KR102347461B1 - Anti-hacking apparatus and method for remote-controlled or autonomously movable vehicle - Google Patents

Abstract

The present invention relates to an apparatus and method for preventing hacking on a remote-controllable or autonomous-drivable vehicle. More specifically, the apparatus is formed in a remote-controllable or autonomous-drivable vehicle to receive only a radio signal from the outside of the vehicle, and to sense and block a radio signal having an abnormal pattern as a signal for a hacking attempt through an interconnection with a vehicle control part controlling the vehicle, thereby preventing hacking on the vehicle. In accordance with the present invention, the apparatus can easily detect an external radio signal, which has abnormal signal intensity or indicates abnormally repetitive patterns, as a signal for a hacking attempt to notify a vehicle control part of the risk of hacking, thereby supporting the vehicle control part so as to block the hacking attempt in advance before the vehicle control part loses its control right due to the hacking, which can lead to an effect of safely protecting the vehicle from the hacking.

Description

Anti-hacking apparatus and method for remote-controlled or autonomously movable vehicle

The present invention relates to an apparatus and method for preventing hacking for a remote-controlled or autonomously movable moving object, and more particularly, to a remote-controlled or autonomously capable moving object, configured to receive only radio signals external to the moving object, and to control the moving object. It relates to a hacking prevention apparatus and method for a remote control or autonomously movable moving object that prevents hacking of a moving object by detecting and blocking a wireless signal having an abnormal pattern as a signal for a hacking attempt through interworking with a vehicle control unit.

An autonomous driving vehicle that senses the state of the vehicle and surrounding objects according to the development of current communication technology and various devices, and collects information about the surrounding traffic situation through communication with various external servers to automatically drive the vehicle without driver intervention is emerging, and such autonomous vehicles are continuously developing according to the high performance of communication technology and various devices.

Since these autonomous vehicles drive the vehicle on their own without driver intervention, it is possible to increase the driver's convenience, but since the driver is not involved in driving, it must make an accurate judgment on its own for all situations that threaten the vehicle and control the behavior of the autonomous vehicle. Thorough protection of the vehicle control unit to be controlled must be made.

However, due to the nature of the autonomous vehicle, which contains a large proportion of electronic components, there are concerns about hacking attempts on the autonomous vehicle, which threatens the safe driving of the autonomous vehicle, thereby reducing the reliability of the autonomous vehicle.

Accordingly, there is a need to develop alternatives for protecting autonomous vehicles from hacking threats.

Korean Patent Publication No. 10-2019-0116192

The present invention collects all wireless signals received from the outside of a moving object capable of remote control or autonomous driving, accurately identifies a radio signal suspected of being hacked among these radio signals, and warns the vehicle control unit that controls the moving object when hacking is suspected. The purpose is to protect the mobile body from hacking by supporting follow-up actions.

An anti-hacking device according to an embodiment of the present invention includes an antenna unit installed outside a moving object capable of remote control or autonomous movement to receive a radio signal, a communication unit communicating with a vehicle control unit controlling the moving object, and the antenna; A signal collection unit that converts to a frequency domain when receiving the wireless signal through the unit to generate signal collection information and a plurality of signal types corresponding to the signal collection information by applying the signal collection information to the learned first learning model A specific signal type is identified, and a specific operation mode currently set in the vehicle control unit among a plurality of operation modes preset in the vehicle control unit respectively corresponding to a plurality of different operating states of the moving object is confirmed through the communication unit, and the signal Comparing the collection information with a preset reference frequency pattern corresponding to the identified specific signal type for each of a plurality of comparison properties including signal strength and overlapping frequency, a difference value for each of the plurality of comparison properties is calculated, and the plurality of comparison properties of each of the at least one comparison property corresponding to the specific operation mode is applied to the difference value to calculate a final value for each of the plurality of comparison properties, and then the final value for each of the plurality of comparison properties is applied to a preset abnormality determination criterion It may include a control unit that analyzes the signal collection information to determine whether the signal collection information is normal, generates warning information to warn of a hacking risk when abnormal, and transmits the generated warning information to the vehicle control unit through the communication unit.

As an example related to the present invention, when the learning of the first learning model is completed, the controller applies each of a plurality of signal types preset in the first learning model to the first learning model to correspond to the plurality of signal types, respectively It may be characterized in that a reference frequency pattern is calculated, and setting information for a plurality of reference frequency patterns each matching the plurality of signal types is generated and stored.

As an example related to the present invention, when the signal type corresponding to the signal collection information is not identified, the control unit generates query information for inquiring whether the signal collection information is normal and then transmits it to the vehicle control unit, When the signal collection information is a normal signal based on response information received in response to the inquiry information from the vehicle control unit, a new signal type is generated and then matched with the signal collection information to learn the first learning model can be done with

As an example related to the present invention, when the signal collection information is abnormal based on the response information, the control unit may transmit warning information about the risk of hacking to the vehicle control unit.

As an example related to the present invention, each time the control unit receives the signal collection information, the specific signal type identified for the signal collection information and the signal collection information and the analysis result information on whether the signal collection information is normal are matched with each other After generating training data including one state, the second learning model is trained in advance, and the abnormal frequency pattern and the normal frequency pattern for each signal type are learned in the second learning model, and the second learning model is learned Upon completion, the signal collection information received from the signal collection unit and a specific signal type corresponding to the signal collection information are applied to the second learning model to generate analysis result information on whether the signal collection information is normal can do.

As an example related to the present invention, the control unit further includes a specific operation mode confirmed when the signal collection information is received in the learning data to teach the second learning model, and when learning of the second learning model is completed The signal collection information and a specific signal type and a specific operation mode corresponding to the signal collection information may be applied to the second learning model to generate analysis result information on whether the signal collection information is normal.

As an example related to the present invention, when the specific operation mode is a stop mode corresponding to a state in which the moving object is stopped, a comparison attribute different from a difference value of a comparison attribute corresponding to an overlapping frequency among the one or more comparison attributes A weight higher than the weight applied to the difference value of It may be characterized in that a weight higher than a weight applied to a difference value of another comparison attribute is applied.

As an example related to the present invention, the control unit may transmit the warning information through the communication unit to a user terminal corresponding to a preset user of the moving object.

According to an embodiment of the present invention, the hacking prevention method of a hacking prevention device that is installed outside of a moving object capable of remote control or autonomous movement, an antenna is installed to receive a radio signal, and communicates with a vehicle control unit that controls the moving object, the antenna generating signal collection information by converting it into a frequency domain when receiving a wireless signal through Identifying, confirming a specific operation mode currently set in the vehicle control unit among a plurality of operation modes preset in the vehicle control unit corresponding to a plurality of different operating states of the moving object, respectively, and sending the signal collection information Comparing with a preset reference frequency pattern corresponding to the identified specific signal type for each of the plurality of comparison properties including intensity and overlapping frequency, a difference value for each of the plurality of comparison properties is calculated, and one or more values corresponding to the specific operation mode are calculated. After applying a preset weight for each comparison attribute to the difference value for each comparison attribute, analyzing it according to a preset abnormality determination criterion to determine whether the signal collection information is normal and the vehicle control unit when abnormal through the analysis step It may include the step of transmitting warning information to warn of the risk of hacking.

As an example related to the present invention, the step of determining whether the normal operation is normal may include communicating with a hacking prevention device of another moving object and checking whether a difference occurs in at least some of the common reception information including GPS reception information. The method may further include, and the transmitting of the warning information may further include transmitting warning information to the vehicle control unit when a difference occurs in at least some of the common received information.

The present invention collects external radio signals that wirelessly access the vehicle control unit to detect hacking attempts against the vehicle control unit that controls a moving object capable of remote control or autonomous driving, and various types of normal signals received by the vehicle control unit are learned. Through the learning model of the state, it is determined whether there is a signal type corresponding to the external wireless signal to determine whether a hacking attempt has been made, and if there is a signal type corresponding to the external wireless signal, the corresponding signal type is determined according to the operation state of the moving object. By comparing the frequency pattern of the external radio signal with the normal frequency pattern of the external radio signal based on the frequency domain, an external radio signal having an abnormal signal strength or an abnormally repeating pattern is easily detected as a signal for a hacking attempt and sent to the vehicle control unit. By notifying the risk of hacking, there is an effect of supporting the vehicle control unit to block hacking attempts in advance before the control authority is seized by hacking, thereby safely protecting the moving object from hacking.

In addition, the present invention can support to easily distinguish between a normal signal and an abnormal signal for hacking based on learning the natural frequency pattern of a wireless signal having a different shape for each different moving object for each signal type, and the operation of the moving object It is possible to easily and accurately detect a wireless signal for a hacking attempt in consideration of the frequency pattern of a wireless signal that changes depending on the state, thereby having an effect of safely protecting a mobile object from hacking.

1 is a block diagram of a hacking prevention system for a remotely controlled or autonomously movable moving object according to an embodiment of the present invention.
2 is a block diagram of a hacking prevention device for a remotely controlled or autonomously movable object according to an embodiment of the present invention;
3 is a diagram illustrating an operation of an anti-hacking device according to an embodiment of the present invention.
4 is an operation flowchart of a hacking prevention method of a hacking prevention apparatus according to an embodiment of the present invention.
5 is an exemplary diagram of a signal type of a radio signal.
6 is an exemplary diagram of a hacking attempt signal.
7 is an exemplary operation diagram for learning-based hacking attempt detection of the hacking prevention device according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Prior to the description, the present invention provides an apparatus and method for preventing hacking by targeting a remote-controlled or autonomously movable moving object. It goes without saying that the hacking prevention device according to the embodiment of the present invention can be applied even to a moving object that can be remotely controlled, such as a drone.

1 is a block diagram of a hacking prevention system for an autonomous vehicle according to an embodiment of the present invention.

As shown, the hacking prevention system according to an embodiment of the present invention includes a vehicle control unit 10 that controls an autonomous driving vehicle (or vehicle) and a hacking prevention device 100 configured to communicate with the vehicle control unit 10 . may be included.

In this case, the vehicle control unit 10 and the hacking prevention device 100 may be configured in an autonomous driving vehicle, and the hacking prevention device 100 may be configured inside the vehicle control unit 10 .

Alternatively, the hacking prevention device 100 may be installed outside the autonomous vehicle.

Also, the vehicle control unit 10 may be configured as a moving body control unit configured in a moving body.

In this case, the vehicle control unit 10 includes various modules such as one or more sensor modules configured in the autonomous vehicle, a location measurement module for measuring the location of the autonomous vehicle, and a communication module for receiving information on external traffic conditions. It is possible to identify the current state of the autonomous vehicle based on the information received from , and control various driving devices for driving the autonomous vehicle to automatically control the operation of the autonomous vehicle without driver intervention.

For the safe operation of the autonomous vehicle operated by the vehicle control unit 10, the vehicle control unit 10 blocks the signal that the vehicle operation authority is stolen or the vehicle operation is interrupted by an outsider who does not have the vehicle operation authority. It is very important to prevent hacking of the self-driving vehicle, such as.

Therefore, the hacking prevention device 100 according to an embodiment of the present invention detects a hacking attempt by an outsider against the vehicle control unit 10 and notifies the vehicle control unit 10 of the hacking attempt when a hacking attempt occurs. By supporting this, it is possible to support the safe operation of the autonomous vehicle, which will be described in detail with reference to the drawings below based on the above-described configuration.

First, FIG. 2 is a configuration diagram of an apparatus 100 for preventing hacking for an autonomous vehicle according to an embodiment of the present invention.

As shown, the hacking prevention device 100 includes an antenna unit 110 for receiving a wireless signal received outside of the autonomous vehicle, and a communication unit for communication with the vehicle controller 10 for controlling the autonomous vehicle. (140), it may be configured to include a signal collection unit 120 that collects the radio signal and converts it into a frequency domain, and a control unit 130 that performs an overall control function of the hacking prevention apparatus 100.

In this case, the control unit 130 performs an overall control function of the hacking prevention device 100 , and the control unit 130 may include RAM, ROM, CPU, GPU, and a bus, and includes RAM, ROM, CPU, and GPU. The lights may be connected to each other via a bus.

Based on the above-described configuration, in the hacking prevention method of the hacking prevention apparatus 100 according to the embodiment of the present invention according to the exemplary operation diagram of the hacking prevention apparatus 100 according to the embodiment of the present invention according to FIG. 3 and FIG. A detailed operation configuration of the hacking prevention device 100 will be described with reference to an operation flowchart.

In this case, the antenna unit 110 is configured to include one or more antennas, and the antenna may be steered and installed to receive only radio signals received from the outside of the autonomous vehicle.

In addition, the signal collection unit 120 may convert the radio signal received through the antenna unit 110 into a frequency domain to generate signal collection information for the radio signal converted into the frequency domain (S1).

In addition, the control unit 130 may apply the signal collection information to a first learning model in which a plurality of frequency patterns for each signal type are learned when the signal collection information is received from the signal collection unit 120 (S2), A specific signal type corresponding to the signal collection information may be identified through the first learning model (S3).

As an example of the plurality of signal types, an external device that transmits information necessary for driving a vehicle to the vehicle control unit 10 includes a second signal corresponding to a signal transmitted to identify a vehicle, such as a unique ID set in the vehicle control unit 10 . It may include one signal type, a second signal type corresponding to a signal related to weather information, a third signal type corresponding to a signal related to multimedia content, and a fourth signal type corresponding to a signal for communication with a surrounding vehicle.

In addition, the plurality of signal types is a signal type corresponding to a radio signal that can be identified as a normal signal in the vehicle control unit 10 , and refers to a signal type for a normal signal that can be identified in the frequency domain by the vehicle control unit 10 . can do.

Accordingly, the control unit 130 applies the antenna unit to the first learning model in which a frequency domain-based frequency pattern is learned for each of a plurality of different signal types corresponding to signals other than signals for hacking in the vehicle control unit 10 . A specific signal type corresponding to the signal collection information may be identified by applying the signal collection information generated by converting the wireless signal received in 110 into the frequency domain through the signal collection unit 120 .

At this time, when the specific signal type is not identified through the first learning model, the control unit 130 queries the vehicle control unit 10 for whether the radio signal according to the signal collection information is a normal signal or not. Inquiry information may be generated and transmitted to the vehicle control unit 10 through the communication unit 140, and the inquiry information may include the signal collection information (S4).

Accordingly, the vehicle control unit 10 may determine whether the signal collection information is normal by analyzing the signal collection information included in the inquiry information when the inquiry information is received, Response information including the determination result may be generated and transmitted to the hacking prevention apparatus 100 .

In addition, when the wireless signal according to the signal collection information is a normal signal based on the determination result of the signal collection information included in the response information received from the vehicle control unit 10, the control unit 130 Case) (S5) After a new signal type is created, it can be matched with the signal collection information and trained in the first learning model (S6).

In this case, the new signal type may mean a new ID corresponding to the new signal type for differentiation from other signal types learned in the first learning model.

Accordingly, the controller 130 may update the first learning model by continuously learning a frequency pattern for a new signal type in the first learning model.

In addition, when the wireless signal according to the signal collection information is an abnormal signal (or abnormal) based on the determination result of the signal collection information included in the response information (S5), the control unit 130 is the vehicle control unit Warning information for notifying the hacking risk to (10) may be generated and transmitted to the vehicle control unit 10 through the communication unit 140 (S14).

On the other hand, when a specific signal type corresponding to the signal collection information is identified by applying the signal collection information to the first learning model, the control unit 130 transmits the signal collection information to the vehicle control unit 10 through the communication unit 140 in advance. A specific operation mode currently set in the vehicle control unit 10 among a plurality of set different operation modes may be checked (S7).

In this case, the plurality of different operation modes may each correspond to a plurality of different operation states of the autonomous vehicle, and the plurality of different operation states include a stop state and a driving state, and the plurality of mutually Other operation modes may include a stop mode corresponding to the stop state, a driving mode corresponding to the driving state, and the like.

For example, when a specific signal type corresponding to the signal collection information is identified through the first learning model, the control unit 130 informs the vehicle control unit 10 through the communication unit 140 to operate the autonomous vehicle. Transmits mode confirmation request information for confirming an operation mode corresponding to the state, and receives mode confirmation information transmitted in response to the mode confirmation request information from the vehicle control unit 10 through the communication unit 140 , A specific operation mode currently set in the vehicle control unit 10 may be checked based on the mode confirmation information.

In addition, the control unit 130 compares the signal collection information with a preset reference frequency pattern corresponding to a specific signal type identified for the signal collection information for each of a plurality of comparison attributes including signal strength and overlapping frequency, and compares the plurality of signals. It is possible to calculate a difference value between the signal collection information and the reference frequency pattern for each comparison attribute (S8).

In this case, the control unit 130 applies each of the plurality of signal types set in advance when the learning of the first learning model is completed as an input to the first learning model, and a plurality of reference frequencies respectively corresponding to the plurality of signal types. pattern can be calculated.

Accordingly, the control unit 130 may generate and store in advance setting information including a plurality of reference frequency patterns each matching a plurality of signal types and the plurality of signal types through the first learning model, and the first learning model. 1 When a new signal type is added to the learning model, a frequency pattern learned corresponding to the new signal type is calculated through the first learning model as a reference frequency pattern, and a reference frequency pattern and a new signal type corresponding to the new signal type are calculated through the first learning model. The setting information may be updated in addition to the setting information.

In addition, the control unit 130 may preset one or more comparison properties to be weighted for each of the plurality of operation modes and weights corresponding to comparison properties to be weighted.

Accordingly, the control unit 130 corresponds to a difference value for one or more comparison attributes, which is a weight application target, which is preset corresponding to a specific operation mode of the vehicle control unit 10 among the plurality of comparison attributes, and corresponds to the comparison attribute that is the weight application target. A final value for each of the plurality of comparison attributes may be calculated by applying a preset weight.

In this case, it goes without saying that the weights set in the comparison attribute, which is the weight application target, may be different between the plurality of operation modes.

For example, when the signal collection information is received (generated), the control unit 130 compares a specific operation mode currently set in the autonomous vehicle to a stop mode (S9), and a weight corresponding to the stop mode is applied. When the attribute is the overlapping frequency, a preset weight corresponding to the overlapping frequency, which is a weighting target corresponding to the stop mode, is applied to the difference value calculated corresponding to the overlapping frequency between the signal collection information and the reference frequency pattern to obtain a final value can be calculated (S11).

That is, the control unit 130 repeats a specific frequency form including one or more frequency components in the signal collection information three times so that the overlapping frequency is 3, and the reference frequency corresponding to the specific signal type identified for the signal collection information When the overlapping frequency according to the pattern is 1, the difference value for the overlapping frequency between the signal collection information and the reference frequency pattern is calculated as 2, and a weight preset corresponding to the overlapping frequency corresponding to the comparison attribute for the stop mode is preset. When is 1.5, 3 may be calculated as a final value corresponding to the overlapping frequency by applying a corresponding weight of 1.5 to the difference value calculated corresponding to the overlapping frequency.

In this case, the control unit 130 may calculate a difference value as a final value for a comparison attribute that is not a weight application target.

Alternatively, when the specific operation mode is a stop mode corresponding to a state in which the autonomous driving vehicle is stopped, the control unit 130 is configured to compare a difference value of a comparison property corresponding to an overlapping frequency among the one or more comparison properties to be weighted. A weight higher than the weight applied to the difference value of the attribute may be applied.

As another example, when the signal collection information is received (when generated), the specific operation mode currently set in the autonomous driving vehicle is the driving mode (S9), and the control unit 130 determines that a weight corresponding to the driving mode is applied. When the comparison property is signal strength, a difference in signal strength between the signal collection information and the reference frequency pattern is calculated, and the calculated difference value for the signal strength corresponds to the signal strength to which a weight corresponding to the driving mode is applied in advance. A final value may be calculated by applying the set weight (S10).

Alternatively, when the specific operation mode is a driving mode corresponding to a state in which the autonomous driving vehicle is driving, the controller 130 compares a difference value of a comparison attribute corresponding to a signal strength among one or more comparison attributes to which a weight is applied. A weight higher than the weight applied to the difference value of the attribute may be applied.

In addition, the control unit 130 may determine whether the signal collection information is normal by analyzing the final values for each of the plurality of comparison attributes in a weighted state according to a preset abnormality determination criterion (S12).

For example, as shown in FIG. 5 , the control unit 130 applies the final value for each of the plurality of comparison attributes calculated to correspond to the signal collection information to the abnormality determination criterion, and the final value for each of the plurality of comparison attributes The final value corresponding to the signal strength-related comparison attribute exceeds a preset reference value in relation to the signal strength-related comparison attribute (corresponding to the signal strength) in the abnormality determination criterion, and signals of all frequency components according to the signal collection information When the intensity exceeds the signal intensity of all frequency components according to the reference frequency pattern, the signal collection information may be determined as an abnormal signal related to hacking.

That is, the control unit 130 controls the signal strength of the radio signal collected through the antenna unit 110 to be abnormally higher than the signal strength of the normal signal, so the signal strength is high enough to interfere with the reception of other normal signals or to cover the normal signal. When is detected, it may be determined that the corresponding wireless signal is a signal for hacking attempts against the vehicle control unit 10 .

Alternatively, the control unit 130 applies the final value for each of the plurality of comparison attributes calculated to correspond to the signal collection information to the abnormality determination criterion, which corresponds to the comparison attribute related to the overlapping frequency among the final values for each of the plurality of comparison attributes. When the final value exceeds a preset reference value in relation to the comparison property related to the overlapping frequency preset in the abnormality determination criterion (corresponding to the overlapping frequency) and the specific frequency shape is abnormally repeated, the unique ID of the vehicle control unit 10 or By determining that it is a repeated hacking attempt, such as an attempt to steal a password, the signal collection information may be determined as an abnormal signal.

In the above configuration, when the autonomous driving vehicle is driving, the signal strength of the signal for the hacking attempt may be measured to be lower than when the autonomous driving vehicle is in a stopped state. By applying a weight to the signal strength in order to more accurately detect the signal for

In addition, since the control unit 130 facilitates repeated hacking attempts on the vehicle control unit 10 when the autonomous vehicle is in a stopped state, in order to more accurately detect repeated hacking attempts when the autonomous vehicle is in a stopped state, By applying a weight to the overlapping frequency, it is possible to increase the sensitivity to detection of hacking attempts when the vehicle is stopped so that the corresponding wireless signal can be detected as a hacking-related signal even when a wireless signal for a small number of hacking attempts is detected.

As described above, the present invention can increase the detection accuracy of the hacking attempt signal by assigning a weight to a comparison property that can increase the detection sensitivity of the signal for the hacking attempt according to the operating state of the autonomous driving vehicle.

On the other hand, the control unit 130 analyzes the final value for each of the plurality of comparison attributes according to the abnormality determination criterion to determine whether the signal collection information is normal (whether the radio signal according to the signal collection information is a normal signal) Result information may be generated, and warning information for warning of a hacking risk may be generated when it is determined as abnormal based on the analysis result information (S13).

In addition, the control unit 130 transmits the warning information to the vehicle control unit 10 through the communication unit 140 , and the vehicle control unit 10 that has received the warning information informs or disables the hacking attempt. can be made (S14).

Meanwhile, the control unit 130 may include setting information in which one or more reference frequency patterns are preset for each of the plurality of operation modes, and the setting information includes one or more types of signals corresponding to one or more reference frequency patterns for each operation mode. can be set.

At this time, as shown in FIG. 6 , the one or more signal types include a first signal type for unique ID identification of the vehicle control unit 10 , a second signal type for weather information, and a third signal type for multimedia content information. It may include a signal type, a fourth signal type for communication with a surrounding vehicle, and the like, and one or more signal types may be learned by the first learning model.

Accordingly, the control unit 130 sets a reference frequency pattern corresponding to both the specific signal type identified by applying the signal collection information to the first learning model and the specific operation mode identified by the vehicle control unit 10 as the setting information After extracting from , a difference value may be calculated by comparing the signal collection information with each of the plurality of comparison attributes.

Through this, the control unit 130 considers the signal type identified through the learning model with respect to the external wireless signal received by the hacking prevention device 100 and the frequency pattern in which the corresponding signal type changes according to the operating state of the autonomous vehicle Thus, it can be compared with the signal collection information, and through this, it is possible to more accurately detect an abnormal signal according to a hacking attempt.

As described above, the present invention collects an external wireless signal that wirelessly accesses the vehicle control unit 10 to detect a hacking attempt against the vehicle control unit 10 that controls the autonomous vehicle, and receives the signal from the vehicle control unit 10 . Through a learning model in which various types of normal signals are learned, it is determined whether a signal type corresponding to the corresponding external wireless signal exists, first determining whether a hacking attempt has been made, and then, if the signal type corresponding to the external wireless signal exists, the vehicle By comparing the normal frequency pattern for the corresponding signal type with the frequency pattern of the external radio signal according to the operation state of the By easily detecting a signal and notifying the vehicle control unit 10 of the risk of hacking, the vehicle control unit 10 supports to block hacking attempts before the control authority is seized by hacking, thereby safely protecting the autonomous vehicle from hacking. We can help you protect it.

On the other hand, the hacking prevention apparatus 100 according to the embodiment of the present invention learns the frequency pattern of the signal collection information determined to be an abnormal signal related to the hacking attempt according to the abnormality determination criterion, and based on this, it further generates a signal related to the hacking attempt. It can support accurate detection, which will be described in detail with reference to FIG. 7 .

As shown, the control unit 130 applies the signal collection information to a preset abnormality determination criterion to correspond to the signal collection information and the signal collection information when calculating the analysis result information and is identified through a first learning model. After generating training data including a specific signal type and analysis result information corresponding to the signal collection information in a state in which they are matched with each other, the training data may be trained in a preset second learning model.

Accordingly, a normal frequency pattern and an abnormal frequency pattern can be learned for each signal type in the second learning model, and the frequency pattern learned in the second learning model is a waveform shape and a repetition frequency (number of repetitions) of similar signals. ) and the duration of the signal, various properties necessary for distinguishing the normal frequency pattern from the abnormal frequency pattern may be reflected.

In addition, when the learning of the second learning model is completed, the controller 130 determines a specific signal type corresponding to the signal collection information when receiving the specific signal collection information from the signal collection unit 120 through the first learning model. It is possible to identify, apply the specific signal collection information and the specific signal type to the second learning model, and calculate the analysis result information corresponding to the specific signal collection information and the specific signal type through the second learning model have.

As an example, when the specific signal collection information and the specific signal type are received as inputs, the second learning model sets the normal frequency pattern and the abnormal frequency pattern learned corresponding to the specific signal type to a specific frequency according to the specific signal collection information. In comparison with the pattern, the degree of similarity to the normal frequency pattern and the degree of similarity to the abnormal frequency pattern are calculated for the specific signal collection information, and when the frequency pattern with the highest similarity to the specific signal collection information is the normal frequency pattern, the Calculating the analysis result information determined that the specific signal collection information is a normal signal, or calculating the analysis result information determined that the specific signal collection information is an abnormal signal when the frequency pattern with the highest similarity to the specific signal collection information is an abnormal frequency pattern can do.

In this case, the similarity may be a correlation or a correlation coefficient between the specific signal collection information and a frequency pattern learned in the learning model.

In addition, the controller 130 generates the warning information when it is determined that the signal collection information is an abnormal signal based on the analysis result information calculated by applying the signal collection information to the second learning model to generate the warning information to the communication unit 140 . can be transmitted to the vehicle control unit 10 through the .

In the above configuration, the control unit 130 includes the specific operation mode confirmed through the vehicle control unit 10 when receiving the signal collection information by matching it with the signal collection information in the learning data, and then includes the specific operation mode further. The included learning data can be trained in the second learning model, and through this, the correlation between the signal collection information and the analysis result information in which a specific operation mode and a specific signal type are reflected in the second learning model is determined by the second learning model. can be learned in

In addition, when the learning of the second learning model is completed, the control unit 130 sets a specific signal type identified based on specific signal collection information and a specific operation mode corresponding to the specific signal collection information together with the specific signal collection information. By applying the second learning model as an input, analysis result information corresponding to the specific signal collection information may be calculated through the second learning model.

That is, the control unit 130 determines the normal frequency pattern and the abnormal frequency pattern in which an arbitrary signal corresponding to a signal type identified for an external radio signal received by the autonomous vehicle is generally displayed in a specific operating state of the autonomous vehicle. By comparing the frequency pattern of the external wireless signal, it is possible to more accurately determine whether an external wireless signal is a signal related to a hacking attempt by comprehensively considering the change in the external wireless signal according to the signal type of the external wireless signal and the operating state of the autonomous vehicle. can

Meanwhile, in the above configuration, when generating the warning information, the control unit 130 may transmit the warning information to a user terminal corresponding to the user of the autonomous vehicle through the communication unit 140 , and the warning information is a text message. It may be in the form of a message.

In this case, the communication unit 140 may transmit warning information to the user terminal through various well-known communication methods.

Alternatively, the controller 130 may communicate with a service server that monitors the autonomous vehicle through the communication unit 140 , and transmits the warning information to the service server, and the service server receives the warning information when receiving the warning information. The warning information may be transmitted to the user terminal, which is a destination of the warning information. In this case, the warning information may be configured in the form of a push message.

Also, in the above configuration, the controller 130 configured in the autonomous driving vehicle may communicate with other hacking prevention devices configured in other autonomous driving vehicles through the communication unit 140 .

Accordingly, the controller 130 communicates with other hacking prevention devices of one or more other adjacent autonomous vehicles to receive common reception information, such as global positioning system (GPS) reception information, and receive one or more other information received in common. It may be determined whether a difference occurs in at least some of the common reception information received from the autonomous vehicle.

Also, when a difference occurs in at least some of the common reception information, the control unit 130 may transmit warning information to the vehicle control unit 10 of the autonomous vehicle through the communication unit 140 .

In this case, the common reception information may be GPS reception information, information on broadcast traffic, accidents, etc.

As described above, the wireless signal received by the autonomous driving vehicle includes various information such as the vehicle type of the autonomous driving vehicle and a unique ID for identification of the autonomous driving vehicle. Accordingly, the wireless signal received by any one autonomous driving vehicle The signal includes information that is different from the information included in the radio signal received by other autonomous vehicles, so that the radio signal received by the autonomous vehicle not only has a unique frequency pattern, but also has a variety of frequency patterns depending on the signal type of the radio signal. do.

Therefore, the present invention can support to easily distinguish between a normal signal and an abnormal signal for hacking based on learning the natural frequency pattern of a wireless signal having a different shape for each different autonomous vehicle for each signal type, and the vehicle It is possible to easily and accurately detect a wireless signal for a hacking attempt by considering the frequency pattern of the wireless signal that changes depending on the operating state of the vehicle, thereby safely protecting the autonomous vehicle from hacking.

The components described in the embodiments of the present invention include, for example, a storage unit such as a memory, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, and a field programmable (FPGA). One or more general-purpose computers or It can be implemented using a special purpose computer.

Those of ordinary skill in the art to which the present invention pertains may modify and modify the above-described contents without departing from the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

10: vehicle control unit 100: hacking prevention device
110: antenna unit 120: signal collecting unit
130: control unit 140: communication unit

Claims (10)

An antenna unit installed outside of a moving object capable of remote control or autonomous movement to receive a radio signal;
a communication unit communicating with a vehicle control unit that controls the moving object;
a signal collecting unit generating signal collection information by converting the radio signal into a frequency domain when receiving the radio signal through the antenna unit; and
A specific signal type corresponding to the signal collection information is identified by applying the signal collection information to a first learning model in which a plurality of frequency patterns for each signal type are learned, and the vehicle corresponding to a plurality of different operating states of the moving object, respectively A specific operation mode currently set in the vehicle control unit among a plurality of operation modes preset in the control unit is checked through the communication unit, and the identified specific signal type for each of a plurality of comparison properties including signal strength and overlap frequency of the signal collection information compares with a preset reference frequency pattern corresponding to , and calculates a difference value for each of the plurality of comparison attributes, and applies a preset weight for each of at least one comparison attribute corresponding to the specific operation mode among the plurality of comparison attributes to the difference value After calculating the final value for each of the plurality of comparison properties, the final value for each of the plurality of comparison properties is analyzed according to a preset abnormality determination criterion to determine whether the signal collection information is normal or not, and to reduce the risk of hacking when determined to be abnormal and a control unit that generates warning information for warning and transmits it to the vehicle control unit through the communication unit,
The control unit transmits the warning information through the communication unit to a preset user terminal corresponding to the user of the moving object.
The method according to claim 1,
When the learning of the first learning model is completed, the control unit applies each of a plurality of signal types preset in the first learning model to the first learning model to calculate a reference frequency pattern corresponding to the plurality of signal types, respectively, An apparatus for preventing hacking for a remote-controlled or autonomously movable moving object, characterized in that it generates and stores setting information for a plurality of reference frequency patterns matched with the plurality of signal types, respectively.
The method according to claim 1,
When the signal type corresponding to the signal collection information is not identified, the control unit generates query information for inquiring whether the signal collection information is normal and transmits it to the vehicle control unit, and receives the query information from the vehicle control unit Remote control or autonomous movement possible, characterized in that when the signal collection information is a normal signal based on the correspondingly received response information, a new signal type is created and then matched with the signal collection information and taught to the first learning model Anti-hacking device for moving objects.
4. The method according to claim 3,
wherein the control unit transmits warning information about the risk of hacking to the vehicle control unit when the signal collection information is abnormal based on the response information.
The method according to claim 1,
The control unit, each time the signal collection information is received, the signal collection information, the specific signal type identified for the signal collection information, and the learning data including the analysis result information on whether the signal collection information is normal in a state in which the signal collection information is matched with each other After generation, the second learning model is trained in advance to learn the abnormal frequency pattern and the normal frequency pattern for each of the plurality of signal types in the second learning model, and when the learning of the second learning model is completed, the signal collection unit receives it Remotely controlled or autonomously movable moving object, characterized in that by applying the collected signal information and a specific signal type corresponding to the signal collection information to the second learning model, analysis result information on whether the signal collection information is normal anti-hacking device for
6. The method of claim 5,
The control unit further includes a specific operation mode confirmed when the signal collection information is received in the training data to train the second learning model, and collects the signal collection information and the signal when learning for the second learning model is completed Hacking prevention for remotely controlled or autonomously movable moving objects, characterized in that by applying a specific signal type and specific operation mode corresponding to the information to the second learning model, analysis result information on whether the signal collection information is normal Device.
The method according to claim 1,
The control unit is
When the specific operation mode is a stop mode corresponding to a state in which the moving object is stopped, a weight higher than a weight applied to a difference value of another comparison attribute is applied to the difference value of the comparison attribute corresponding to the overlapping frequency among the one or more comparison attributes. apply,
When the specific operation mode is a driving mode corresponding to a state in which the moving object is driving, a weight higher than a weight applied to a difference value of another comparison attribute is applied to a difference value of a comparison attribute corresponding to a signal strength among the one or more comparison attributes. A hacking prevention device for a remote control or autonomously movable moving object, characterized in that it is applied.
delete In the hacking prevention method of a hacking prevention device that is installed outside of a moving object capable of remote control or autonomous movement, an antenna is installed to receive a radio signal, and communicates with a vehicle control unit that controls the moving object,
generating signal collection information by converting it into a frequency domain when receiving a radio signal through the antenna;
identifying a specific signal type corresponding to the signal collection information by applying the signal collection information to a learning model in which frequency patterns for each signal type are learned;
checking a specific operation mode currently set in the vehicle controller from among a plurality of operation modes preset in the vehicle controller respectively corresponding to a plurality of different operational states of the moving object;
Comparing the signal collection information with a preset reference frequency pattern corresponding to the identified specific signal type for each of a plurality of comparison properties including signal strength and overlapping frequency, a difference value for each of the plurality of comparison properties is calculated, and the specific operation determining whether the signal collection information is normal by applying a preset weight for each at least one comparison attribute corresponding to the mode to the difference value for each comparison attribute and analyzing it according to a preset abnormality determination criterion; and
When it is determined that the signal collection information is abnormal, transmitting warning information for warning of a risk of hacking to the vehicle control unit, and transmitting the warning information to a user terminal corresponding to a preset user of the moving object
A hacking prevention method for a remotely controlled or autonomously movable vehicle comprising a.
10. The method of claim 9,
The step of determining whether the normal
Further comprising the step of communicating with the hacking prevention device of another moving object and confirming whether a difference occurs in at least some of the common reception information including GPS reception information,
Transmitting the warning information includes:
The hacking prevention method for a remotely controlled or autonomously movable moving object, further comprising the step of transmitting warning information to the vehicle control unit when a difference occurs in at least some of the common reception information.

Images