data learning unit 141 may learn a criterion for classifying an accident type, analyzing an cause of an accident, and the like. According to the learned criteria, the processor 140 may calculate the likelihood of an accident from the accident data. In addition, the processor 140 may classify the accident data into each type according to the learned criteria.

The data learner 141 may determine what data to use to build an accident prediction model. The data learner 141 acquires data to be used for learning and applies the acquired data to a data recognition model to be described later to learn criteria for accident occurrence possibility, accident cause analysis, and accident type classification.

The data recognizer 142 may recognize a situation from predetermined data by using the learned data recognition model. The data recognizing unit 142 may acquire predetermined data according to a predetermined criterion by learning, and use the data recognition model by using the acquired data as an input value. For example, using the learned accident prediction model, the data recognizer 142 may calculate the similarity between the current driving situation and the situation in which the accident occurred. The data recognizing unit 142 may update the accident prediction model by using the data acquired in the current driving situation and the new accident occurrence situation as input values again. As such, the data recognizer 142 may collect the accident data from the big data and the self-collection data.

At least one of the data learner 141 and the data recognizer 142 may be manufactured in the form of one or a plurality of hardware chips and mounted on the electronic device 100. For example, at least one of the data learner 141 and the data recognizer 142 may be manufactured in the form of a dedicated hardware chip for artificial intelligence (AI), or may be a conventional general purpose processor (eg, It may be manufactured as a part of an IP for a CPU or an application processor) or a specific function and mounted on the aforementioned various electronic devices 100.

In the embodiment of FIG. 3, the data learner 141 and the data recognizer 142 are both mounted on the electronic device 100, but they may be mounted on separate devices. For example, one of the data learner 141 and the data recognizer 142 may be included in the electronic device 100, and the other may be included in the server 200. In addition, the data learner 141 and the data recognizer 142 may be connected to each other by wire or wirelessly, so that model information constructed by the data learner 141 may be provided to the data recognizer 142, and data recognition may be performed. The data input to the unit 142 may be provided to the data learning unit 141 as additional learning data.

Meanwhile, at least one of the data learner 141 and the data recognizer 142 may be implemented as a software module. When at least one of the data learner 141 and the data recognizer 142 is implemented as a software module (or a program module including instructions), the software module may be stored in a non-transitory computer readable recording medium. At least one software module may be provided by an operating system (OS) or by a predetermined application. Alternatively, some of the at least one software module may be provided by the OS, and some of the at least one software module may be provided by a predetermined application.

4A is a block diagram of a data learner 141, according to some embodiments of the present disclosure. Referring to FIG. 4A, the data learner 141 may include a data acquirer 141-1, a preprocessor 141-2, a training data selector 141-3, and a model learner 141. -4) and the model evaluator 141-5.

The data acquirer 141-1 may acquire data necessary for determining a situation. The data acquirer 141-1 may receive training data through a network. For example, the data acquisition unit 141-1 may receive traffic accident-related big data classified by accident type as learning data. In addition, the data acquisition unit 141-1 may collect information on the current driving situation and use the data for learning.

The preprocessor 141-2 may preprocess the acquired data so that the data acquired for learning for situation determination may be used. The preprocessor 141-2 may process the acquired data into a predetermined format so that the model learner 141-4, which will be described later, uses the acquired data for learning for situation determination.

The training data selector 141-3 may select data necessary for learning from the preprocessed data. The selected data may be provided to the model learner 141-4. The training data selector 141-3 may select data necessary for learning from preprocessed data according to a predetermined criterion for determining a situation. In addition, the training data selector 141-3 may select data according to a predetermined criterion by learning by the model learner 141-4, which will be described later.

For example, in the early stage of learning, the learning data selector 141-3 may configure the learning data set using only data having different accident types. That is, for the initial learning, the learning data selector 141-3 may select accident data included in a type having low similarity to learn a criterion that is easy to distinguish.

In addition, the learning data selector 141-3 may select accident data that satisfies one of the criteria set by learning in common. In this way, the model learner 141-4 may learn another criterion different from the previously learned criterion.

The model learner 141-4 may learn a criterion for identifying what type of accident is based on the training data. In addition, the model learner 141-4 may learn a criterion about what training data to use for classification of an accident type.

The model learner 141-4 may train the data recognition model used for the situation determination using the training data. In this case, the data recognition model may be a pre-built model. For example, the accident prediction model, which is a data recognition model that recognizes traffic accident data, may be a model built in advance by receiving basic training data (for example, traffic accident data in which a death accident occurs).

The data recognition model may be constructed in consideration of the application field of the recognition model, the purpose of learning, or the computer performance of the device. The data recognition model may be, for example, a model based on a neural network. For example, a model such as a deep neural network (DNN), a recurrent neural network (RNN), and a bidirectional recurrent deep neural network (BRDNN) may be used as the data recognition model, but is not limited thereto.

According to various embodiments of the present disclosure, when there are a plurality of pre-built data recognition models, the model learner 141-4 may be a data recognition model for learning a data recognition model having a large correlation between input training data and basic training data. You can decide. In this case, the basic training data may be enjoyed for each type of data, and the data recognition model may be built in advance for each type of data. For example, the basic training data may be mood based on various criteria such as the region where the training data is generated, the time at which the training data is generated, the size of the training data, the genre of the training data, the creator of the training data, the types of objects in the training data, and the like. It may be.

In addition, the model learner 141-4 may train the data recognition model using, for example, a learning algorithm including an error back-propagation method or a gradient descent method. .

For example, the model learner 141-4 may train the data recognition model through supervised learning using the training data as an input value. As another example, the model learning unit 141-4 learns a data recognition model through unsupervised learning that finds a criterion for situation determination by learning a kind of data necessary for situation determination without a separate guidance. I can learn. As another example, the model learner 141-4 may train the data recognition model through reinforcement learning using feedback on whether the result of the situation determination according to the learning is correct.

In addition, when the data recognition model is trained, the model learner 141-4 may store the trained data recognition model. In this case, the model learner 141-4 may store the learned data recognition model in the memory 180 of the electronic device 100. Alternatively, the model learner 141-4 may store the learned data recognition model in a memory of the server 200 connected to the electronic device 100 through a wired or wireless network.

In this case, the memory 180 in which the learned data recognition model is stored may also store commands or data related to at least one other element of the electronic device 100. In addition, the memory 180 may store software and / or a program. For example, the program may include a kernel, middleware, an application programming interface (API) and / or an application program (or “application”), and the like.

The model evaluator 141-5 may input the evaluation data into the data recognition model, and cause the model learner 141-4 to relearn when the recognition result output from the evaluation data does not satisfy a predetermined criterion. have. In this case, the evaluation data may be preset data for evaluating the data recognition model.

In the initial recognition model construction stage, the evaluation data may be accident data having different accident types and damage scales. The evaluation data can then be replaced by an accident data set with increasingly similar similarity of accident types. Through this, the model evaluator 141-5 may gradually verify the performance of the data recognition model (eg, an accident prediction model).

For example, the model evaluator 141-5 may determine a predetermined criterion when the number or ratio of the evaluation data that is not accurate among the recognition results of the learned data recognition model for the evaluation data exceeds a preset threshold. It can be evaluated as not satisfied. For example, when a predetermined criterion is defined at a ratio of 2%, the model evaluator 141-5 when the learned data recognition model outputs an incorrect recognition result for more than 20 evaluation data out of a total of 1000 evaluation data. Can be judged that the learned data recognition model is not suitable.

On the other hand, when there are a plurality of trained data recognition models, the model evaluator 141-5 evaluates whether each learned data recognition model satisfies a predetermined criterion, and recognizes the final data as a model that satisfies the predetermined criterion. Can be determined as a model. In this case, when there are a plurality of models satisfying a predetermined criterion, the model evaluator 141-5 may determine any one or a predetermined number of models which are preset in the order of the highest evaluation score as the final data recognition model.

Meanwhile, the data acquisition unit 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141 in the data learner 141. At least one of -5) may be manufactured in the form of at least one hardware chip and mounted on the electronic device. For example, at least one of the data acquirer 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141-5. One may be manufactured in the form of a dedicated hardware chip for artificial intelligence (AI), or may be manufactured as an existing general purpose processor (eg, a CPU or an application processor) or part of an IP for a specific function. It may be mounted on the electronic device 100.

In addition, the data acquisition unit 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141-5 are electronic components. It may be mounted on the device, or may be mounted on separate electronic devices, respectively. For example, some of the data acquirer 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141-5. May be included in the electronic device 100, and some of them may be included in the server 200.

Meanwhile, at least one of the data acquirer 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141-5 is provided. It may be implemented as a software module. At least one of the data acquirer 141-1, the preprocessor 141-2, the training data selector 141-3, the model learner 141-4, and the model evaluator 141-5 is a software module. (Or, a program module including instructions), the software module may be stored on a non-transitory computer readable recording medium. At least one software module may be provided by an operating system (OS) or by a predetermined application. Alternatively, some of the at least one software module may be provided by the OS, and some of the at least one software module may be provided by a predetermined application.

4B is a block diagram of the data recognizer 142 according to some embodiments of the present disclosure. Referring to FIG. 4B, the data recognizer 142 according to some embodiments may include a data acquirer 142-1, a preprocessor 142-2, a recognition data selector 142-3, and a recognition result provider ( 142-4) and a model updater 142-5.

The data acquirer 142-1 may acquire data necessary for situation determination, and the preprocessor 142-2 may preprocess the acquired data so that the acquired data may be used for situation determination. The preprocessing unit 142-2 may process the acquired data into a predetermined format so that the recognition result providing unit 142-4, which will be described later, may use the obtained data for determining a situation.

The recognition data selector 142-3 may select data required for situation determination from among the preprocessed data. The selected data may be provided to the recognition result providing unit 142-4. The recognition data selector 142-3 may select some or all of the preprocessed data according to a predetermined criterion for determining the situation. In addition, the recognition data selector 142-3 may select data according to a predetermined criterion by learning by the model learner 142-4 to be described later.

The recognition result providing unit 142-4 may determine the situation by applying the selected data to the data recognition model. The recognition result providing unit 142-4 may provide a recognition result according to the recognition purpose of the data. The recognition result provider 142-4 may apply the selected data to the data recognition model by using the data selected by the recognition data selector 142-3 as an input value. In addition, the recognition result may be determined by the data recognition model.

For example, the recognition result providing unit 142-4 may recognize the input current driving state data according to the accident type classification criteria determined in the accident prediction model (data recognition model). In addition, using the accident prediction model, the recognition result providing unit 142-4 may calculate the similarity between the current driving state and the driving state at the time of the occurrence of the accident. Based on the calculated similarity, the processor 140 may determine whether an accident occurs and determine whether to provide a warning message.

The model updater 142-5 may cause the data recognition model to be updated based on the evaluation of the recognition result provided by the recognition result provider 142-4. For example, the model updater 142-5 may provide the model learner 141-4 with the recognition result provided by the recognition result provider 142-4 so that the model learner 141-4 can receive the recognition result. The data recognition model can be updated.

Meanwhile, the data acquisition unit 142-1, the preprocessor 142-2, the recognition data selection unit 142-3, the recognition result providing unit 142-4, and the model updating unit in the data recognition unit 142 ( At least one of the 142-5 may be manufactured in the form of at least one hardware chip and mounted on the electronic device. For example, among the data acquiring unit 142-1, the preprocessor 142-2, the recognition data selecting unit 142-3, the recognition result providing unit 142-4, and the model updating unit 142-5. At least one may be fabricated in the form of a dedicated hardware chip for artificial intelligence (AI), or may be fabricated as part of an existing general purpose processor (e.g., a CPU or application processor) or an IP for a particular function as described above. It may be mounted on various electronic devices 100.

In addition, one data acquisition unit 142-1, the preprocessor 142-2, the recognition data selection unit 142-3, the recognition result providing unit 142-4, and the model updater 142-5 It may be mounted on the device, or may be mounted on separate electronic devices, respectively. For example, among the data acquiring unit 142-1, the preprocessor 142-2, the recognition data selecting unit 142-3, the recognition result providing unit 142-4, and the model updating unit 142-5. Some may be included in the electronic device 100 and others may be included in the server 200.

Meanwhile, at least one of the data acquirer 142-1, the preprocessor 142-2, the recognition data selector 142-3, the recognition result provider 142-4, and the model updater 142-5. May be implemented as a software module. At least one of the data obtaining unit 142-1, the preprocessor 142-2, the recognition data selecting unit 142-3, the recognition result providing unit 142-4, and the model updating unit 142-5 is software. If implemented as a module (or a program module containing instructions), the software module may be stored on a non-transitory computer readable recording medium. At least one software module may be provided by an operating system (OS) or by a predetermined application. Alternatively, some of the at least one software module may be provided by the OS, and some of the at least one software module may be provided by a predetermined application.

5 is a diagram for describing a method of constructing an accident prediction model of the electronic device 100 according to an exemplary embodiment. The processor 140 may learn an accident pattern using information that can be obtained from a vehicle when an accident occurs. The processor 140 may build an accident prediction model by using the deep learning method.

If it is determined that an accident has occurred, the electronic device 100 may collect data at the time of the accident. For example, the electronic device 100 may determine whether an accident occurs based on malfunction information of parts of a linked vehicle, whether an air bag is operated, collision information collected through the camera 150 or the sensor unit 160, and the like. have.

As another example, the electronic device 100 may constantly monitor driving conditions. When the monitoring is suddenly stopped, the electronic device 100 may regard the final monitored driving situation as data at the time of the accident.

For example, the collected accident data may include at least one of vehicle information at the time of the accident, operation information of the vehicle, location information, road shape information, weather information, distance between vehicles, image information, acceleration information, and steering information. .

By using a machine learning method such as deep learning, the processor 140 may build an accident prediction model based on the collected accident data. Also, according to the accident type classification criteria learned through deep learning, the processor 140 may determine which type of accident data is the accident data. In the example of FIG. 5, an accident prediction model classifying accident types into five types is illustrated. For example, the processor 140 may classify each accident data into types of speeding, overtaking, slipping, and the like.

The processor 140 may build a general accident prediction model available in any region and a special accident prediction model specific to a specific region. For example, the processor 140 may set a specific area where a lot of accidents occur as an accident occurrence area. In addition, the processor 140 may analyze an accident type for each set accident area, and build an accident prediction model specialized for a specific accident type.

6 is a view for explaining a method of setting an accident-prone area according to an embodiment of the present disclosure. If it is determined that an accident occurs more than a predetermined number of times in an area having a predetermined area, the processor 140 may set the area as an accident occurrence area. For example, FIG. 6 illustrates an embodiment in which three

places

610, 620, and 630 are set as an accident occurrence area.

The processor 140 may determine the frequency for each type of accident occurring in each accident area. In addition, the processor 140 may build a special accident prediction model that weights the type of accident according to the frequency.

For example, as a result of analyzing accident data (ex. Weather information, image information, etc.) collected in the first accident area 610, the processor 140 may determine that the fog is generally dense at the time of the accident. Can be. In addition, the processor 140 may collect information on the first accident area 610 through the Internet. Through this, the processor 140 may recognize that the first accident area 610 is an area in which the fog is severely momentarily.

As another example, as a result of analyzing accident data (eg, image information, acceleration information, accelerator / brake information, road shape information, etc.) collected in the second accident area 620, the processor 140 may analyze the second accident area. It can be recognized that 620 is an area where a lot of interruptions occur.

As another example, as a result of analyzing accident data (eg, image information, accelerator / brake information, steering information, etc.) collected in the third accident area 630, the processor 140 may determine the third accident area 630. It can be recognized that this road killing area is frequent. The processor 140 may verify the recognized result by collecting information related to the third accident area 630 through the Internet. For example, the processor 140 may retrieve news related to a load kill generated in the third accident area 630.

Based on the recognized results, the processor 140 may build a special accident prediction model suitable for each

accident area

610, 620, 630.

FIG. 7 is a diagram for describing changing an accident prediction model applied according to a current position of the electronic device 100. Based on the current location of the electronic device 100 determined by the location determiner 110, the processor 140 may select an accident prediction model to be applied from among a plurality of constructed accident prediction models.

As shown in FIG. 7, when it is determined that the electronic device 100 enters an accident-prone area from a general area, the processor 140 is a special accident prediction model corresponding to an accident-prone area that enters from the general accident prediction model. You can change the accident prediction model that applies. If the electronic device 100 operates by downloading the accident prediction model from the server 200, the processor 140 requests the server 200 to request a special accident prediction model when entering the accident-prone region. ) Can be controlled.

The processor 140 may estimate the estimated time to enter the accident-prone area by referring to the current location, the map information, and the driving speed information determined by the location determiner 110. The special accident prediction model may be loaded from the memory 180 (or downloaded from the server 200) before entering the accident occurrence area, so that the processor 140 may perform the accident prediction model conversion at the same time as the accident occurrence area conversion.

FIG. 8 is a diagram for explaining a difference when a general accident prediction model and a special accident prediction model are respectively applied. When driving conditions other than the location are the same, whether or not a warning message is provided by the electronic device 100 may vary depending on the accident prediction model applied. This is because there are different things to be careful about in an area to prevent accidents.

The processor 140 may collect a current driving situation. In the embodiment of FIG. 8, the processor 140 may collect external brightness, average hourly speed, road shape, weather information, and distance information with a preceding vehicle. Using the general accident prediction model, the processor 140 may determine that the probability of occurrence of an accident is lower than a preset value. Thus, processor 140 may determine not to provide a warning message.

As illustrated in FIG. 8, the processor 140 may build a special accident prediction model for an accident occurrence area in which a number of stepless accidents occur. Using the special accident prediction model, the processor 140 may determine that the probability of an accident is higher than a preset value for the same driving situation. The processor 140 may control the output unit 130 to provide a warning message to the user.

FIG. 9 is a diagram for describing a method of providing, by an electronic device 100, an alert message for each accident type, according to an exemplary embodiment. The processor 140 may continuously collect driving information while the vehicle is driving. For example, the driving information collected may include vehicle information, steering wheel operation information, excel / brake operation information, gear information, location information, road shape information, weather information, distance information with surrounding vehicles, image information, acceleration information, and steering information. It may be at least one of.

The processor 140 may analyze the collected current driving situation by using an accident prediction model. The accident prediction model to be applied may be determined according to the location of the current electronic device 100. The processor 140 may calculate a similarity degree between the accident occurrence situation learned by the accident prediction model and the collected current driving situation.

For example, the processor 140 may calculate the similarity between each of the various accident types classified in the accident prediction model and the current driving situation. In the embodiment of FIG. 8, the processor 140 may calculate a similarity degree between the accident occurrence situation of each of the accident types 1 to 5 and the collected current driving situation.

Assuming that the preset value is 0.7, the processor 140 may provide a warning message as the similarity between the current driving situation and the accident type 4 accident occurrence situation is 0.81. The processor 140 may provide a warning message corresponding to the accident type 4. For example, if accident type 4 is the interrupting accident type as the road becomes narrower, processor 140 may generate a warning message specific to the accident type, such as "Please note that interruption of vehicles in other lanes is expected." The output unit 130 may be controlled to output.

According to an embodiment of the present disclosure, the electronic device 100 may provide different warning messages according to the degree of accident possibility. The processor 140 may set a plurality of thresholds and compare the similarity values between the current driving situation and the accident occurrence situation with the respective threshold values.

For example, if the similarity value exceeds 0.5, the processor 140 may provide a warning message for an item similar to an accident occurrence situation. The processor 140 may provide a voice message, such as "An accident may occur when the driving speed exceeds 70 km / h."

If the similarity value exceeds 0.7, the processor 140 may provide a stronger warning message. For example, processor 140 may provide a commanded voice message such as "Please slow down." In addition, the processor 140 may output a visual warning message together with a voice message by using the display 133 or an LED (not shown).

As the likelihood of an accident increases as described above, the processor 140 may control the output unit 130 to provide a stronger warning message.

10 is a diagram for explaining reinforcement learning with different weights for a general accident prediction model and a special accident prediction model when an accident occurs. FIG. 10 illustrates a case where an accident occurs in an accident occurrence area in which a number of stepless accidents occur.

When an accident occurs, the processor 140 may reinforce the accident prediction model using data at the time of the accident. For example, the processor 140 may reinforce the accident prediction model using input values such as external brightness at the time of an accident, average speed, road shape, weather information, distance information with a preceding vehicle, and the like.

If an accident occurs in an accident-prone area, it is necessary to reinforce both the special accident prediction model that applies to the accident-prone area and the general accident prediction model that applies to the entire area. Therefore, the processor 140 may reinforce and learn both the general accident prediction model and the special accident prediction model. However, the processor 140 may reinforce each accident prediction model by assigning different weights to the accident data.

The driving situation at the time of the accident occurrence illustrated in FIG. 10 may be determined to have a low probability of occurrence of an accident when analyzed by a general accident prediction model. If the accident data of FIG. 10 is input to the general accident prediction model by giving the same weight as other accidents, a problem may occur in statistics due to accident data corresponding to a rare case. Accordingly, the processor 140 may reinforce and learn a general accident prediction model by giving a low weight to accident data generated in the accident occurrence area.

On the contrary, if the analysis is conducted by a special accident prediction model, it may be judged that the probability of occurrence of an accident is high. In particular, when an accident corresponding to an accident type having a high frequency occurs in the accident occurrence area, the processor 140 may reinforce and learn a special accident prediction model by giving a high weight to the accident data.

11 is a flowchart illustrating a warning message providing method of the electronic device 100 according to an embodiment of the present disclosure.

Referring to FIG. 11, the electronic device 100 may learn accident data to build a plurality of accident prediction models (S1110). The plurality of accident prediction models may include a general accident prediction model and at least one special accident prediction model applicable to the entire region. A special accident prediction model can be built, one for each incident area.

If it is determined that an accident occurs more than a predetermined number of times in an area having a preset area, the electronic device 100 may set the corresponding area as an accident-prone area. The electronic device 100 may build a special accident prediction model based only on data on accidents occurring in the set accident area. The electronic device 100 may determine which type of accident has a high frequency by analyzing accident data generated in the accident occurrence area. In addition, the electronic device 100 may establish a special accident prediction model specialized for preventing a type of accident determined to have a high frequency.

The electronic device 100 may select an accident prediction model to be applied among the plurality of accident prediction models based on the current location in operation S1120. For example, the electronic device 100 may determine the current location by using a satellite navigation device such as GPS. As another example, the electronic device 100 may receive current location information by communicating with an external navigation device.

When the determined current location enters the accident-prone area, the electronic device 100 may apply a special accident prediction model corresponding to the entered accident-prone area. In contrast, when the determined current location leaves the accident occurrence area, the electronic device 100 may apply a general accident prediction model.

Subsequently, the electronic device 100 may determine a possibility of an accident by using the selected accident prediction model (S1130). The electronic device 100 may collect a current driving situation. The electronic device 100 may measure the similarity by comparing the accident occurrence situation learned by the accident prediction model with the collected current driving situation. The electronic device 100 may determine the possibility of an accident based on the measured similarity. For example, as the current driving speed, the weather, the road shape, and the like are similar to the case where a large number of accidents occur, the electronic device 100 may determine a higher probability of the accident.

The electronic device 100 may set a threshold value in advance and compare it with the likelihood of an accident (S1140). If the likelihood of an accident is greater than or equal to a preset value (S1140-Y), the electronic device 100 may provide a warning message to the user (S1150). The electronic device 100 may provide different warning messages according to the type of accident and the possibility of the occurrence of the accident. For example, the electronic device 100 may provide a warning message about a method (eg, deceleration, watching a side lane, etc.) to be dealt with according to an accident type. As another example, the electronic device 100 may provide a stronger warning message as the possibility of an accident increases.

12 is a diagram for describing an electronic device 100 interworking with the server 200 according to another exemplary embodiment. Referring to FIG. 12, the electronic device 100 and the server 200 may interoperate to learn and recognize data.

The server 200 may learn a criterion for recognizing an accident situation, and the electronic device 100 may determine a possibility of an accident and an accident type based on the learning result by the server 200.

In this case, the data learner 240 of the server 200 may perform the function of the data learner 141 illustrated in FIG. 4A. For example, the data learner 240 of the server 200 may learn a criterion for analyzing an accident type. The server 200 may build an accident prediction model by analyzing accident data according to the learned criteria. The data learner 240 may determine what data to use to learn / reinforce the accident prediction model. In addition, the data learner 240 may learn a criterion for determining a possibility of an accident, an accident type, and the like using the determined data. The data learner 240 acquires data to be used for learning and learns criteria for feature analysis by applying the obtained data to a data recognition model to be described later.

In addition, the recognition result providing unit 142-4 of the electronic device 100 may determine the situation by applying the data selected by the recognition data selecting unit 142-3 to the accident prediction model generated by the server 200. Can be. In addition, the recognition result providing unit 142-4 receives the accident prediction model generated by the server 200 from the server 200, and performs image analysis, content type determination, etc. using the received accident prediction model. Can be. The electronic device 100 may receive a special accident prediction model corresponding to the case where the electronic device 100 is adjacent to the accident occurrence area. The model updater 142-5 of the electronic device 100 may update the accident prediction model by providing the model learner 240-4 of the server 200.

That is, the electronic device 100 may use an accident prediction model generated by using the computing power of the server 200. In addition, the accident data learned or recognized by the plurality of electronic devices 100 is transmitted to the server 200, so that the server 200 may update the accident prediction model. In addition, the server 200 transmits the accident data and the driving habit data learned or recognized in each of the plurality of electronic devices 100 to the server 200 so as to generate an accident prediction model personalized for each electronic device 100. You can also create a.

FIG. 13 is a sequence diagram illustrating an operation of the electronic device 100 and the server 200 according to another exemplary embodiment.

In this case, the electronic device 100 may include a general purpose processor, and the server 200 may include an artificial intelligence processor. Alternatively, the electronic device 100 may include at least one application, and the server 200 may include an operating system. The server 200 is a component that is more integrated, dedicated, has a smaller delay, has better performance, or has more resources than the electronic device 100, and is required to generate, update, or apply a recognition model. Many operations that can be faster and more efficient than the electronic device 100 can be a component.

In this case, an interface for transmitting / receiving data between the electronic device 100 and the server 200 may be defined.

For example, an application program interface (API) having training data to be applied to the recognition model as an argument value (or, a parameter value or a transfer value) may be defined. An API is a set of subroutines or functions that can be called for processing of one protocol (eg, a protocol defined in the electronic device 100) to another protocol (eg, a protocol defined in the server 200). Can be defined. That is, an API may provide an environment in which an operation of another protocol may be performed in one protocol.

The server 200 may build an accident prediction model by collecting accident data through various paths (S1310). For example, the server 200 may build an accident prediction model using accident data obtained from a police agency, a navigation company, and the like and accident data transmitted from each electronic device 100.

The server 200 may build a general accident prediction model applicable to the entire region. In addition, the server 200 may set an accident occurrence area, and may build a special accident prediction model specialized for the set accident occurrence area. As the accident prediction model that can be applied according to the region is different as described above, the electronic device 100 must measure the current location (S1320).

The electronic device 100 may transmit the measured current location data to the server 200 (S1330). In operation S1340, the server 200 may transmit an accident prediction model corresponding to the location of the electronic device 100.

The electronic device 100 may determine a possibility of an accident by using the received accident prediction model in operation S1350. If the possibility of an accident is greater than or equal to a preset value, the electronic device 100 may output a warning message to the user (S1360).

According to various embodiments of the present disclosure as described above, the electronic device 100 analyzes a traffic accident situation that has occurred previously by using artificial intelligence, and provides a warning message to a user when the current driving situation is similar to the accident situation. Similar accidents can be prevented from occurring.

As used herein, the term "unit" includes a unit composed of hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. The module may be an integrally formed part or a minimum unit or part of performing one or more functions. For example, the module may be configured as an application-specific integrated circuit (ASIC).

Various embodiments of the present disclosure may be implemented in software that includes instructions stored in a machine-readable storage media. A device capable of calling and operating in accordance with the called command may include an electronic device according to the disclosed embodiments (for example, the electronic device A.) When the command is executed by the processor, the processor directly, Alternatively, other components may be used to perform functions corresponding to the instructions under the control of the processor, and the instructions may include code generated or executed by a compiler or an interpreter. It may be provided in the form of a non-transitory storage medium, where "non-transitory" means that the storage medium does not contain a signal and does not actually contain a signal. It is meant to be tangible but does not distinguish that data is stored semi-permanently or temporarily on a storage medium.

According to one embodiment, a method according to various embodiments disclosed herein may be provided included in a computer program product. The computer program product may be traded between the seller and the buyer as a product. The computer program product may be distributed online in the form of a device-readable storage medium (eg compact disc read only memory (CD-ROM)) or through an application store (eg Play StoreTM). In the case of an online distribution, at least a portion of the computer program product may be stored at least temporarily on a storage medium such as a server of a manufacturer, a server of an application store, or a relay server, or may be temporarily created.

Each component (eg, a module or a program) according to various embodiments may be composed of a singular or plural number of objects, and some of the above-described subcomponents may be omitted, or other subcomponents may be omitted. It may be further included in various embodiments. Alternatively or additionally, some components (eg, modules or programs) may be integrated into one entity to perform the same or similar functions performed by each corresponding component prior to integration. According to various embodiments, operations performed by a module, program, or other component may be executed sequentially, in parallel, repeatedly, or heuristically, or at least some of the operations may be executed in a different order, omitted, or another operation may be added. Can be.

Claims

In the method for providing a warning message of the electronic device,

Learning accident data to construct a plurality of accident prediction models;

Selecting an accident prediction model to be applied from the plurality of accident prediction models based on a current position of the electronic device;

Determining a possibility of an accident using the selected accident prediction model; And

And providing a warning message when the determined likelihood of occurrence of an accident is equal to or greater than a predetermined value.
The method of claim 1,

If it is determined that an accident occurs more than a predetermined number of times in a region having a predetermined area, setting the region as an accident-prone region; further comprising:

The plurality of accident prediction models,

A warning message providing method comprising a general accident prediction model applicable to an entire region and at least one special accident prediction model applicable to at least one accident-prone region.
The method of claim 2,

The building step,

A warning message providing method for building the general accident prediction model based on the entire accident data, and building the special accident prediction model based only on the accident data generated in the accident occurrence area.
The method of claim 1,

The building step,

A method of providing a warning message for classifying the accident data by accident type to determine a frequency and learning the accident data by weighting the accident data.
The method of claim 2,

The selecting step,

Determining a current location of the electronic device; And

Selecting the special accident prediction model when the determined current location is the accident-prone region, and selecting the general accident prediction model in the remaining cases.
The method of claim 1,

Determining the possibility of the accident,

Collecting a current driving situation; And

The warning message providing method for determining the likelihood of accident by calculating the similarity between the accident occurrence situation learned by the accident prediction model and the collected current driving situation.
The method of claim 6,

Providing the warning message,

How to provide a warning message that provides different warning messages in stages as the probability of an accident increases.
The method of claim 2,

And strengthening the general accident prediction model and the special accident prediction model with different weights when an accident occurs in the accident occurrence area.
The method of claim 1,

The accident data,

And at least one of vehicle information, vehicle operation information, location information, road shape information, weather information, inter-vehicle distance information, image information, acceleration information, and steering information at the time of an accident.
The method of claim 1,

The accident prediction model is an artificial neural network (Neural Network) model, a warning message providing method.
In an electronic device,

A location determiner that determines a current location of the electronic device;

Communication unit for receiving the accident data and driving situation;

An output unit for outputting a warning message; And

Constructs a plurality of accident prediction models by learning the received accident data, selects an accident prediction model to be applied among the plurality of accident prediction models based on the current position determined by the position determining unit, and selects the selected accident prediction model And a processor configured to determine a possibility of an accident and to provide a warning message when the determined probability of occurrence of the accident is greater than or equal to a preset value.
The method of claim 11,

The processor,

If it is determined that an accident has occurred more than a predetermined number of times in an area having a predetermined area, the area is set as an accident-prone area,

The plurality of accident prediction models,

An electronic device comprising a general accident prediction model applicable to an entire region and at least one special accident prediction model applicable to at least one accident-prone region.
The method of claim 12,

The processor,

And building the general accident prediction model based on total accident data and building the special accident prediction model based only on accident data generated in the accident occurrence area.
The method of claim 12,

The processor,

And classifying the accident data by accident type to determine a frequency and learning the accident data by weighting the accident data.
The method of claim 12,

The processor,

And selecting the special accident prediction model when the determined current location is the accident-prone region, and selecting the general accident prediction model in the remaining cases.