KR102191169B1 - Method for preventing possible malfunctions of dcu occuring during autonomous driving by referring to ads using outputs of heterogeneous dcus and method using the same - Google Patents

Abstract

The present invention relates to a method for preventing a misjudgment situation by domain control units (DCUs), which may occur in autonomous driving through an anomaly detection system (ADS) using an output value of a heterogeneous DCU. According to the present invention, the method comprises the steps of: (a) obtaining first situation information to N^th situation information; (b) generating at least some of K_1^th determination information to K_M^th determination information; and (c) supporting at least some of a first DCU to an M^th DCU to be analyzed. According to the present invention, it is possible to prevent misjudgment of DCUs.

Description

METHOD FOR PREVENTING POSSIBLE MALFUNCTIONS OF DCU OCCURING DURING AUTONOMOUS DRIVING BY REFERRING TO ADS USING OUTPUTS OF HETEROGENEOUS DCUS through ADS using output values of heterogeneous DCUs AND METHOD USING THE SAME}

The present invention relates to a method for preventing a misjudgment situation of DCUs that may occur in autonomous driving through ADS using output values of heterogeneous DCUs, and to an apparatus using the same.

In recent years, autonomous driving technology is attracting attention as a great development. The self-driving technology that develops every year shows more and more advanced and sophisticated autonomous driving, but the reality is that there are still questions about its safety. For example, Tesla's self-driving vehicle, which is evaluated as having the most advanced autonomous driving technology, has reportedly reported deaths of occupants due to autopilot errors.

This risk of autonomous driving is because too many variables that may occur in an actual driving situation cannot be taken into account when performing a trial driving. In fact, Shashua, CEO of Mobileye, a leading company in the field of autonomous driving, needs a pilot mileage of more than 30 billion km to ensure safety in a statistical way, considering the frequency of accidents caused by autonomous vehicles driving on the road. And, such a pilot mileage is a level that can only be achieved when 4 million autonomous vehicles drive on the road for 20 hours every day, and even considering only the traffic, vehicle purchase cost, and computing cost resulting from this, it is $2.5 trillion. It has been argued that the above is not the level that general companies can bear.

Therefore, there is a need for a method that allows many variables of the actual driving situation to be considered even if the pilot driving is not performed by investing excessively large resources as described above, but it is true that such a method has not been studied much.

An object of the present invention is to solve the above-described problems.

In addition, an object of the present invention is to support the analysis of the DCUs in order to prevent misjudgment of DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs). .

Another object of the present invention is to provide a method of comparing determination information output from heterogeneous DCUs and detecting a problem of a specific DCU corresponding to specific determination information whose similarity to other determination information is less than or equal to a threshold. .

Another object of the present invention is to provide a method for an administrator to check a problem of a specific DCU and correct an algorithm of a specific DCU.

The characteristic configuration of the present invention for achieving the object of the present invention as described above and realizing the characteristic effects of the present invention described later is as follows.

According to an aspect of the present invention, support to analyze the DCUs to prevent the misjudgment situation of the DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs) In the method, (a) the computing device operating in conjunction with the target vehicle performing autonomous driving, a predetermined first to N-th time point corresponding to the autonomous driving state-N is an integer greater than or equal to 1-every time, the Causing a sensor module mounted in a target vehicle to acquire first situation information to an Nth situation information about a situation around the target vehicle; (b) the computing device allows at least some of the first DCU to M-th DCU operating in conjunction with the computing device-M is an integer of 2 or more-which is one of the first context information to the N-th context information. Generating at least a part of K_1th determination information to K_Mth determination information with reference to K context information-K is an integer of 1 or more and N or less; And (c) the computing device causes the ADS to operate in conjunction with the K_1-th determination information to at least part of the first DCU to the M-th DCU with reference to at least a portion of the K_1th determination information to the K_Mth determination information. At least a part of the first DCU to the M-th DCU can be analyzed by calculating the K-th decision match degree for the K-th decision match, and causing the edge logger to tag and store the K-th situation information with reference to the K-th decision match degree. Disclosed is a method comprising the step of assisting to perform.

As an example, in step (c), the computing device causes the ADS to apply a Dynamic Time Warping algorithm according to the following equation to at least a part of the K_1-th determination information to the K_M-th determination information, and the first DCU To calculate the K-th determination agreement for at least a portion of the M-th DCU,

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 적어도 하나에 대한 정보를 포함하는 제1 특정 시계열 벡터를 의미하고,

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 다른 하나에 대한 정보를 포함하는 제2 특정 시계열 벡터를 의미하는 것을 특징으로 하는 방법이 개시된다. In the above formula,

Denotes a first specific time series vector including information on at least one of the K_1th determination information to the K_Mth determination information,

Disclosed is a method, characterized in that denotes a second specific time series vector including information on another one of the K_1-th determination information to the K_M-th determination information.

As an example, (d) the computing device, when the edge logger causes the tag information of the K-th situation information to indicate that at least a part of the matching degree of the K-th determination is less than or equal to a threshold, the log analysis system The K-th situation information and the K_1-th determination information to the K_M-th determination information are transmitted, and the log analysis system transmits the K-th situation information and the K_1-th determination information to the K_M-th determination information through a predetermined display device. The method further comprising the step of supporting the manager to analyze at least some of the process processes of the first DCU to the M-th DCU at the K-th time point corresponding to the K-th context information by transmitting to the manager. Is initiated.

As an example, (e) when the computing device obtains analysis information about a problem in a specific DCU among the first DCU to the M-th DCU from the log analysis system, the specific DCU The method further comprising the step of modifying the algorithm of is disclosed.

As an example, in the step (a), the computing device obtains the first situation information to the N-th situation information by the sensor module including at least a part of a camera, a radar, a lidar, a GPS, and a V2X communication module. A method is disclosed, characterized in that the

As an example, at least some of the first DCU to the M-th DCU are implemented in the form of a neural network consisting of a plurality of layers each including a plurality of virtual neurons, and other parts are implemented in the form of a rule-based algorithm, Disclosed is a method, wherein each of the first DCU to the Mth DCU outputs a result according to a different logic.

As an example, in the step (b), the computing device includes at least one of the K_1th determination information to the K_Mth determination information generated from at least one preset main DCU among the first DCU to the Mth DCU. Disclosed is a method, characterized in that, by transmitting determination information to an actuator of the target vehicle, the target vehicle supports the autonomous driving according to the main determination information.

According to another aspect of the present invention, support for analyzing the DCUs to prevent misjudgment of the DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs). A computing device comprising: one or more memories for storing instructions; And one or more processors configured to perform the instructions, wherein the processor includes: (I) a predetermined first to N-th time point corresponding to the autonomous driving state-N is an integer greater than or equal to 1-each mounted on the target vehicle A process of causing the sensor module to acquire first situation information to Nth situation information about a situation around the target vehicle; (II) the first DCU to the M-th DCU operating in conjunction with the computing device-M is an integer greater than or equal to 2-K-th situation information, which is one of the first situation information to the N-th situation information-K Is an integer equal to or greater than 1 and equal to or less than N; And (III) the ADS operating in conjunction therewith, with reference to at least a part of the K_1-th determination information to the K_M-th determination information, and the K-th determination match degree with respect to at least a portion of the first DCU to the M-th DCU. A process for supporting at least part of the first DCU to the M DCU to be analyzed by calculating and storing the K-th context information by tagging the edge logger with reference to the K-th determination match degree. An apparatus is disclosed, characterized in that it performs.

As an example, in the (III) process, the processor causes the ADS to apply a Dynamic Time Warping algorithm according to the following equation to at least a part of the K_1-th determination information to the K_M-th determination information, and the first DCU to To calculate the K-th determination match degree for at least a portion of the M-th DCU,

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 적어도 하나에 대한 정보를 포함하는 제1 특정 시계열 벡터를 의미하고,

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 다른 하나에 대한 정보를 포함하는 제2 특정 시계열 벡터를 의미하는 것을 특징으로 하는 장치가 개시된다. In the above formula,

Denotes a first specific time series vector including information on at least one of the K_1th determination information to the K_Mth determination information,

A device is disclosed, characterized in that denotes a second specific time series vector including information on another one of the K_1-th determination information to the K_M-th determination information.

As an example, (IV) the processor, when the edge logger causes the tag information of the K-th context information to indicate that at least a part of the matching degree of the K-th determination is less than or equal to a threshold value, the The K situation information and the K_1th determination information to the K_Mth determination information are transmitted, and the log analysis system manages the Kth situation information and the K_1th determination information to the K_Mth determination information through a predetermined display device. The apparatus further performs a process of supporting the manager to analyze at least some of the process processes of the first DCU to the M-th time point corresponding to the K-th context information by transmitting to Is initiated.

As an example, when the processor (V) obtains analysis information about a problem in a specific DCU among the first DCU to the M-th DCU from the log analysis system, the specific DCU is An apparatus is disclosed which further performs the process of modifying the algorithm.

As an example, in the (I) process, the processor causes the sensor module including at least some of a camera, a radar, a lidar, a GPS, and a V2X communication module to obtain the first situation information to the Nth situation information. Disclosed is an apparatus, characterized in that.

As an example, at least some of the first DCU to the M-th DCU are implemented in the form of a neural network consisting of a plurality of layers each including a plurality of virtual neurons, and other parts are implemented in the form of a rule-based algorithm, An apparatus is disclosed, wherein each of the first DCU to the Mth DCU outputs a result according to a different logic.

As an example, in the (II) process, the processor may determine at least one of K_1-th determination information to K_M-th determination information generated from at least one preset main DCU among the first DCU to the M-th DCU Disclosed is an apparatus, characterized in that by transferring information to an actuator of the target vehicle, the target vehicle supports the autonomous driving according to the main determination information.

In addition, the present invention has an effect of supporting the analysis of the DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs). .

In addition, the present invention has the effect of providing a method of comparing the determination information output from heterogeneous DCUs and detecting a problem of a specific DCU corresponding to specific determination information whose similarity with other determination information is less than or equal to a threshold. have.

In addition, the present invention has an effect of providing a method for a manager to check a problem of a specific DCU and correct an algorithm of a specific DCU.

FIG. 1 is to analyze DCUs to prevent misjudgment of DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs) according to an embodiment of the present invention. A diagram showing a configuration of a computing device that performs a supporting method.
FIG. 2 is a diagram of a computing device performing a method of supporting analysis of DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing an interlocking relationship between components and components of a target vehicle.
3 is a flowchart illustrating a method of supporting analysis of DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention.
FIG. 4 is a dynamic time used while performing a method of supporting DCUs to be analyzed in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing the principle of the warping algorithm.
FIG. 5 is a display device used while performing a method of supporting DCUs to be analyzed in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing the information shown in.

DETAILED DESCRIPTION OF THE INVENTION The detailed description of the present invention to be described later refers to the accompanying drawings, which illustrate specific embodiments in which the present invention may be practiced. These embodiments are described in detail sufficient to enable a person skilled in the art to practice the present invention. It is to be understood that the various embodiments of the present invention are different from each other, but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the present invention in relation to one embodiment. In addition, it is to be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the present invention. Accordingly, the detailed description to be described below is not intended to be taken in a limiting sense, and the scope of the present invention, if properly described, is limited only by the appended claims, along with all scopes equivalent to those claimed by the claims. Like reference numerals in the drawings refer to the same or similar functions over several aspects.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those of ordinary skill in the art to easily implement the present invention.

FIG. 1 is to analyze DCUs to prevent misjudgment of DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs) according to an embodiment of the present invention. A diagram showing a configuration of a computing device that performs a supporting method.

Referring to FIG. 1, the computing device 100 may include at least one Anomaly Detection System (ADS) 130 and at least one edge logger 140. In this case, input/output and operation processes of the ADS 130 and the edge logger 140 may be performed by the communication unit 110 and the processor 120, respectively. However, in FIG. 1, a specific connection relationship between the communication unit 110 and the processor 120 is omitted. At this time, the memory 115 may be in a state in which various instructions to be described later are stored, and the processor 120 is set to execute instructions stored in the memory 115, but the processor 120 performs processes to be described later. By doing this, the present invention can be carried out. The description of the computing device 100 as described above does not exclude the case where the computing device 100 includes an integrated processor in which a medium, a processor, and a memory are integrated for implementing the present invention.

As described above, of the computing device 100 performing a method of supporting analysis of DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. I learned about the composition. Hereinafter, the interlocking relationship between this and the target vehicle 200 will be examined with reference to FIG. 2.

FIG. 2 is a diagram of a computing device performing a method of supporting analysis of DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing an interlocking relationship between components and components of a target vehicle.

Referring to FIG. 2, when K-th surrounding information to be described later is acquired through the sensor module 210 included in the target vehicle 200, the first DCU to the M-th DCU 220 included in the target vehicle 200 -1, ??, 220-M), after the K-th surrounding information is calculated, the K_1-th determination information or K_M-th determination information generated accordingly is transmitted to the ADS 130, and the ADS 130 uses this After calculating the K-th determination match, the configuration of generating tag information for the K-th state information and transmitting it to the edge logger 140 may be confirmed using this. In addition, the configuration in which the main determination information, which is one of the K_1th determination information to the K_Mth determination information, is transmitted to the actuator 230 included in the target vehicle 200 to perform autonomous driving may also be confirmed.

Under the assumption of such a relationship, using ADS for heterogeneous DCUs according to an embodiment of the present invention, a method of supporting DCUs to be analyzed to prevent misjudgment of DCUs that may occur in autonomous driving It will be described with reference to FIG. 3.

3 is a flowchart illustrating a method of supporting analysis of DCUs in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention.

Referring to FIG. 3, the computing device 100 includes a sensor module 210 mounted in the target vehicle 200 for each predetermined first to N-th time point corresponding to the autonomous driving state-N is an integer greater than or equal to 1- ) To obtain the first situation information to the Nth situation information about the situation around the target vehicle 200 (S01). Thereafter, the computing device 100 causes at least a portion of the first DCU to the M-th DCU 220-1, ??, 220-M to transmit the K-th context information, which is one of the first context information to the N-th context information. For reference, it is possible to generate at least some of the K_1th determination information to the K_Mth determination information (S02). In addition, the computing device 100 causes the ADS 130 to refer to at least a portion of the K_1th determination information to the K_Mth determination information, and the first DCU to the Mth DCU 220-1, ??, 220 -M) can be calculated to calculate the K-th determination match for at least some (S03). Finally, the computing device 100 causes the edge logger 140 to tag and store the K-th context information with reference to the K-th determination matching degree, so that the first DCU to the M-th DCU 220-1, ??, 220-M) may be supported so that at least some of them can be analyzed (S04).

Hereinafter, this will be described in more detail.

First, in step S01, the computing device uses the sensor module 210 described above, and the sensor module 210 may include at least some of a camera, a radar, a lidar, a GPS, and a V2X communication module. The first situation information to the N-th situation information at each point in time obtained through this is the information of the surrounding vehicle obtained through radar and the rider, the information of the surrounding vehicle obtained by calculating the information obtained through the camera by a predetermined neural network, and It may include a result of synthesizing information on surrounding vehicles acquired through the V2X communication module and location information of the target vehicle 200 acquired through GPS. The combination of the above sensors is not essential or limited, and other sensors may be included in the sensor module 210.

In addition, in step S02, the above-described context information, that is, the K-th context information, which is one of the first context information to the N-th context information, may be calculated. The first DCU to the M-th DCU 220-1, ??, 220-M) may be modules that output results according to different logics. That is, the first DCU to the M-th DCU (220-1, ??, 220-M) may be basically software or hardware modules designed to perform autonomous driving, and they are designed to output results according to different logics. As a result, it is designed to be a comparison group with each other. For example, at least some of the first DCU to the M-th DCU (220-1, ??, 220-M) are implemented in the form of a neural network consisting of a plurality of layers each including a plurality of virtual neurons, and other parts Can be implemented in the form of a Rule-based algorithm. DCUs implemented in the form of a neural network may have different layer configurations from each other, or may be learned in different ways.

And, in step S03, the computing device 100, the ADS (130), the first DCU to the M-th DCU (220-1, ??, 220-M) from at least some of the generated K_1 determination The K-th judgment matching degrees for at least some of the information to the K_M-th judgment information may be calculated by applying a dynamic time warping algorithm. The Dynamic Time Warping algorithm is an algorithm that is widely used in the futures trading field of stocks to measure the similarity between time series vectors of different lengths. Specifically, a Dynamic Time Warping algorithm may be performed according to the following equation.

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 적어도 하나에 대한 정보를 포함하는 제1 특정 시계열 벡터를 의미하고,

는 상기 제K_1 판단 정보 내지 상기 제K_M 판단 정보 중 다른 하나에 대한 정보를 포함하는 제2 특정 시계열 벡터를 의미할 수 있다. 이와 같은 Dynamic Time Warping 알고리즘이 수행됨에 따라, 각각의 제1 DCU 내지 제M DCU(220-1, ??, 220-M)들이 각각 판단한 결과가 비교될 수 있다. 이들 중 튀는 결과가 있을 경우, 해당 판단 정보와 다른 판단 정보 간의 제K 판단 유사도가 낮게 도출될 것이므로, 이를 찾아낼 수 있을 것이다. 제K 상황 정보는, 이와 같은 제K 판단 유사도들 중 임계치 이하인 것들이 존재할 경우, 이를 나타내도록 태그될 수 있다.In the formula,

Denotes a first specific time series vector including information on at least one of the K_1th determination information to the K_Mth determination information,

May mean a second specific time series vector including information on another one of the K_1-th determination information to the K_M-th determination information. As such a dynamic time warping algorithm is performed, the results determined by each of the first DCU to M-th DCUs 220-1, ??, and 220-M may be compared. If there is a bouncing result among these, the K-th judgment similarity between the corresponding judgment information and other judgment information will be derived, so that it will be possible to find it. The K-th situation information may be tagged to indicate when there are those of the K-th determination similarity that are less than or equal to a threshold.

Here, the reason why the Dynamic Time Warping algorithm should be performed will be described with reference to FIG. 4.

FIG. 4 is a dynamic time used while performing a method of supporting DCUs to be analyzed in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing the principle of the warping algorithm.

As described above, since each of the first DCU to M-th DCU 220-1, ??, 220-M is implemented with different logics, determination information will be generated in different styles. The use of the Dynamic Time Warping algorithm is to calculate the similarity between judgment information generated in different styles, that is, the K-th judgment match. For example, each determination information may include indication information for a different time range. 4 is a diagram illustrating a form of a graph that may appear when the indication information included in such determination information is digitized. In this case, referring to FIG. 4, the K_1th determination information may include instruction information for accelerating straight ahead for 2 seconds out of 5 seconds and turning left at a constant speed for 3 seconds, and the K_2 determination information is in 6 seconds. It can include instruction information to accelerate straight ahead for 2.5 seconds and turn left at constant speed for 3.5 seconds. In this case, both determination information may be viewed as somewhat similar, but it is difficult to conclude that both determination information are similar with a general Euclidean distance-based similarity calculation method. The Dynamic Time Warping algorithm was used because it can relatively accurately generate similarity, that is, the K-th judgment agreement, even for time series vectors having different time ranges.

The edge logger 140 stores the tagged K-th situation information to which information reflecting this, when at least a part of the K-th determination coincidence is less than the threshold, and then interlocks with a predetermined log analysis system (not shown) to perform the first At least some of the DCU to M-th DCUs 220-1, ??, 220-M may be supported to be analyzed.

Specifically, when the edge logger 140 indicates that the tag information of the K-th situation information indicates that at least a part of the matching degree of the K-th determination is less than or equal to a threshold value, the The K-th situation information and the K_1-th determination information to the K_M-th determination information may be transmitted. Thereafter, the computing device 100 causes the log analysis system to transmit the K-th situation information and the K_1-th determination information or the K_M-th determination information to the administrator through a predetermined display device, so that the administrator is the K-th time point corresponding to the K-th situation information. It may support to analyze at least some of the process processes of the first DCU to the Mth DCU 220-1, ??, 220-M. This will be described with reference to FIG. 5.

FIG. 5 is a display device used while performing a method of supporting DCUs to be analyzed in order to prevent a misjudgment situation of DCUs that may occur in autonomous driving by using ADS for heterogeneous DCUs according to an embodiment of the present invention. It is a diagram showing the information shown in.

Referring to FIG. 5, the administrator makes a determination that at least one of the first DCU to the M-th DCU 220-1, ??, 220-M is different from the others through the display device interlocked with the log analysis system. Information on how each of the first DCU to M-th DCU 220-1, ??, 220-M determined this, together with the K-th situation information corresponding to one situation, may be checked according to a predetermined UI. That is, at 22:05 on November 19th, when cars, trucks, and pedestrians are located according to the coordinates in the drawing, after checking the result of each DCU's decision through the above screen, the manager made the right decision by which DCU and which other It is possible to determine whether the DCU has made an incorrect judgment, and input analysis information on a specific DCU that has made an incorrect judgment into the log analysis system. Thereafter, the log analysis system transmits the analysis information to the computing device 100 again, and the computing device 100 may modify the algorithm of a specific DCU with reference to this.

However, since the above process is performed during the pilot driving, the process in which the pilot driving is performed will also be described. That is, at least one of the first DCU to the M-th DCU 220-1, ??, 220-M may be designated as the main DCU for performing the trial driving. In this case, the computing device 100 transmits the main determination information generated by the main DCU calculating the first situation information to the Nth situation information to the actuator of the target vehicle 200 so that the target vehicle 200 It can be supported to perform a trial run according to the following.

Through the above process, it is possible to obtain a feedback result considering various variables that may occur in an actual driving situation while properly performing the pilot driving.

The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded in the computer-readable recording medium may be specially designed and configured for the present invention, or may be known and usable to those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magnetic-optical media such as floptical disks. media), and a hardware device specially configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform processing according to the present invention, and vice versa.

In the above, the present invention has been described by specific matters such as specific elements and limited embodiments and drawings, but this is provided only to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , Anyone with ordinary knowledge in the technical field to which the present invention pertains can make various modifications and variations from these descriptions.

Therefore, the spirit of the present invention is limited to the above-described embodiments and should not be defined, and all modifications that are equally or equivalent to the claims as well as the claims to be described later fall within the scope of the spirit of the present invention. I would say.

Claims (14)

(a) A computing device operating in conjunction with a target vehicle performing autonomous driving, a predetermined first point to an Nth point in time corresponding to the autonomous driving state-N is an integer greater than or equal to 1-Each sensor mounted on the target vehicle Allowing the module to acquire first situation information to Nth situation information about a situation around the target vehicle;
(b) the computing device allows at least some of the first DCU to M-th DCU operating in conjunction with the computing device-M is an integer of 2 or more-which is one of the first context information to the N-th context information. Generating at least a part of K_1th determination information to K_Mth determination information with reference to K context information-K is an integer of 1 or more and N or less; And
(c) The computing device causes the ADS operating in conjunction therewith to refer to at least a part of the K_1th determination information to the K_Mth determination information to determine at least a portion of the first DCU to the Mth DCU. Support for at least part of the first DCU to the M-th DCU to be analyzed by allowing the K-th decision match to be calculated, and by causing the edge logger to tag and store the K-th situation information with reference to the K-th match to be determined. Steps to
The method comprising a. The method of claim 1,
The step (c),
The computing device causes the ADS to apply a Dynamic Time Warping algorithm according to the following equation to at least a portion of the K_1-th determination information to the K_M-th determination information to determine at least a portion of the first DCU to the M-th DCU. Calculate the degree of agreement of the K-th judgment,

In the above formula,

Denotes a first specific time series vector including information on at least one of the K_1th determination information to the K_Mth determination information,

Means a second specific time series vector including information on the other one of the K_1th determination information to the K_Mth determination information. The method of claim 1,
(d) When the computing device causes the edge logger to indicate that the tag information of the K-th situation information indicates that at least a part of the matching degree of the K-th determination is equal to or less than a threshold, the K-th situation Information and the K_1th determination information to the K_Mth determination information, and causes the log analysis system to transmit the Kth situation information and the K_1th determination information to the K_Mth determination information to an administrator through a predetermined display device. Supporting the manager to analyze at least some of the process processes of the first DCU to the M DCU at the Kth point of time corresponding to the Kth context information by transmitting
The method further comprising a. The method of claim 3,
(e) When the computing device obtains analysis information about a problem in a specific DCU among the first DCU to the M-th DCU from the log analysis system, the algorithm of the specific DCU is determined by referring to the analysis information. Steps to fix
The method further comprising a. The method of claim 1,
The step (a),
Wherein the computing device causes the sensor module including at least some of a camera, a radar, a lidar, a GPS, and a V2X communication module to obtain the first situation information to the Nth situation information. The method of claim 1,
At least part of the first DCU to the M-th DCU is implemented in the form of a neural network consisting of a plurality of layers each including a plurality of virtual neurons, and the other part is implemented in the form of a rule-based algorithm, each of the The first DCU to the M-th DCU outputting results according to different logics. The method of claim 1,
The step (b),
The computing device transmits at least one of the K_1th determination information to the K_Mth determination information, generated from at least one preset main DCU among the first DCU to the Mth DCU, to the actuator of the target vehicle Thereby supporting the target vehicle to perform the autonomous driving according to the main determination information. In a computing device that supports analysis of the DCUs to prevent a misjudgment situation of the DCUs that may occur in autonomous driving by using an Anomaly Detection System (ADS) for heterogeneous Domain Control Units (DCUs),
One or more memories for storing instructions; And
Including one or more processors configured to perform the instructions,
The processor, (I) a predetermined first time point to the N-th time point corresponding to the autonomous driving state-N is an integer greater than or equal to 1-each time, the sensor module mounted in the target vehicle to determine the situation around the target vehicle A process of obtaining 1 context information to an Nth context information; (II) the first DCU to the M-th DCU operating in conjunction with the computing device-M is an integer greater than or equal to 2-K-th situation information, which is one of the first situation information to the N-th situation information-K Is an integer equal to or greater than 1 and equal to or less than N; a process of generating at least a part of K_1th determination information to K_Mth determination information with reference to; And (III) the ADS operating in conjunction therewith, with reference to at least some of the K_1-th determination information to the K_M-th determination information, and the K-th determination match degree for at least a portion of the first DCU to the M-th DCU. A process for supporting at least some of the first DCU to the M DCU to be analyzed by calculating and storing the K-th context information by tagging the edge logger with reference to the K-th determination agreement
Device, characterized in that to perform. The method of claim 8,
The (III) process,
The processor causes the ADS to apply a Dynamic Time Warping algorithm according to the following equation to at least a portion of the K_1-th determination information to the K_M-th determination information, and the first DCU to the M-th DCU for at least a portion of the Calculate the K-th judgment agreement,

In the above formula,

Denotes a first specific time series vector including information on at least one of the K_1th determination information to the K_Mth determination information,

Means a second specific time series vector including information on the other one of the K_1-th determination information to the K_M-th determination information. The method of claim 8,
(IV) When the processor causes the edge logger to indicate that the tag information of the K-th situation information indicates that at least a part of the matching degree of the K-th determination is less than or equal to a threshold, the K-th situation information And transmitting the K_1th determination information to the K_Mth determination information, and causing the log analysis system to transmit the Kth situation information and the K_1th determination information to the K_Mth determination information to an administrator through a predetermined display device. By doing so, a process for supporting the manager to analyze at least some of the process processes of the first DCU to the M DCU at the Kth time point corresponding to the Kth context information
The device, characterized in that to further perform. The method of claim 10,
The processor,
(V) A process of modifying an algorithm of the specific DCU with reference to the analysis information when analysis information on a problem exists in a specific DCU among the first DCU to the M-th DCU from the log analysis system
The device, characterized in that to further perform. The method of claim 8,
The (I) process,
The apparatus, wherein the processor causes the sensor module including at least some of a camera, a radar, a lidar, a GPS, and a V2X communication module to obtain the first situation information to the Nth situation information. The method of claim 8,
At least part of the first DCU to the M-th DCU is implemented in the form of a neural network consisting of a plurality of layers each including a plurality of virtual neurons, and the other part is implemented in the form of a rule-based algorithm, each of the The device, characterized in that the first DCU to the Mth DCU output results according to different logics. The method of claim 8,
The (II) process,
The processor transmits at least one of the K_1-th determination information to the K_M-th determination information, generated from at least one preset main DCU among the first DCU to the M-th DCU, to the actuator of the target vehicle. And supporting the target vehicle to perform the autonomous driving according to the main determination information.

Images