KR20220048515A - Method And Apparatus for Vehicle State Transition Learning Based on Vehicle State Based Model - Google Patents

Abstract

The present disclosure provides a method and apparatus for learning a state transition of a vehicle based on a vehicle state-based model. According to one aspect of the present disclosure, provided are the method and device for learning the state transition in the vehicle state-based model so that a vehicle operates adaptively to an event occurring in relation to the vehicle. The method comprises: a process of recognizing the current event; a process of determining whether or not the state transition is possible; a process of determining the transition state of the vehicle; and a process of learning the vehicle state-based model.

Description

Method And Apparatus for Vehicle State Transition Learning Based on Vehicle State Based Model

The present invention relates to a method and apparatus for learning a state transition of a vehicle based on a vehicle state-based model.

The content described in this section merely provides background information for the present embodiment and does not constitute the prior art.

Autonomous driving is generally performed based on a vehicle state based model generated by software installed in a vehicle. This vehicle state-based model is modeling the state and state transformation of the vehicle, and by defining the operation or event to be performed in order to transform the vehicle state, the state transition of the vehicle is abstracted. to provide. At this time, the vehicle state is defined as a value of a system variable at a specific time point, and a state transition is defined as the occurrence of an event and the operation of the vehicle according to the occurrence of the event.

1 is an exemplary diagram of a vehicle state-based model.

1 shows an example of a state transition for driving speed control. Referring to FIG. 1 , when the accelerator pedal (acc in FIG. 1) is pressed while the engine is running (Engine_On in FIG. 1), the throttle valve (tv in FIG. 1) of the engine is opened, and the vehicle is in a moving state (FIG. 1) is transferred. In FIG. 1 , possible operations and corresponding states are defined in a state in which the vehicle is moving. The first is a state in which the vehicle is accelerated by pressing the accelerator pedal (peed_Up in FIG. 1 ), and the second is the brake (in FIG. 1 ). brk) is pressed and the vehicle is decelerated (Speed_Down in FIG. 1).

As described above, the general on-vehicle software controls the driving operation by changing the state of the vehicle according to an event occurring in the vehicle, based on a predefined state-based model. However, when an event not defined in such a state-based model occurs, there is a problem in that general vehicle-mounted software cannot properly control the driving operation of the vehicle.

Therefore, there is a problem that the driving safety of an autonomous driving vehicle cannot be guaranteed with only the conventional in-vehicle software, and even when a dangerous situation occurs that the in-vehicle software does not consider, the autonomous driving vehicle can be quickly and fluidly Therefore, it is necessary to devise a plan to respond to dangerous situations. Since there is no limit to such dangerous situations, learning-based responses must be made in order for the autonomous vehicle to efficiently respond to dangerous situations that the software does not consider.

Laid-open Patent Publication No. 10-2019-0045511 (2019.05.03.)

According to one aspect of the present disclosure, a main object is to provide a method and apparatus for learning a state transition in a vehicle state-based model so that the vehicle operates adaptively to an event occurring in relation to the vehicle.

According to one aspect of the present disclosure, in a method for learning a vehicle state based model to cause a vehicle state to undergo a state transition, a process of recognizing a current event that is an event; determining whether a state transition is possible from a current state of the vehicle based on the vehicle state-based model; determining a transition state of the vehicle based on the monitoring state when a state transition is not possible in a monitoring state that is a result of monitoring the system of the vehicle; and learning the vehicle state-based model so that the vehicle state-based model reflects a new state transition based on the current event, the current state, and the transition state. to provide.

According to another aspect of the present disclosure, in an apparatus for learning a vehicle state based model to cause a vehicle state to undergo a state transition, generated inside or outside the vehicle an event recognition unit for recognizing a current event that is an event; a transition determination unit that determines whether a state transition is possible from a current state of the vehicle based on the vehicle state-based model; a transition state generator configured to determine a transition state of the vehicle based on the monitoring state when a state transition is not possible in a monitoring state that is a result of monitoring the system of the vehicle; and a learning unit for learning the vehicle state-based model so that the vehicle state-based model reflects a new state transition based on the current event, the current state, and the transition state. to provide.

According to another aspect of the present disclosure, in the above-described process of learning the vehicle state-based model, the current state transitioned to the transition state due to a vehicle action of the vehicle according to the current event. to provide a state transition learning method, characterized in that it is taught to model using a corresponding node and edge.

According to another aspect of the present disclosure, when the transition state cannot be determined, the transition state is a new state created to transfer the control right of the vehicle to the driver, and the vehicle action of the vehicle performed by the driver ) as a cause, it provides a vehicle state transition learning method, characterized in that it learns to model the transition from the current state to the new state to the corresponding node and edge.

According to another aspect of the present disclosure, the above-described state transition learning apparatus further includes an operation driving unit for requesting a vehicle action from the vehicle so that the state of the vehicle changes from the current state to the transition state. It provides an apparatus for learning state transition of a vehicle, characterized in that.

According to an aspect of the present disclosure, by providing a method and an apparatus for learning state transition in a vehicle state-based model, there is an effect of allowing a vehicle to adaptively operate in response to an event occurring in relation to the vehicle.

According to another aspect of the present disclosure, when the transition state cannot be determined, the control right of the superior vehicle is transferred to the driver and a new state transition is learned in the vehicle state-based model based on the driver's vehicle motion, thereby making the existing vehicle state-based model There is an effect that the vehicle can quickly and efficiently respond to this difficult situation.

Accordingly, the method and apparatus for learning state transition according to various aspects of the present disclosure support autonomous driving of a vehicle and have an effect of improving driving stability of an autonomous driving vehicle.

1 is an exemplary diagram of a vehicle state-based model.
2 is a conceptual diagram illustrating an apparatus for learning state transition of a vehicle according to an exemplary embodiment of the present disclosure.
3 is a list showing system state variables and their values according to an embodiment of the present disclosure.
4 is a flowchart illustrating a method for learning state transition of a vehicle according to an exemplary embodiment of the present disclosure.
5 is a diagram illustrating a state transition of a vehicle state-based model to which a vehicle state transition learning method according to an embodiment of the present disclosure is applied.

Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. It should be noted that, in adding the reference numerals to the components of each drawing, the same components are to have the same reference numerals as much as possible even if they are displayed on different drawings. In addition, in describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description thereof will be omitted.

In addition, in describing the components of the present disclosure, terms such as second and first may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms. Throughout the specification, when a part 'includes' or 'includes' a certain element, this means that other elements may be further included, rather than excluding other elements, unless otherwise stated. . In addition, the '... Terms such as 'unit' and 'module' mean a unit that processes at least one function or operation, which may be implemented as hardware or software or a combination of hardware and software.

DETAILED DESCRIPTION The detailed description set forth below in conjunction with the appended drawings is intended to describe exemplary embodiments of the present disclosure and is not intended to represent the only embodiments in which the present disclosure may be practiced.

2 is a conceptual diagram illustrating an apparatus for learning state transition of a vehicle according to an exemplary embodiment of the present disclosure.

The apparatus 200 for learning state transition of a vehicle according to an embodiment of the present disclosure includes an event recognizer 202, a transition determining unit 204, a state monitoring unit 210, A transition state generator 212 , an action operator 220 , and a learning unit 230 are included in whole or in part. The apparatus 200 for learning state transition of a vehicle shown in FIG. 2 is according to an embodiment of the present disclosure, and not all components shown in FIG. 2 are essential components, and some components are added, changed, or may be deleted.

2 shows the vehicle state transition learning apparatus 200 as a device for convenience of explanation, this is for convenience of description, and in another embodiment, the learning device performs the functions of each component 202-230. It may be implemented as a software module or a processor.

The event recognition unit 202 recognizes an event occurring outside or inside the vehicle. These events may include, for example, the accelerator pedal press, brake press, obstacle detection, distance between vehicles, changes in traffic signals, weather changes, etc. An event that is triggered by a situation.

The transition determination unit 204 determines whether the state of the vehicle can be transitioned based on the occurrence event. Here, the state of the vehicle is defined as values or categories of system state variables at a specific point in time, and a state transition is a change from a specific state of the vehicle to another state by the occurrence of an event and a corresponding vehicle operation. For example, when the vehicle is in a stationary state (when the vehicle speed, which is a system state variable, is 0 km) and an event in which the accelerator pedal is pressed occurs, the vehicle may be transitioned to a driving state (vehicle speed exceeds 0 km). These system state variables will be described later in detail with reference to FIG. 3 .

The determination of the transition determination unit 204 is determined based on a vehicle state based model. Hereinafter, the vehicle state-based model will be described on the premise of a model in which the state of the vehicle is a node and the possibility of state transition between vehicle states is expressed as an edge. Any of these may be a vehicle state model of the present disclosure. These vehicle state-based models further define the events and/or vehicle behaviors required for the vehicle state to change, and these events and/or vehicle behaviors are set as data of the edges in the vehicle state-based model consisting of node-edges. do.

The transition determining unit 204 is a node of the vehicle state-based model and, with respect to a current node, which is a node corresponding to the current state of the vehicle, the event recognition unit 202 among the outgoing edges of the current node. It is determined whether the state of the vehicle can be transitioned by determining whether a trunk line corresponding to the recognized event exists. This outgoing edge is an edge between nodes with a direction, which establishes a relationship between nodes indicating that it can move from one vehicle state to another, and is not limited to a specific expression method such as an arrow. The existence of the trunk line corresponding to the event may be determined based on whether the event set in the outgoing trunk coincides with the event recognized by the event recognition unit 202 or whether it can be determined to be the same within a preset threshold range. When it is determined that a matching event exists, the transition determining unit 204 is an operation driving unit such that a vehicle state of a node (hereinafter, an adjacent node) connected to a corresponding outgoing edge (hereinafter, referred to as a transition edge) is obtained. By the operation of 220, the state of the vehicle is changed. That is, the actual vehicle state is changed to the vehicle state defined in the adjacent node by requesting the operation driver 220 to operate the vehicle set in the transition trunk. The state transition of the vehicle state-based model according to the present disclosure will be described in detail later with reference to FIG. 5 .

When it is determined that there is no matching event, that is, when it is determined that transition is not possible, the transition determination unit 204 does not request the operation driver 220 to operate the vehicle, or the current vehicle state-based model indicates that transition is not possible. The signal is transmitted to the operation driving unit 220 . In this case, the operation driving unit 220 receives a request to operate the vehicle from the transition state generating unit 212 .

The state monitoring unit 210 monitors the state of the vehicle. Specifically, the state monitoring unit 210 monitors the vehicle system and stores and provides values of various system state variables as a monitoring state. The state monitoring unit 210 monitors various configurations of the vehicle system. The monitoring state at the time when the transition determining unit 204 determines that the state cannot be transferred may be provided to the transition state generating unit 212 .

When the transition determining unit 204 determines that the state cannot be transferred, the transition state generating unit 212 selects a state that matches the monitoring state or has the highest similarity as the state to be transitioned. If the state to transition to cannot be selected, a new state corresponding to the monitoring state is added to the vehicle state-based model. That is, the transition state generating unit 212 determines the node most similar to the monitoring state among the nodes of the vehicle state-based model as the adjacent node, or generates a new state node based on all or part of the system variable values of the monitoring state to adjacent provided as a node.

An example in which the transition state generating unit 212 determines the adjacent node is by calculating the difference between the monitoring state and the matching system variable between the nodes of the vehicle state basic model, and selecting the node with the smallest difference as adjacent. It may be decided by node. When all of the above-described differences exceed a preset threshold, the transition state generator 212 determines that a state to be transitioned cannot be selected and generates a new state node.

Another example in which the transition state generator 212 determines the adjacent node is to determine the node having the largest number of matching system variable values (or ranges) between the monitoring state and the node of the vehicle state-based model as the adjacent node. can When the number of matching system variable values (or ranges) is less than a preset number, the transition state generator 212 determines that a transition state cannot be selected and creates a new state node.

When the transition state cannot be selected, the transition state generator 212 may generate a new state for transferring the control right of the vehicle to the driver and provide it as an adjacent node.

The operation driving unit 220 executes a vehicle operation to be performed in order to transition the state of the vehicle, thereby transferring the actual state of the vehicle to the state of the adjacent node. Such a vehicle operation may be set as data on the outgoing trunk corresponding to the state transition. When the transition state generator 212 cannot select a state to be transitioned to and thus generates a new state in which the control authority of the vehicle is transferred to the driver, these vehicle operations may be operations controlled by the driver. In this case, it is preferable that the state transition learning apparatus 200 of the vehicle learns so that the new state of the vehicle changed as a result of the driver's control and the operations controlled by the driver are reflected in the vehicle state-based model.

The learning unit 230 generates a vehicle state-based model by reflecting the added state transition based on the state to be transitioned determined by the transition state generating unit 212 , that is, the state of the adjacent node and the vehicle operation of the operation driver 220 . As much as possible, machine learning for model generation is performed.

This learning is, for example, if the adjacent node is a node of the existing vehicle state-based model, an outgoing edge from the current node to the adjacent node is generated, and the event recognized by the event recognition unit 202 and/or the operation driving unit 220 is performed. It may be to learn to set the vehicle motion as data of the generated outgoing edge. If the adjacent node is a newly created node, a new node defined as the state of the newly created node in the vehicle state-based model is created, and an outgoing edge from the current node to the new node is generated, and the event recognition unit 202 recognizes The event and/or vehicle motion of the motion driver 220 is set as data of the generated outgoing trunk line. It also trains the new node to generate more outgoing edges from the new node to the current node. In this case, a cycle between the current node and the new node is formed.

3 is a list showing system state variables and their values according to an embodiment of the present disclosure.

내지 +90

), 차량 방향의 경우 차량이 주행하는 방향으로서, 예컨대 전방(forward)/후방(backward), 센서 상태의 경우 센서 활성화(active)/비활성화(inactive) 또는 센서의 감지 결과가 시스템 상태(도 3의 Sensor_status) 변수의 값 또는 범주가 될 수 있다.The system state variable set as the vehicle state may be a value or a range corresponding to the state of each component of the vehicle system. For example, the vehicle system may include various sensors such as an engine, a throttle valve, a steering wheel, a radar, and a lidar. System state variables include engine state (Engine_state in FIG. 3), throttle valve state (Throttle_valve_state in FIG. 3), steering angle (Steering_angle in FIG. 3), vehicle direction (Car_direction in FIG. 3), sensor state (Sensor_state in FIG. 3), etc. can be Referring to FIG. 3 , the values or categories of these system state variables are on/off in the case of the engine state, open/increase/decrease/close in the case of the throttle valve state, and the angle within the steerable range for the steering angle (- 90

to +90

), in the case of the vehicle direction, as the direction in which the vehicle travels, for example, forward/backward, in the case of a sensor state, sensor activation (active)/inactive (inactive) or the detection result of the sensor is the system state (FIG. 3). Sensor_status) variable value or category.

The values or categories of these system state variables are not limited to the values or categories shown in the list of FIG. 3 . For example, a value of a new state variable generated by combining the system state variables shown in FIG. 3 may be set as some of the variables representing the vehicle state.

4 is a flowchart illustrating a method for learning state transition of a vehicle according to an exemplary embodiment of the present disclosure.

When the vehicle state transition learning apparatus recognizes whether an event has occurred in the vehicle (S400), it is determined whether the vehicle state transition is possible due to the corresponding event (S402, S410). Such determination may be made on the basis of whether a trunk line corresponding to an event occurring in the vehicle exists in the vehicle state-based model among outgoing trunks of the current node corresponding to the current state of the vehicle.

When a state transition is possible, that is, when a transition edge that is an edge corresponding to an event occurring in the vehicle exists, the state transition learning device of the vehicle uses the ECU (Electronic Control Unit) to enable the vehicle to perform the vehicle operation set on the corresponding transition edge. Or ask the configuration of the vehicle in question. Accordingly, the actual state of the vehicle is changed according to the operation of the vehicle, and the current node of the vehicle in the vehicle state-based model becomes a node connected to the outgoing trunk (S412).

When the state transition is not possible, the device for learning the state transition of the vehicle secures the value or category of the system state variable of the vehicle ( S420 ). These system state variables are variables representing the state of each component (eg, engine, throttle valve, steering wheel, etc.) constituting the vehicle system. The state transition learning device of the vehicle monitors the system state variable. Can store and provide state variable values or categories.

Based on the value or category of the secured system state variable, the vehicle state transition learning apparatus determines whether there is a node with the highest state similarity to the vehicle state-based model, and determines whether a node with the highest state similarity and the secured system state exists. When the difference between the value or the category of the variable is less than or equal to a preset threshold, the state of the corresponding node is determined as the state to be transitioned (S _J ) ( S422 , S430 ). In this case, the vehicle state transition learning apparatus adds the transition edge T _J-1 between the current node and the determined node on the vehicle state-based model, and changes the actual state of the vehicle (S432, S434).

When the above-described difference exceeds a preset threshold or when it is not possible to determine the state to be transitioned to for other reasons, the state transition learning apparatus of the vehicle is based on the value or category of the system state variable secured in the vehicle state-based model. A new state (S _K ) is defined to create a node (S440). In this case, the vehicle state transition learning apparatus adds the transition edge T _K-1 between the current node and the generated node and the transition edge T _K+1 between the generated node and the current node, respectively (S442, S444). .

The state transition learning apparatus of the vehicle requests the vehicle to operate the vehicle so that the actual state of the vehicle is changed, and the current node of the vehicle in the vehicle state-based model becomes a node connected to the outgoing trunk, that is, the generated node (S412).

The vehicle state transition learning apparatus performs machine learning on model generation so that the vehicle state-based model is generated by reflecting the state transition added in steps S432, S434, S440, S442, and S444.

Although it is described that each process is sequentially executed in FIG. 4 , this is merely illustrative of the technical idea of an embodiment of the present disclosure. In other words, those of ordinary skill in the art to which an embodiment of the present disclosure pertain may change the order described in FIG. 4 within the scope not departing from the essential characteristics of an embodiment of the present disclosure, or perform one or more of each process Since it may be applied by various modifications and variations to be executed in parallel, it is not limited to the time series sequence of FIG. 4 .

5 is a diagram illustrating a state transition of a vehicle state-based model to which a vehicle state transition learning method according to an embodiment of the present disclosure is applied.

FIG. 5A illustrates a case in which an event defined in the existing vehicle state-based model occurs and the state of the vehicle is transferred.

When the vehicle state transition learning device recognizes the event _e3 occurring in the vehicle in S _H , which is a node representing the current state of the vehicle, since there is an edge corresponding to _e3 in the outgoing edge of S _H , the state of the vehicle is It can be transferred only by the model. Accordingly, the state transition learning apparatus of the vehicle requests the vehicle operation a ₃ to be performed, which is the vehicle operation a ₃ set on the trunk line corresponding to e ₃ , and the state of the vehicle is the state defined in the S _I node, which is an adjacent node. change

FIG. 5B illustrates a case in which an event not defined in the existing vehicle state-based model occurs and the state of the vehicle is transferred by determining a transitionable adjacent node.

When the state transition learning device of the vehicle recognizes the event e _{J -1} occurring in the vehicle in S _H , which is a node representing the current state of the vehicle, there is no edge corresponding to e _{J -1} in the out-going edge of S _H. The state cannot be transferred only by the current model. Accordingly, the state transition learning device of the vehicle defines the state with the highest similarity to the system variable value based on the system variable value of the vehicle, S _J. Determining a node as an adjacent node, S _H Node and S _J Set the T _{J -1} edge indicating that node-to-node transition is possible. Event e _{J - 1} is set on the T _{J -1} edge.

FIG. 5C illustrates a case in which a vehicle state is transferred by creating a new node when an event not defined in the existing vehicle state-based model occurs and a transitionable adjacent node cannot be determined.

When the state transition learning device of the vehicle recognizes the event e _{K -1} occurring in the vehicle in S _H , which is a node representing the current state of the vehicle, there is no edge corresponding to e _{K -1} in the out-going edge of S _H , so the The state cannot be transferred only by the current model. In addition, when the system variable value of the vehicle and the state with the highest similarity to the system variable value within the preset threshold cannot be determined, the vehicle state transition learning apparatus selects the _SK node that newly defines the state based on the vehicle system variable value. Created and decided as an adjacent node, S _H Node and S _K Set the T _K-1 edge indicating that node-to-node transition is possible. Event e _{K - 1} is set on the T _K-1 edge. In addition, a _K - ₁ , which is a vehicle operation to be performed in order for the vehicle to enter the node _SK state, is further set in the T _K-1 trunk line. These S _K nodes are S _H In order to indicate that it is possible to transition to a node, a T _K+1 edge is further set.

Various implementations of the devices, units, processes, steps, etc., described herein may include digital electronic circuits, integrated circuits, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), computer hardware, firmware, software, and/or a combination thereof. These various implementations may include being implemented in one or more computer programs executable on a programmable system. The programmable system includes at least one programmable processor (which may be a special purpose processor) coupled to receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device. or may be a general-purpose processor). Computer programs (also known as programs, software, software applications or code) contain instructions for a programmable processor and are stored on a "computer-readable recording medium".

The computer-readable recording medium includes all types of recording devices in which data readable by a computer system is stored. These computer-readable recording media are non-volatile or non-transitory, such as ROM, CD-ROM, magnetic tape, floppy disk, memory card, hard disk, magneto-optical disk, and storage device. It may further include a medium or a transitory medium such as a data transmission medium. In addition, the computer-readable recording medium may be distributed in network-connected computer systems, and computer-readable codes may be stored and executed in a distributed manner.

Various implementations of the systems and techniques described herein may be implemented by a programmable computer. Here, the computer includes a programmable processor, a data storage system (including volatile memory, non-volatile memory, or other types of storage systems or combinations thereof), and at least one communication interface. For example, the programmable computer may be one of a server, a network appliance, a set-top box, an embedded device, a computer expansion module, a personal computer, a laptop, a Personal Data Assistant (PDA), a cloud computing system, or a mobile device.

The above description is merely illustrative of the technical idea of this embodiment, and a person skilled in the art to which this embodiment belongs may make various modifications and variations without departing from the essential characteristics of the present embodiment. Accordingly, the present embodiments are intended to explain rather than limit the technical spirit of the present embodiment, and the scope of the technical spirit of the present embodiment is not limited by these embodiments. The protection scope of this embodiment should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be interpreted as being included in the scope of the present embodiment.

200: vehicle state transition learning device 202: event recognition unit
204: transition determination unit 210: state monitoring unit
212: transition state generating unit 220: operation driving unit
230: study unit

Claims (16)

In a method for learning a vehicle state based model in order to make a state transition of a vehicle state,
Recognizing a current event that is an event occurring inside or outside the vehicle;
determining whether a state transition is possible from a current state of the vehicle based on the vehicle state-based model;
determining a transition state of the vehicle based on the monitoring state when a state transition is not possible in a monitoring state that is a result of monitoring the system of the vehicle; and
Learning the vehicle state-based model so that the vehicle state-based model reflects a new state transition based on the current event, the current state, and the transition state
A vehicle state transition learning method comprising a. According to claim 1,
The process of learning the vehicle state-based model is,
Learning that the current state has transitioned to the transition state due to the vehicle action of the vehicle according to the current event is modeled using corresponding nodes and edges A vehicle state transition learning method, characterized in that it. According to claim 1,
The vehicle state is defined based on a value or category of a system state variable,
The system state variable includes all or part of an engine status, a throttle valve state, a steering angle, a car direction, and one or more sensor status A vehicle state transition learning method, characterized in that. 3. The method of claim 2,
The process of determining the transition state is
and determining based on a difference between the monitoring state and a vehicle state corresponding to a node of the vehicle state-based model. 5. The method of claim 4,
When the node of the vehicle state-based model in which the monitoring state and the vehicle state are equal to or less than a preset threshold does not exist, the transition state is a new state generated based on the monitoring state. . 3. The method of claim 2,
The process of determining the transition state is
When there is no node of the vehicle state-based model in which the monitoring state and the vehicle state match by more than a preset threshold, the transition state is a new state generated based on the monitoring state. How to learn. 7. The method according to claim 5 or 6,
The process of learning the state-based model is,
The vehicle state-based model generates a new node that is a node corresponding to the new state, and generates an outgoing edge from the node corresponding to the current state to the new node, and the current event is set as an event of the generated outgoing edge, and learning to generate an outgoing edge from the new node to a node corresponding to the current state is performed. 3. The method of claim 2,
The process of learning the state-based model is,
An outgoing edge from a node corresponding to the current state to a node corresponding to the transition state is generated, and the current event is taught to be set as an event corresponding to the outgoing edge. transfer learning method. 9. The method of claim 8,
The vehicle operation is further set in the outgoing trunk line,
The vehicle operation is a state transition learning method of a vehicle, characterized in that set based on the monitoring state. 10. The method of claim 9,
The vehicle operation is an operation method of the vehicle, which is an operation method required to change the current state to the monitoring state. According to claim 1,
If the transition state cannot be determined, the transition state is
State transition learning method of a vehicle, characterized in that the new state is created to transfer the control right of the vehicle to the driver. 12. The method of claim 11,
The process of learning the state-based model is,
a node corresponding to that the current state of the vehicle has transitioned to the new state due to the vehicle operation according to the current event based on the vehicle action of the vehicle performed by the driver; and A state transition learning method of a vehicle, characterized in that it is taught to model as an edge. A computer program stored in a computer-readable recording medium to execute each process included in the method for learning state transition of a vehicle according to any one of claims 1 to 12. In an apparatus for learning a vehicle state based model in order to make a state transition of a vehicle state,
an event recognition unit for recognizing a current event that is an event occurring inside or outside the vehicle;
a transition determination unit that determines whether a state transition is possible from a current state of the vehicle based on the vehicle state-based model;
a transition state generator configured to determine a transition state of the vehicle based on the monitoring state when a state transition is not possible in a monitoring state as a result of monitoring the system of the vehicle; and
A learning unit for learning the vehicle state-based model so that the vehicle state-based model reflects a new state transition based on the current event, the current state, and the transition state
State transition learning apparatus of a vehicle, characterized in that it comprises a. 15. The method of claim 14,
The learning unit,
Learning that the current state has transitioned to the transition state due to the vehicle action of the vehicle according to the current event is modeled using corresponding nodes and edges State transition learning device characterized by the. 15. The method of claim 14,
and an operation driver for requesting a vehicle action from the vehicle so that the state of the vehicle changes from the current state to the transition state.

Images