Electronic control unit (ECU), autonomous vehicle having the ECU, and method for detecting a nearby vehicle therefor

Abstract

A method for detecting a nearby vehicle for an autonomous vehicle (10) having an electronic control unit (ECU), the method comprising: Extracting, by the electronic control unit (ECU), nearby vehicle information (S20) by detecting at least one nearby vehicle from a distance detected at each angle within a certain angle range, Calculating, by the electronic control unit (ECU), a false detection index (S30) indicating a degree of proximity of a shape of the at least one nearby vehicle to a shape of an exhaust gas based on the nearby vehicle information, Comparing, by the electronic control unit (ECU), the false detection index with a threshold false detection index, Executing, by the electronic control unit (ECU), an exhaust removal algorithm for correcting the nearby vehicle information (S60) when the false detection index is greater than the threshold false detection index, and Calculating, by the electronic control unit (ECU), a movement of the at least one nearby vehicle after a current frame (S40) by using the nearby vehicle information or the nearby vehicle information obtained by means of the exhaust removal algorithm, wherein the false detection index comprises at least one of a lateral movement index, a longitudinal movement index, a width change index and a length change index of the at least one nearby vehicle, where the lateral movement index is a product of a difference between the lateral change rate (y diff (t)) of at least one nearby vehicle and the lateral change rate (v y(t)) of a host vehicle with a ratio of the vehicle width (w(t)) of the at least one nearby vehicle to an average vehicle width (w car ) is, where the longitudinal motion index is a product of a difference between the longitudinal rate of change (x diff (t)) of the at least one nearby vehicle and the longitudinal rate of change (v x (t)) of a host vehicle with a ratio of the vehicle length (l(t)) of the at least one nearby vehicle to an average vehicle length (l bump ) is, where the width change index is an absolute value of the vehicle width change rate (w diff (t)) of at least one nearby vehicle, and where the length change index is an absolute value of the vehicle length change rate (l diff (t)) of at least one nearby vehicle.

Description

Background area

[0001] The present invention relates to an electronic control unit (ECU) for autonomous driving of a vehicle, an autonomous vehicle including the ECU, and a nearby vehicle detection method. More specifically, the present invention relates to a control unit capable of improving detection accuracy by considering exhaust gas characteristics when detecting a nearby vehicle, an autonomous vehicle including the ECU, and a nearby vehicle detection method therefor. Description of related technology

[0002] In recent years, there has been growing interest in autonomous navigation technology. Autonomous navigation technology refers to a technology that enables a vehicle to be driven automatically without driver intervention. Additionally, when the vehicle is traveling autonomously, the distance to a preceding vehicle is detected based on distance information obtained using a distance measurement sensor such as LiDAR, and operations such as lane change and smart cruise control (SCC) are executed based on such detection.

[0003] However, when detecting the distance from / to a preceding vehicle, exhaust gas emitted by the preceding vehicle may be misunderstood as a preceding vehicle (e.g., erroneously recognized as a preceding vehicle). If the exhaust gas is detected as a preceding vehicle, the behavior of the host vehicle (hereinafter also referred to as "host vehicle" or "subject vehicle") may be controlled without considering the actual movement of the preceding vehicle, which may cause a serious safety hazard to the driver or endanger a following vehicle due to sudden braking or the like. Therefore, there is a need for a technology for preventing the exhaust gas from being erroneously recognized as a vehicle.

[0004] Furthermore, US 2016 / 0154094 A1 discloses a method for detecting a vehicle located near an autonomous vehicle, the method comprising: autonomous driving logic of an electronic control unit (ECU) extracting nearby vehicle information by detecting at least one nearby vehicle from a distance detected at each angle within a certain angle range, calculating a false detection index indicating a degree of proximity of a shape of the nearby vehicle to a shape of an exhaust gas based on the nearby vehicle information, executing an exhaust removal algorithm for correcting the nearby vehicle information if the false detection index is greater than a threshold false detection index, and calculating a movement of the nearby vehicle after a current frame by using the nearby vehicle information or the corrected nearby vehicle information,which is obtained using the exhaust gas removal algorithm. Description of the invention

[0005] Accordingly, it is an object of the present invention to provide an ECU, an autonomous vehicle (e.g., a vehicle that can be driven automatically without driver intervention) having the ECU, and a method for detecting a nearby vehicle (hereinafter "near vehicle") therefor, which substantially obviate one or more problems caused by limitations and disadvantages of the related art.

[0006] Another object of the present invention is to provide an ECU capable of minimizing an influence of exhaust gas causing false detection during travel of an autonomous vehicle, and a nearby vehicle detection method therefor.

[0007] Additional advantages, objects, and features of the invention will be set forth in the following description and will become apparent to those skilled in the art upon examination of the following or may be learned from practice of embodiments of the invention. The features and advantages of the invention may be realized and appreciated by the arrangement particularly pointed out in the written description and claims as well as the appended figures.

[0008] To achieve advantages in accordance with embodiments of the invention, the present invention provides a method for detecting a vehicle located near an autonomous vehicle (e.g., autonomous motor vehicle, in particular, autonomous passenger vehicle) according to claim 1, an electronic control unit (ECU) for an autonomous vehicle according to claim 8, and an autonomous vehicle according to claim 13 with the inventive ECU. Advantageous further developments are described in the dependent claims.

[0009] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. Short description of the characters

[0010] The accompanying figures, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate one embodiment(s) of the invention and, together with the description, serve to explain features of the invention. In the drawings: is Fig. 1 is a block diagram schematically illustrating a vehicle according to an embodiment of the present invention, presents Fig. 2 shows an example of a false detection of an exhaust gas, is Fig. 3 a flowchart showing an operation / procedure of the autonomous driving logic, which in Fig. 1 is shown, presents Fig. 4 shows an example of a calculation of a false detection index, presents Fig. 5 illustrates an embodiment of an exhaust removal algorithm, and presents Fig. 6 an example of removing exhaust gas using the exhaust gas removal algorithm of Fig. 5. Detailed description of the invention

[0011] Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying figures. Wherever possible, the same reference numerals will be used throughout the figures to refer to the same or equivalent parts. The suffixes or terms "...module", "-module", "...unit", and "-unit" are added or used interchangeably to simplify the preparation of this description and are not intended to suggest particular meanings and functions. Accordingly, the terms "module" and "unit" may be used interchangeably.

[0012] According to one aspect of the present invention, a system and method for detecting a nearby vehicle are provided. At least one control unit or control device (hereinafter referred to as "control unit") of the vehicle processes information / signals from sensors installed in the vehicle, identifies at least one nearby vehicle, determines (repeatedly, periodically monitors) at least one parameter (location, speed, latitude, longitude) of the identified nearby vehicle, and generates signals for controlling the vehicle using the at least one parameter of the nearby vehicle.

[0013] In embodiments, the control unit calculates (periodically, repeatedly) by using information about the surroundings of the vehicle at a first time point (t1, a first time frame) for each nearby vehicle a parameter representing at least one of (1) a relative movement (distance, speed) between the subject vehicle and the nearby vehicle in a direction along the lane (longitudinal), (2) a relative movement (distance, speed) between the subject vehicle and the nearby vehicle in a direction across the lane (lateral), (3) a width of the nearby vehicle, and (4) a length of the nearby vehicle at the first time point. In embodiments, the control unit calculates the parameter for a second time point (t2, or a second time frame) later than the first time point and tracks a change in the parameter.

[0014] In embodiments, using the changes in the calculated parameter from the first time point to the second time point, the control unit calculates a misdetection index representing a probability that the exhaust from a nearby vehicle was misinterpreted as a structure of the nearby vehicle (or another) for the second time point, or at a second time point later than the second time point (e.g., a third time point). If the calculated misdetection index is greater than a predetermined threshold, the control unit applies an additional process (S60) to correct / adjust information from the sensors or to correct / adjust at least one parameter calculated from the information from the sensors.If the false detection index is less than the predetermined threshold, a location and / or speed of nearby vehicles undergo a normal process without applying the additional process / algorithm.

[0015] In embodiments, the greater the lateral relative motion (speed) between the subject vehicle and the nearby vehicle, the greater the false detection index. In embodiments, the smaller the longitudinal relative motion (speed), the smaller the false detection index. In embodiments, the greater the change in the width of the nearby vehicle, the greater the false detection index. In embodiments, the greater the change in the length of the nearby vehicle, the greater the false detection index.

[0016] Fig. 1 is a block diagram schematically illustrating a vehicle according to an embodiment of the present invention.

[0017] With reference to Fig. 1, the vehicle 10 is a vehicle (e.g., a motor vehicle, e.g., a passenger car) to which autonomous navigation technology is applied. Autonomous navigation technology refers to a technology that enables a vehicle to be driven automatically without driver intervention. Autonomous navigation technology aims to improve user comfort and safety by preventing accidents.

[0018] The vehicle 10 may include an information extraction unit 100, an autonomous driving logic 200, and a driving unit 300.

[0019] The information extraction unit 100 is configured to collect information about the (e.g., immediate) surroundings of the vehicle 10 and may include: a distance measurement sensor 110 configured to detect distance information about an object located in the surroundings of the vehicle 10, a camera 120 configured to obtain image information about the surroundings of the vehicle 10, and a speed sensor 130 configured to detect (e.g., measure) a speed at which the vehicle 10 is currently moving.

[0020] In particular, the distance measurement sensor 110 can be implemented with LiDAR, radar, or an ultrasonic sensor. The distance measurement sensor 110 can transmit (e.g., radiate) a signal of a specific wavelength forward in a time-of-flight (ToF) manner and then measure the distance to a preceding object using the difference in the detection time of a reflected signal of a specific wavelength.

[0021] The autonomous driving logic 200 may be software, hardware, or a combination thereof to implement the autonomous driving function. The autonomous driving logic 200 may be implemented as part of an electronic control unit (ECU) of the vehicle 10, but the present invention is not limited thereto.

[0022] The main logic (also main control logic) 210 may be responsible for overall control of the autonomous driving function. The main control logic 210 may generate a signal to control the driving unit 300 based on the information provided by a nearby vehicle information extraction unit 220, an exhaust misdetection elimination unit 230, and / or a nearby vehicle information calculation unit (e.g., nearby vehicle information determination unit, nearby vehicle information estimation unit) 240.

[0023] The nearby vehicle information extraction unit 220 may generate various types of information about a nearby vehicle based on at least one of the distance information from the distance measurement sensor 110, the image from the camera 120, and the speed information about the vehicle 10. The nearby vehicle information extraction unit 220 may also determine whether to execute the exhaust removal algorithm using various types of information. The exhaust removal algorithm refers to an algorithm executed by the exhaust false detection elimination unit 230.

[0024] The exhaust misdetection eliminating unit 230 can determine whether or not the detected nearby vehicle is an exhaust gas (e.g., an exhaust plume that is mistakenly detected as a nearby vehicle) and the type of exhaust gas by using various types of information provided from the nearby vehicle information extracting unit 220, and can correct the information about the corresponding nearby vehicle according to the determination.

[0025] The nearby vehicle information calculation unit 240 may use the position and speed of the nearby vehicle from the nearby vehicle information extraction unit 220 or the exhaust false detection elimination unit 230 to calculate (e.g., predict or estimate) a subsequent movement of the nearby vehicle.

[0026] The driving unit 300, which is configured to execute driving of the vehicle 10 according to the control signal of the main control logic 220, may comprise elements for substantially controlling driving of the vehicle, such as a brake, an accelerator (e.g., a drive unit, for example, an internal combustion engine and / or an electric motor), a gearshift (e.g., transmission), and a steering system.

[0027] For example, if the control signal of the main control logic 220 is a signal instructing a lane change to the left lane by means of acceleration, a driving control may be executed in which acceleration is applied by the accelerator of the driving unit 300 and a torque to the left is applied by the steering system.

[0028] Fig. 2 illustrates an example of false detection of exhaust gas.

[0029] Referring to Fig. 2, the left figure illustrates a case in which the exhaust gas from a nearby vehicle traveling on the same lane as the host vehicle (also referred to as the “subject vehicle” in some places herein) is not discharged at a first time point (t=t1) (which means that the exhaust gas from a moving vehicle is naturally discharged, but the amount of exhaust gas is not so large as to cause false detection by the distance measuring sensor), and the right figure illustrates a case in which the exhaust gas from the nearby vehicle traveling on the same lane is discharged at a second time point (t=t2, t2>t1) which is later than the first time point (t1).

[0030] First, in the left figure, the distance measuring sensor 110 emits a signal of a specific wavelength and detects a reflected signal during the host vehicle's movement. Since a reflection occurs at the rear body of the nearby vehicle (in Fig. 2, four reflection points are shown (filled circles), the distance of the nearby vehicle traveling in the same lane and the shape of the nearby vehicle, as detected by the distance measurement sensor 110, match the actual nearby vehicle using a distance from a first point of the subject vehicle to each of the four reflection points. In this case, controlling the driving unit 300 based on the distance information from the distance measurement sensor 110 does not cause any problem. A computer system of the subject vehicle repeatedly and / or periodically monitors the distance from t1 to t2.

[0031] However, in the right image (t=t2), the signal of a specific wavelength emitted forward from the distance measurement sensor 110 of the host vehicle is reflected by the exhaust gas emitted by the nearby vehicle, not by the vehicle body of the nearby vehicle. Accordingly, the distance to the nearby vehicle traveling in the same lane detected by the distance measurement sensor 110 (calculated by using a distance from the first point of the subject vehicle to each of the ten reflection points—solid circles) is shorter than the actual distance (and than the distance measured at the first time t1), and the detected shape of the nearby vehicle may be further broadened to the right side.In this case, when the traveling unit 300 is controlled based on the distance information from the distance measuring sensor 110, a control operation such as rapid deceleration or the like may be unnecessarily performed, which may be dangerous for the host vehicle and other nearby vehicles.

[0032] Fig. 3 is a flowchart showing an operation of the Fig. 1 represents the autonomous driving logic shown. Fig. Figure 4 shows an example of calculating a false detection index. Fig. 5 illustrates an embodiment of an exhaust removal algorithm. Fig. 6 shows an example of removing an exhaust gas by the exhaust removal algorithm Fig. 5.

[0033] Referring to the Fig. 3 to 6, during travel of the host vehicle (hereinafter, the vehicle 10 is referred to as a host vehicle or subject vehicle), the host vehicle information extraction unit 100 may collect information about the environment(s) around the host vehicle (S10). The collected information may include: distance information about a distance detected at each angle (e.g., with an (angular) increment of one degree) within a certain angular range (e.g., 120 degrees) in front of the distance measuring sensor 110, captured image information about the front (e.g., the area in front of the vehicle), and the travel speed of the host vehicle. The collected environment information may be transmitted to the near-vehicle information extraction unit 220.

[0034] The near-vehicle information extraction unit 220 may cluster distances detected at each angle within a certain angle range to at least one nearby vehicle.

[0035] For example, if objects are detected to have distances within a certain distance range within a continuous angular range, the objects can be grouped into a single nearby vehicle.

[0036] The nearby vehicle information extraction unit 220 may acquire two-dimensional coordinates of the corners (e.g., vertices of a polygon) (rear left corner, rear right corner, front left corner, and front right corner) of each nearby vehicle in a coordinate system centered on the host vehicle. If it fails to acquire the coordinates of some of the corners of the nearby vehicle directly from the distance information, the nearby vehicle information extraction unit 220 may determine (e.g., estimate) the shape of the nearby vehicle to obtain the coordinates of the corners for which it failed to acquire them. Various parameters (e.g., the minimum longitudinal coordinate, the minimum lateral coordinate, the maximum longitudinal coordinate, the maximum lateral coordinate, the longitudinal center coordinate (e.g., longitudinal center coordinate), the lateral center coordinate (e.g.,The lateral center coordinate, a vehicle width, and a vehicle length) of the nearby vehicle can be calculated based on the corner coordinates of the nearby vehicle. In this description, the longitudinal direction refers to the forward direction (or advancing direction) of the vehicle or the extending direction of the X-axis, and the lateral direction refers to the direction perpendicular to the forward direction (advancing direction) of the host vehicle or the extending direction of the Y-axis. Assuming that the position of the host vehicle is the origin (0, 0), the X-coordinate (e.g., the value of the longitudinal coordinate) increases when a point moves from left to right, and the Y-coordinate (e.g., the value of the lateral coordinate) increases when the point moves upward.

[0037] The nearby vehicle information extraction unit 220 may perform a calculation operation on the nearby vehicle parameters in the previous frame (e.g., previous data frame) and the nearby vehicle parameters in the current frame (e.g., current data frame) to calculate various parameters (e.g., a longitudinal change rate, a lateral change rate, a vehicle width change rate, and a vehicle length change rate). Here, the frame may refer to a set of nearby vehicle information generated at a specific time.

[0038] The nearby vehicle information extraction unit 220 may acquire the lane width of the lane in which the host vehicle is traveling and the lane information about the host vehicle and the nearby vehicles from an image from the camera 120. According to another embodiment, the nearby vehicle information extraction unit 220 may acquire the lane width of the lane in which the host vehicle is traveling and the lane information about the host vehicle and the nearby vehicles from a navigation program.

[0039] The near-vehicle information extraction unit 220 may obtain information about the traveling speed (longitudinal speed and lateral speed) of the host vehicle from the speed sensor 130.

[0040] That is, the nearby vehicle information extraction unit 220 can extract various kinds of information (hereinafter referred to as “nearby vehicle information”) about the nearby vehicles and the host vehicle required in the subsequent steps (S20).

[0041] The near-vehicle information extraction unit 220 may calculate a false detection index using the extracted information and may compare the false detection index with a threshold false detection index to determine whether to execute the exhaust removal algorithm (S30). The threshold false detection index may be a value experimentally determined taking into account a sensor error, etc. For example, the threshold false detection index may be 1.5, but the present invention is not limited thereto.

[0042] Fig. 4 shows an embodiment of a calculation of a false detection index (e.g., error detection index) P gas .

[0043] The false detection index P gas can be calculated for each nearby vehicle and can be calculated by calculating and summing the lateral movement index (e.g., sideways movement index), the longitudinal movement index (e.g., longitudinal movement index), the width change index, and the length change index of the nearby vehicle. According to one embodiment, the false detection index P gas not have at least one of the lateral movement index, the longitudinal movement index, the width change index and the length change index of the nearby vehicle.

[0044] The lateral movement index of the nearby vehicle is represented as the product (e.g. multiplication result) of the difference (absolute value) between the lateral change rate y diff(t) of the nearby vehicle and the lateral velocity v y (t) of the host vehicle with the ratio of the vehicle width w(t) of the nearby vehicle to the average vehicle width w car .

[0045] The lateral change rate y diff (t) of the nearby vehicle is the result obtained by subtracting the lateral coordinate of the center of the previous frame from the lateral coordinate of the center of the current frame (the average of the maximum coordinate and the minimum coordinate in the lateral direction) and dividing the subtraction result by an inter-frame time (e.g., time between two frames; English: "inter-frame time"), and indicates the speed (e.g., lateral speed) at which the nearby vehicle is moving in the lateral direction. Therefore, the difference (absolute value) between the lateral change rate y diff (t) of the nearby vehicle and the lateral velocity v y(t) of the host vehicle indicates the difference in lateral velocity between the nearby vehicle and the host vehicle.

[0046] Considering the case that the lateral movement speed (e.g., sideways movement speed) of the nearby vehicle would be almost equal to the lateral movement speed of the host vehicle, except in a special situation (e.g., during a lane change), when the nearby vehicle is moving in the same direction as the host vehicle, then a large difference between the lateral change rate y diff (t) of the nearby vehicle and the lateral velocity v y (t) of the host vehicle means that the shape of the nearby vehicle is close to the shape of exhaust gas (e.g., is similar to the shape of exhaust gas).

[0047] In addition, the average vehicle width w carthe average vehicle width of the nearby vehicle, which may be predetermined as the typical vehicle width of an average vehicle. According to a further embodiment, if the vehicle type of the nearby vehicle can be identified using other information (for example, image information), the vehicle width of the nearby vehicle can be determined by referring to a table in which pre-stored vehicle types are assigned to the corresponding vehicle widths. The vehicle width w(t) of the nearby vehicle can be calculated from the difference (absolute value) between the maximum lateral coordinate and the minimum lateral coordinate. The ratio of the vehicle width w(t) of the nearby vehicle to the average vehicle width w car can indicate whether the vehicle width of the nearby vehicle is abnormal or not.

[0048] Considering the case that the vehicle width of the nearby vehicle would be approximately equal to the average vehicle width of a normal nearby vehicle, then a large ratio of the vehicle width w(t) of the nearby vehicle to the average vehicle width w car indicate that the shape of the nearby vehicle is / is close to the shape of exhaust gas (e.g. is similar to the shape of exhaust gas).

[0049] That is, the lateral movement index of the nearby vehicle corresponds to an index indicating how close (e.g., similar) the shape of the nearby vehicle is to that of exhaust gas in terms of lateral movement (or shape (e.g., in the lateral direction)).

[0050] The longitudinal motion index of the nearby vehicle can be expressed as a product of the difference between the longitudinal change rate x diff (t) of the nearby vehicle and the longitudinal velocity v x(t) of the host vehicle with the ratio of the vehicle length l(t) of the nearby vehicle to the average vehicle length l bump .

[0051] The longitudinal rate of change x diff (t) of the nearby vehicle is the result obtained by subtracting the longitudinal coordinate of the center of the previous frame from the longitudinal coordinate of the center of the current frame (the average of the maximum coordinate and the minimum coordinate in the longitudinal direction) and dividing the subtraction result by an inter-frame time, and indicates the speed (e.g., longitudinal speed) at which the nearby vehicle is moving in the longitudinal direction. Therefore, the difference (absolute value) between the longitudinal change rate x diff (t) of the nearby vehicle and the longitudinal velocity v x(t) of the host vehicle indicates the difference in longitudinal velocity between the nearby vehicle and the host vehicle.

[0052] Considering the case where the longitudinal movement speed of the nearby vehicle is almost equal to the longitudinal movement speed of the host vehicle, except in a special situation (for example, during a lane change) when the nearby vehicle is moving in the same direction as the host vehicle, a small difference between the longitudinal change rate x diff (t) of the nearby vehicle and the longitudinal velocity v x (t) of the host vehicle means that the shape of the nearby vehicle is close to the shape of exhaust gas (e.g., is similar to the shape of exhaust gas).

[0053] Here, unlike the lateral movement index, the longitudinal movement index is the absolute value of the difference between the longitudinal change rate x diff(t) of the nearby vehicle and the longitudinal velocity v x (t) of the host vehicle. A negative value of the longitudinal motion index means that the nearby vehicle is moving rapidly away from the host vehicle when the difference between the longitudinal change rate x diff (t) of the nearby vehicle and the longitudinal velocity v x (t) of the host vehicle increases, which in turn means that executing the exhaust removal algorithm is less beneficial, regardless of the safety of the host vehicle, and therefore omitting (e.g., skipping) the algorithm may contribute to system efficiency.

[0054] In addition, the average vehicle length l bumpthe average vehicle length of the nearby vehicle, which may be predetermined as the typical vehicle length of an average vehicle. According to a further embodiment, if the vehicle type of the nearby vehicle can be identified using other information (for example, image information), the vehicle length of the nearby vehicle can be determined by referring to a table in which pre-stored vehicle types are assigned to the corresponding vehicle lengths. The vehicle length l(t) of the nearby vehicle can be calculated from the difference (absolute value) between the maximum longitudinal coordinate and the minimum longitudinal coordinate. The ratio of the vehicle length l(t) of the nearby vehicle to the average vehicle length l bump can indicate whether the vehicle length of the nearby vehicle is abnormal or not.

[0055] Considering the case that the vehicle length of the nearby vehicle is approximately the same as the average vehicle length of a normal (nearby) vehicle, a high ratio of the vehicle length l(t) of the nearby vehicle to the average vehicle length l bump indicate that the shape of the nearby vehicle is / is close to the shape of exhaust gas (e.g. is similar to the shape of exhaust gas).

[0056] That is, the longitudinal motion index of the nearby vehicle corresponds to an index indicating how close (e.g., similar) the shape of the nearby vehicle is to that of exhaust gas in terms of longitudinal motion (or shape (e.g., in the longitudinal direction)).

[0057] The width change index represents the absolute value of the vehicle width change rate w diff(t) of the nearby vehicle and can be calculated by dividing the difference between the vehicle width w(t) of the nearby vehicle in the current frame and the vehicle width w(t - 1) of the nearby vehicle in the previous frame by an inter-frame time.

[0058] Considering that a normal nearby vehicle will hardly show a change of length over time, except in a special situation (e.g. lane change), then a large absolute value of the vehicle width change rate w diff (t) of the nearby vehicle mean that the shape of the nearby vehicle is close to (e.g. similar to) that of exhaust gas.

[0059] The length change index represents the absolute value of the vehicle length change rate L diff(t) of the nearby vehicle and can be calculated by dividing the difference between the vehicle length l(t) of the nearby vehicle in the current frame and the vehicle length l(t - 1) of the nearby vehicle in the previous frame by an inter-frame time.

[0060] Considering that a normal nearby vehicle will hardly show any length change over time, except in a special situation (e.g. lane change), then a high absolute value of the vehicle length change rate l diff (t) of the nearby vehicle means that the shape of the nearby vehicle is close to the shape of exhaust gas (e.g. is similar to the shape of exhaust gas).

[0061] Accordingly, the false detection index P gas the degree of proximity (e.g., similarity) of the shape of the nearby vehicle to that of exhaust gas, and the nearby vehicle information extraction unit 220 compares the false recognition index P gaswith a threshold false discovery index. If the false discovery index P gas exceeds the threshold false detection index ("Yes" in S30), the information extraction unit 220 controls the exhaust false detection removal unit 230 to execute the exhaust removal algorithm. If the false detection index P gas is less than or equal to the threshold false detection index ("No" in S30), the information extraction unit 220 performs a control operation so that step S40 is executed without executing the exhaust gas removal algorithm.

[0062] The nearby vehicle information calculation unit 240 may calculate (e.g., predict or estimate) a movement of the nearby vehicle after the current frame by using the position and speed of the nearby vehicle among the various types of nearby vehicle information (or nearby vehicle information) of the nearby vehicle information extraction unit 220 and / or various types of corrected nearby vehicle information (or corrected nearby vehicle information). The nearby vehicle information calculation unit 240 may calculate the position of the nearby vehicle in the next frame, that is, a predicted longitudinal coordinate x p (t) of the center and a predicted lateral coordinate y p (t) of the center, predict.

[0063] Here, the calculation or determination process (e.g., estimation process) can be carried out by calculating the speed (the longitudinal change rate and the lateral change rate) from the detected position (the longitudinal coordinate and the lateral coordinate of the center) of the nearby vehicle and the inter-frame time and by calculating the predicted longitudinal coordinate x p (t) and the predicted lateral coordinate y p (t) of the center in the next frame, but embodiments of the present invention are not limited thereto.

[0064] The nearby vehicle information calculation unit 240 may calculate not only the position of the nearby vehicle in the next frame but also the position of the nearby vehicle in subsequent frames, and may also calculate other information (e.g., speed) in addition to the position of the nearby vehicle.

[0065] The nearby vehicle information calculation unit 240 may store the calculated information about the nearby vehicle after the current frame, that is, the predicted vehicle information (S50).

[0066] Fig. 5 shows an exhaust removal algorithm for correcting the nearby vehicle information, which is executed by the exhaust false detection removal unit 230.

[0067] The exhaust false detection eliminating unit 230 can determine whether the vehicle length change rate l diff (t) of the nearby vehicle exceeds a first threshold index δ1 (S100). The first threshold index δ1 may be a value experimentally determined taking into account a sensor error, a change in vehicle length at the time (e.g., at the time) of a lane change, and the like.

[0068] If the vehicle length change rate l diff(t) of the nearby vehicle exceeds a first threshold index δ1 (“Yes” in S100), this may mean that the longitudinal information about the nearby vehicle is unreliable, and the exhaust false detection eliminating unit 230 may determine whether the vehicle width change rate w diff (t) of the nearby vehicle exceeds a second threshold index δ2 (S110). The second threshold index δ2 may be a value experimentally determined taking into account a sensor error, a change in vehicle length at the time (e.g., at the time) of a lane change, and the like.

[0069] If the vehicle width change rate w diff(t) of the nearby vehicle is less than or equal to the second threshold index δ2 (“No” in S110), this means that the lateral information about the nearby vehicle is reliable, and therefore the exhaust false detection eliminating unit 230 will correct only the longitudinal information about the nearby vehicle.

[0070] If the X coordinate x rr (t) of the back right corner and the X-coordinate x rl (t) of the rear left corner are essentially equal to each other (for example, with a difference of several tens of centimeters) and the minimum longitudinal coordinate x min (t) is less than the X-coordinate x rl (t) of the rear left corner, the exhaust false detection eliminating unit 230 can determine the minimum longitudinal coordinate x min (t) to the average of the X-coordinate x rr (t) of the back right corner and the X-coordinate x rl(t) of the back left corner. This allows the longitudinal center coordinate x c (t) of the nearby vehicle can be calculated as the average of the corrected minimum longitudinal coordinate x min (t) and the maximum longitudinal coordinate x max (t) (S120).

[0071] Thereby, according to step S120, corruption of the longitudinal information about the nearby vehicle caused by misrecognizing the exhaust gas discharged from the rear center of the vehicle and accumulating, such as the exhaust gas shown in Figure (a) of Fig. 6, can be corrected by reducing the size (e.g., minimizing) of the nearby vehicle.

[0072] If the vehicle width change rate w diff(t) of the nearby vehicle exceeds the second threshold index δ2 ("Yes" in S110), this means that the lateral information about the nearby vehicle is also unreliable, and the exhaust false detection eliminating unit 230 therefore corrects both the longitudinal information and the lateral information about the nearby (e.g., neighboring) vehicle in step S130.

[0073] Since both the longitudinal information and the lateral information about the nearby vehicle acquired in the current frame are unreliable, the exhaust false detection eliminating unit 230 may use the predicted longitudinal center coordinate x p (t - 1) and the predicted lateral center coordinate y p (t - 1), which were calculated and stored in the previous frame, each assigned as the longitudinal center coordinate x c (t) and the lateral center coordinate y c(t) of the nearby vehicle in the current frame (e.g., set) (S130).

[0074] Thereby, according to step S130, corruption of the longitudinal information and the lateral information about the nearby vehicle caused by detection of the exhaust gas discharged in an irregular shape, such as the exhaust gas shown in Figure (b) of Fig. 6, as a nearby vehicle is conditioned, can be minimized.

[0075] If the vehicle length change rate l diff (t) of the nearby vehicle is less than or equal to the first threshold index δ1 (“No” in S100), this means that the longitudinal information about the nearby vehicle is reliable, and the exhaust false detection eliminating unit 230 can determine whether the vehicle width change rate w diff (t) of the nearby vehicle exceeds the second threshold index δ2 or not (S140).

[0076] If the vehicle width change rate w diff (t) of the nearby vehicle exceeds the second threshold index δ2 ("Yes" in S140), this means that the lateral information about the nearby vehicle is unreliable, and therefore the exhaust misdetection eliminating unit 230 will only correct the lateral information about the nearby vehicle in step S150.

[0077] If the maximum lateral coordinate y max (t) half the lane width of the host vehicle's lane wlane2 exceeds, the exhaust false detection elimination unit 230 can determine the maximum lateral coordinate y max (t) so that it is equal to half the lane width wlane2 If furthermore the minimum lateral coordinate y min (t) less than a negative value of half the lane width of the host vehicle's lane −wlane2 is, the exhaust false detection elimination unit 230 can determine the minimum lateral coordinate y min (t) so that it is equal to the negative value of half the lane width −wlane2 is (S150).

[0078] Here, the correction conditions and corrected values ​​are given assuming that the nearby vehicle is in the same lane as the host vehicle. If the nearby vehicle is in a different lane than the host vehicle, the correction conditions and corrected values ​​can be changed using the nearby vehicle's lane information.

[0079] For example, if the host vehicle is moving in the k-th lane (where k is an integer greater than or equal to 1) and the nearby vehicle is moving in the m-th lane (where m is an integer greater than or equal to 1), then if the maximum lateral coordinate y max (t) a value obtained by adding half the lane width of the host vehicle’s lane wlane2 to the product of the lane width of the lane and (mk), the exhaust false detection elimination unit 230 may determine the maximum lateral coordinate y max (t) so that it is equal to the value obtained by adding half the lane width of the host vehicle's lane wlane2 to the product of the lane width of the lane and (mk). Furthermore, if the minimum lateral coordinate y min(t) is less than a value obtained by adding the negative value of half the lane width of the host vehicle lane −wlane2 to the product of the lane width of the lane and (mk), then the exhaust false detection elimination unit 230 can determine the minimum lateral coordinate y min (t) so that it is equal to the value obtained by adding the negative value of half the lane width of the host vehicle's lane −wlane2 to the product of the lane width of the lane and (mk).

[0080] This allows the lateral center coordinate y c (t) of the nearby vehicle as the average of the corrected minimum lateral coordinate y min (t) and the corrected maximum lateral coordinate y max (t) can be calculated (S150).

[0081] Thereby, according to step S150, corruption of the longitudinal information about the nearby vehicle, which is caused by a detection of the exhaust gas which is discharged in a laterally spreading manner, such as the exhaust gas shown in Figure (c) of Fig. 6, as a nearby vehicle is conditioned, can be minimized.

[0082] If the vehicle width change rate w diff (t) of the nearby vehicle is less than or equal to the second threshold index δ2 ("No" in S140), this means that the lateral information about the nearby vehicle is also reliable. However, corruption of the longitudinal information about the nearby vehicle may cause a significant safety risk. Accordingly, step S160 is additionally performed.

[0083] That is, the exhaust false detection eliminating unit 230 can determine whether the longitudinal change rate x diff(t) of the nearby vehicle is less than a third threshold index δ3 (S160). The third threshold index δ3 may have a value experimentally determined taking into account a sensor error, the speed of the host vehicle, and the like. The third threshold index δ3 may be a negative number, but embodiments of the present invention are not limited thereto.

[0084] If the longitudinal rate of change x diff(t) of the nearby vehicle exceeds the third threshold index δ3, it means that the nearby vehicle is moving rapidly toward the host vehicle (e.g., the nearby vehicle and the host vehicle are rapidly approaching each other). Therefore, moving toward the host vehicle at a speed outside the normal range can be regarded as indicating exhaust gas discharged from the nearby vehicle and remaining where the gas was discharged, rather than the nearby vehicle.

[0085] If the longitudinal rate of change x diff (t) of the nearby vehicle is less than the third threshold index δ3 (“Yes” in S160), the exhaust false detection eliminating unit 230 may therefore, considering that the longitudinal information about the nearby vehicle acquired in the current frame is unreliable, calculate the predicted longitudinal center coordinate x p(t - 1), which was calculated and stored in the previous frame, as the longitudinal center coordinate x c (t) of the nearby vehicle in the current frame (e.g., set) (S170).

[0086] Thereby, according to step S170, corruption of the longitudinal information and the lateral information about the nearby vehicle, which is caused by a detection of the exhaust gas discharged from the nearby vehicle and remains where the gas is discharged, such as the exhaust gas shown in Figure (d) of Fig. 6, as a nearby vehicle is conditioned, can be minimized.

[0087] If the longitudinal rate of change x diff(t) of the near vehicle is greater than or equal to the third threshold index δ3 (“No” in S160), then the longitudinal information and the lateral information about the near vehicle acquired in the current frame can be regarded as reliable and thus step S40 can be performed using the uncorrected longitudinal information and the uncorrected lateral information about the near vehicle.

[0088] That is, in a vehicle according to an embodiment of the present invention, corruption of information about a nearby vehicle, such as the position and speed of the nearby vehicle, due to misrecognition of the exhaust gas discharged from the nearby vehicle as a part of the nearby vehicle can be minimized.

[0089] The method described above can be implemented in a computer-readable recording medium as code that is readable by a computer. The computer-readable recording medium includes all types of recording media configured to store data readable by the computer system. Examples of the computer-readable media include ROMs (read-only memories), RAMs (random access memories), magnetic tapes, magnetic floppy disks, flash memories, and optical data storage devices. The computer-readable recording media can also be distributed among computer systems connected via a network, and thus the computer-readable code can be stored and executed in a distributed manner.

[0090] As is apparent from the above description, embodiments of the present invention have effects as follows: An ECU, an autonomous vehicle including the ECU, and a nearby vehicle detection method therefor according to an embodiment of the present invention can minimize corruption of information, such as the position, speed, and the like, of a nearby vehicle due to misrecognition of the exhaust gas discharged from the nearby vehicle as a part of the nearby vehicle.

[0091] The effects that can be achieved by embodiments of the present invention are not limited to the above-mentioned effects, and other effects not mentioned herein may be clearly apparent to those skilled in the art from the above description.

[0092] Logical blocks, modules, or units described in connection with embodiments disclosed herein may be implemented or performed by a data processing device having at least one processor, at least one memory, and at least one communication interface. The components of a method, process, or algorithm described in connection with embodiments disclosed herein may be implemented directly in hardware, in a software module executed by at least one processor, or in a combination of the two. Computer-readable instructions for implementing a method, process, or algorithm disclosed in connection with embodiments disclosed herein may be stored in a non-transitory computer-readable storage medium.

[0093] Those skilled in the art will recognize that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Accordingly, it is intended that the present invention cover the modifications and variations of this invention provided they come within the scope of the appended claims.

Claims

[1] A method for detecting a nearby vehicle for an autonomous vehicle (10) having an electronic control unit (ECU), the method comprising: Extracting, by the electronic control unit (ECU), nearby vehicle information (S20) by detecting at least one nearby vehicle from a distance detected at each angle within a certain angle range, Calculating, by the electronic control unit (ECU), a false detection index (S30) indicating a degree of proximity of a shape of the at least one nearby vehicle to a shape of an exhaust gas based on the nearby vehicle information, Comparing, by the electronic control unit (ECU), the false detection index with a threshold false detection index, Executing, by the electronic control unit (ECU), an exhaust removal algorithm for correcting the nearby vehicle information (S60) when the false detection index is greater than the threshold false detection index, and Calculating, by the electronic control unit (ECU), a movement of the at least one nearby vehicle after a current frame (S40) by using the nearby vehicle information or the nearby vehicle information obtained by means of the exhaust removal algorithm, wherein the false detection index comprises at least one of a lateral movement index, a longitudinal movement index, a width change index and a length change index of the at least one nearby vehicle, where the lateral movement index is a product of a difference between the lateral change rate (y diff (t)) of at least one nearby vehicle and the lateral change rate (v y(t)) of a host vehicle with a ratio of the vehicle width (w(t)) of the at least one nearby vehicle to an average vehicle width (w car ) is, where the longitudinal motion index is a product of a difference between the longitudinal rate of change (x diff (t)) of the at least one nearby vehicle and the longitudinal rate of change (v x (t)) of a host vehicle with a ratio of the vehicle length (l(t)) of the at least one nearby vehicle to an average vehicle length (l bump ) is, where the width change index is an absolute value of the vehicle width change rate (w diff (t)) of at least one nearby vehicle, and where the length change index is an absolute value of the vehicle length change rate (l diff (t)) of at least one nearby vehicle. [2] The method of claim 1, wherein extracting the nearby vehicle information comprises: Grouping the at least one vehicle into a single nearby vehicle when it is detected that the at least one vehicle has a distance within a certain distance range within a continuous angular range. [3] The method according to claim 1 or 2, wherein the nearby vehicle information includes at least one of corner coordinates, a minimum longitudinal coordinate, a minimum lateral coordinate, a maximum longitudinal coordinate, a minimum lateral coordinate, a longitudinal center coordinate, a lateral center coordinate, a vehicle width, a vehicle length, a longitudinal change rate, a lateral change rate, a vehicle width change rate, and a vehicle length change rate of the at least one nearby vehicle. [4] The method according to any one of claims 1 to 3, wherein executing the exhaust gas removal algorithm (S60) comprises: Correct (S120) the minimum longitudinal coordinate (x min (t)), so that it is an intersection of an X-coordinate (x rr (t)) a back right corner and an X-coordinate (x rl (t)) of a back left corner if the X coordinate (x rr (t)) of the back right corner and the X coordinate (x rz (t)) of the back left corner must be within a certain range and the minimum longitudinal coordinate (x min (t)) is less than the X-coordinate (x rl (t)) the rear left corner, and Correcting the longitudinal center coordinate (x c (t)) of the at least one nearby vehicle (S120) such that it is an average of the minimum longitudinal coordinate and the maximum longitudinal coordinate (x max (t)) is, where correcting the minimum longitudinal coordinate (x min (t)) and correcting the longitudinal center coordinate (x c(t)) are executed (S120) when a vehicle length change rate (L diff (t)) of the at least one nearby vehicle exceeds a first threshold index (δ1) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle is less than or equal to a second threshold index (δ2). [5] The method according to any one of claims 1 to 4, wherein executing the exhaust gas removal algorithm (S60) comprises: Determine when a vehicle length change rate (L diff (t)) of the at least one nearby vehicle exceeds a first threshold index (δ1) (S100) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle exceeds a second threshold index (δ2) (S110), a predicted longitudinal center coordinate (x p (t - 1)) and a predicted lateral center coordinate (y p(t - 1)), which were calculated and stored in a previous frame, each assigned as a longitudinal center coordinate (x c (t)) and a lateral direction center coordinate (y c (t)) of the at least one nearby vehicle in a current frame (S130). [6] The method according to any one of claims 1 to 5, wherein executing the exhaust gas removal algorithm (S60) comprises: Correcting the maximum lateral coordinate (y max (t)) so that it is equal to half the lane width (wlane2) is (S150) when the maximum lateral coordinate (y max (t)) half of a lane width (wlane2) a lane on which a host vehicle is moving, and Correcting the minimum lateral coordinate (y min (t)) so that it is equal to a negative value of half the lane width (−wlane2) is (S150) if the minimum lateral coordinate (y min (t)) less than the negative value of half the lane width (−wlane2) is, where correcting the maximum lateral coordinate (y max (t)) and correcting the minimum lateral coordinate (y min (t)) (S150) when a vehicle length change rate (l diff (t)) of the at least one nearby vehicle is less than or equal to a first threshold index (δ1) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle exceeds a second threshold index (δ2). [7] The method according to any one of claims 1 to 6, wherein executing the exhaust gas removal algorithm (S60) comprises: Determine when a vehicle length change rate (L diff(t)) of the at least one nearby vehicle is less than or equal to a first threshold index (δ1) (S100), a vehicle width change rate (w diff (t)) of the at least one nearby vehicle is less than or equal to a second threshold index (δ2) (S140) and a longitudinal change rate (x diff (t)) of the at least one nearby vehicle is less than a third threshold index (δ3) (S160), a predicted longitudinal center coordinate (x p (t - 1)), which were calculated and stored in a previous frame, as a longitudinal center coordinate (x c (t)), of the at least one nearby vehicle in a current frame (S170). [8] Electronic control unit (ECU) for an autonomous vehicle (10), the ECU comprising: a nearby vehicle information extraction unit (220) configured to extract nearby vehicle information by detecting at least one nearby vehicle from a distance detected at each angle within a certain angle range, and to calculate a false detection index (P gas ) indicating a degree of proximity of a shape of the at least one nearby vehicle to a shape of an exhaust gas, based on the nearby vehicle information, an exhaust misdetection removing unit (230) configured to execute an exhaust removal algorithm for correcting the nearby vehicle information when the misdetection index (P gas ) is greater than a threshold false detection index, and a nearby vehicle information calculation unit (240) configured to calculate a movement of the at least one nearby vehicle after a current frame by using the nearby vehicle information or the corrected nearby vehicle information obtained by the exhaust removal algorithm, where the false detection index (P gas ) has at least one of a lateral movement index, a longitudinal movement index, a width change index and a length change index of the nearby vehicle, where the lateral movement index is a product of a difference between the lateral change rate (y diff (t)) of the nearby vehicle and the lateral change rate (v y (t)) of a host vehicle with a ratio of the vehicle width (w(t)) of the nearby vehicle to an average vehicle width (w car ) is, where the longitudinal motion index is a product of a difference between the longitudinal rate of change (x diff (t)) of the nearby vehicle and the longitudinal rate of change (v x (t)) of the host vehicle with a ratio of the vehicle length (l(t)) of the nearby vehicle to an average vehicle length (l bump ) is, where the width change index is an absolute value of the vehicle width change rate (w diff (t)) of the nearby vehicle, where the length change index is an absolute value of the vehicle length change rate (l diff (t)) of the nearby vehicle. [9] ECU according to claim 8, wherein when a vehicle length change rate (l diff (t)) of the at least one nearby vehicle exceeds a first threshold index (δ1) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle is less than or equal to a second threshold index (δ2), the exhaust gas false detection elimination unit (230) determines the minimum longitudinal coordinate (x min (t)) so that it is an average of an X-coordinate (x rr (t)) a back right corner and an X-coordinate (x rl (t)) of a back left corner when the X coordinate (x rr (t)) of the back right corner and the X coordinate (x rz (t)) of the back left corner are within a certain range and the minimum longitudinal coordinate (x min (t)) is less than the X-coordinate (x rz (t)) of the back left corner, and the longitudinal center coordinate (x c (t)) of the at least one nearby vehicle is corrected so that it is an average of the minimum longitudinal coordinate (x min (t)) and the maximum longitudinal coordinate (x max (t)). [10] ECU according to one of claims 8 and 9, wherein when a vehicle length change rate (l diff(t)) of the at least one nearby vehicle exceeds a first threshold index (δ1) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle exceeds a second threshold index (δ2), the exhaust false detection elimination unit (230) has a predicted longitudinal center coordinate (x p (t - 1)) and a predicted lateral center coordinate (y p (t - 1)), which were calculated and stored in a previous frame, each assigned as a longitudinal center coordinate (x c (t)) and a lateral direction center coordinate (y c (t)) of at least one nearby vehicle in a current frame. [11] ECU according to one of claims 8 to 10, wherein when a vehicle length change rate (l diff(t)) of the at least one nearby vehicle is less than or equal to a first threshold index (δ1) and a vehicle width change rate (w diff (t)) of the at least one nearby vehicle exceeds a second threshold index (δ2), the exhaust false detection elimination unit (230) the maximum lateral coordinate (y max (t)) so that it is equal to half the lane width (wlane2) is when the maximum lateral coordinate (y max (t)) half of a lane width (wlane2) a lane on which a host vehicle is moving, and the minimum lateral coordinate (y min (t)) so that it is equal to a negative value of half the lane width (−wlane2) is when the minimum lateral coordinate (y min (t)) less than the negative value of half the lane width (−wlane2) is. [12] ECU according to one of claims 8 to 11, wherein when a vehicle length change rate (L diff (t)) of the at least one nearby vehicle is less than or equal to a first threshold index (δ1), a vehicle width change rate (w diff (t)) of the at least one nearby vehicle is less than or equal to a second threshold index (δ2) and the longitudinal rate of change (x diff (t)) of the at least one nearby vehicle is less than a third threshold index (δ3), the exhaust false detection elimination unit (230) has a predicted longitudinal center coordinate (x p (t - 1)), which was calculated and stored in a previous frame, as a longitudinal center coordinate (x c (t)) of at least one nearby vehicle in a current frame. [13] Autonomous vehicle (10), comprising: the ECU according to any one of claims 8 to 12, and a driving unit (300) configured to control the driving of the vehicle according to a control signal generated by the ECU.

Images