Method and system for identifying a turning object

Abstract

The present disclosure relates to a method for identifying a turning object, comprising detecting a front driving vehicle by an environmental sensor mechanism of a host vehicle; determining a turning probability of the front driving vehicle; and labeling the front driving vehicle as a turning object when the turning probability is equal to or greater than a predetermined threshold.

Description

Technical Field

[0001] The present disclosure relates to a method for identifying a turning object, a storage medium for executing the method, a system for identifying a turning object, and a vehicle having such a system. The present disclosure particularly relates to reliable identification of turning vehicles, for example for adaptive cruise control of the vehicle. Background Art

[0002] Driver assistance systems for automated driving are becoming increasingly important. Automated driving can be performed at different levels of automation. Exemplary levels of automation are assisted, partially automated, highly automated, or fully automated driving. These levels of automation are defined by the Federal Highway Research Institute (BASt) (see the BASt publication "Forschung kompakt," published in November 2012). For example, Level 4 vehicles are fully autonomous in urban operations.

[0003] Driver assistance systems for autonomous driving use sensors that perceive the environment based on vision, both within and beyond the human-visible range. These sensors can be cameras, radar, and / or lidar. In addition to high-precision maps, these sensors are the primary signal sources for driver assistance systems used in autonomous driving.

[0004] Adaptive cruise control (ACC) is now commonly used in vehicles. ACC is a cruise control system that takes the distance to the vehicle ahead into account as an additional feedback and control variable. With ACC, the position and speed of the vehicle ahead are determined by sensors, and the speed and distance are adaptively controlled through engine and brake intervention. Depending on the movement pattern of the vehicle ahead, this longitudinal control can result in unnecessary braking of the vehicle, which can be uncomfortable for the occupants. Summary of the Invention

[0005] One object of the present disclosure is to provide a method for identifying a turning object, a storage medium for executing the method, a system for identifying a turning object, and a vehicle having such a system, which enable improved turn detection. In particular, the present disclosure is directed to improving adaptive cruise control.

[0006] According to an independent aspect of the present disclosure, a method for identifying a turning object is provided. The method includes: detecting a leading vehicle using an environment sensor of a host vehicle; determining a turning probability of the leading vehicle; and marking the leading vehicle as a turning object when the turning probability is equal to or greater than a predetermined threshold.

[0007] The predetermined threshold value may be, for example, 50% or more, 60% or more, 70% or more, 80% or more, or 90% or more.

[0008] According to the present invention, it is possible to identify whether the vehicle ahead is turning. This is achieved using a turn probability. When determining the turn probability or identifying the turning object, a number of criteria can be used, which are designed to increase or optimize the probability that the vehicle ahead is actually turning.

[0009] Specifically, when the turn probability is equal to or greater than a predetermined threshold, the turn probability of the vehicle ahead of the host vehicle is determined and the vehicle is marked as a turning target. For example, the marked vehicle's turn can be taken into account when regulating or controlling the vehicle's speed. The turn probability can be used, for example, to adjust controller weights. This allows the controller to reset positive torque as soon as a turn is detected. This prevents excessive braking and prolonged braking during a turn, allowing the host vehicle to accelerate earlier.

[0010] Preferably, the turning probability of the preceding vehicle is determined based on at least one first criterion. The at least one first criterion may be selected from the group consisting of: turning out, turn signal, lateral speed, relative speed, object speed, and heading.

[0011] Preferably, the method further comprises scaling the turn probability using at least one second criterion. The at least one second criterion may be selected from the group consisting of: turn distance, turn curvature, longitudinal turn speed, lateral turn speed, distance to a roundabout, distance to a T-junction, lateral turn distance, and lateral turn acceleration.

[0012] Preferably, the method further comprises applying an impact value to at least one of the at least one first criterion and the at least one second criterion. The impact value may be between 0 and 1, where 0 specifies no impact and 1 specifies maximum impact.

[0013] Preferably, the method further comprises debouncing at least one of the at least one first standard and the at least one second standard. The debouncing is performed using a debouncing time which can be suitably selected and can be in the range of a few seconds depending on the standard.

[0014] Preferably, only when the direction of the road allows the steering identification, the vehicle ahead is marked as the steering object. If not allowed, the vehicle ahead will not be marked as the steering object and therefore, for example, will not be taken into account in the speed control of the vehicle.

[0015] According to another independent aspect, a software (SW) program is presented. The SW program can be designed to execute on one or more processors to perform the method for identifying a turning object described herein.

[0016] According to another independent aspect, a storage medium is presented. The storage medium can comprise a SW program designed to execute on one or more processors to perform the method for identifying a turning object described herein.

[0017] According to another independent aspect of the present disclosure, a system for identifying a turning object is presented. The system comprises a detection module designed to detect a vehicle driving ahead by means of environmental sensors of the ego vehicle, at least one processor unit designed to determine a turning probability of the vehicle driving ahead and to label the vehicle driving ahead as a turning object if the turning probability is equal to or greater than a predetermined threshold.

[0018] The system can implement various aspects of the method for identifying a turning object described herein. Similarly, the method can implement various aspects of the system for identifying a turning object described herein.

[0019] Preferably, the system comprises a regulation module designed to adjust a speed of the vehicle based on the turning probability of the vehicle driving ahead.

[0020] Preferably, the control module can be designed to adjust the speed of the vehicle based on the turning probability only if the vehicle driving ahead is labeled as turning. If the vehicle driving ahead is not labeled as turning, speed control can be performed without turning, thereby improving safety of the system.

[0021] The detection module and / or the at least one processor unit and / or the regulation module can be implemented in a common software module and / or hardware module. Alternatively, the detection module and / or the at least one processor unit and / or the regulation module can be implemented in separate software modules and / or hardware modules, respectively.

[0022] The system is preferably designed for autonomous driving. In particular, the driver assistance system can be an adaptive cruise control (ACC).

[0023] The adaptive cruise control is designed to maintain a safe distance to the vehicle driving ahead. In some embodiments, the adaptive cruise control can comprise a speed setpoint mode and a time headway mode. In the speed setpoint mode, a speed set by a driver or a desired speed is to be maintained. In the time headway mode, a time advantage over the vehicle driving ahead is to be maintained.

[0024] In the context of this document, the term "automated driving" can be understood as driving with automatic longitudinal or lateral guidance or automated driving with automatic longitudinal and lateral guidance. This can include, for example, driving on motorways for extended periods or for limited periods during parking or maneuvering. The term "automated driving" encompasses automated driving with any degree of automation. Exemplary degrees of automation are assisted, partially automated, highly automated, or fully automated driving. These degrees of automation are defined by the Federal Highway Research Institute (BASt) (see the BASt publication "Forschungkompakt," published in November 2012).

[0025] With assisted driving, the driver continuously provides longitudinal or lateral guidance, while the system takes over other functions within certain limits. With partially automated driving (TAF), the system takes over longitudinal and lateral guidance for a period of time and / or in specific situations, with the driver constantly monitoring the system, just as with assisted driving. With highly automated driving (HAF), the system takes over longitudinal and lateral guidance for a period of time, with the driver constantly monitoring the system; however, the driver must be able to take over the vehicle within certain periods of time. With fully automated driving (VAF), the system automatically handles driving in all situations for the specific application; in this case, the driver is no longer required.

[0026] The four levels of automation mentioned above correspond to SAE Levels 1 to 4 in the SAE J3016 standard (SAE - Society of Automotive Engineering). For example, highly automated driving (HAF) corresponds to Level 3 in the SAE J3016 standard. Furthermore, SAE J3016 specifies SAE Level 5 as the highest level of automation, which is not included in the BASt definition. SAE Level 5 corresponds to autonomous driving, in which the system can automatically handle all situations throughout the entire drive, just like a human driver; a driver is generally no longer required.

[0027] According to another independent aspect of the present invention, a vehicle, in particular a motor vehicle, is provided, comprising a system for identifying a turning object according to an embodiment of the present disclosure.

[0028] The term vehicle includes cars, trucks, buses, RVs, motorcycles, etc., which are used to transport people, goods, etc. In particular, the term includes motor vehicles used to transport people. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Embodiments of the present disclosure are shown in the accompanying drawings and are described in more detail below.

[0030] Figure 1Ais a flowchart of a method for identifying a turning object according to an embodiment of the present disclosure;

[0031] Figure 1B is schematically shown a system for identifying a turning object according to an embodiment of the present disclosure;

[0032] Figures 2 to 4 is shown a criterion for turning identification according to an embodiment of the present disclosure;

[0033] Figure 5 is shown a driving path according to an embodiment of the present disclosure; and

[0034] Figures 6 to 14 is shown a further criterion for turning identification according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0035] In the following, identical reference signs are used for identical and identically acting components, unless stated otherwise.

[0036] According to an embodiment of the present disclosure, it is identified whether a front driving vehicle is turning. This is achieved by a turning probability. A number of criteria can be employed in determining the turning probability or identifying a turning object, which are designed to be able to increase or optimize the probability that a front driving vehicle is actually turning.

[0037] Figure 1A is shown a flowchart of a method 100 for identifying a turning object according to an embodiment of the present disclosure. The method 300 can be implemented by a corresponding software, which can be executed by one or more processors (e.g. a CPU).

[0038] The method 100 comprises: in a block 110, identifying a front driving vehicle by means of environmental sensors of the ego vehicle; in a block 120, determining a turning probability of the front driving vehicle; and in a block 130, marking the front driving vehicle as a turning object, if the turning probability is equal to or greater than a predetermined threshold value.

[0039] A system 200 corresponding to the method is shown in Figure 1B and comprises a detection module 210 designed for detecting a front driving vehicle by means of environmental sensors of the ego vehicle, and at least one processor unit 220 designed for determining a turning probability of the front driving vehicle and marking the front driving vehicle as a turning object, if the turning probability is equal to or greater than a predetermined threshold value.

[0040] The driver assistance system 100 can be designed for adaptive cruise control (ACC), for example. Adaptive cruise control is a cruise control device that takes the distance to a vehicle driving ahead as an additional feedback and control variable when controlling. In adaptive cruise control, the position and speed of the vehicle driving ahead are determined by sensors, and the speed and distance are controlled adaptively (longitudinal control) by engine and brake intervention.

[0041] For planning and implementing such automated driving, the driver assistance system 200 obtains environmental information from environmental sensors that monitor the environment of the vehicle. In particular, the vehicle can comprise at least one environmental sensor that is designed to acquire environmental data indicative of the environment of the vehicle. The at least one environmental sensor can comprise a laser radar system, one or more radar systems and / or one or more cameras, for example.

[0042] Exemplary criteria for the turn recognition are explained below. The following criteria, aspects and partial aspects can be combined and used in a suitable manner here. For example, at least partially different criteria can be employed for the swerving out (low lateral speed and lateral acceleration) and the turning (high lateral speed and increased lateral acceleration).

[0043] An exemplary function for the turn recognition according to the present application can be defined as follows:

[0044] Turn probability * turn scale + road clear

[0045] The product of the turn probability and the turn scale can be maximally 1 and can have a hysteresis.

[0046] Turn probability (turn sum)

[0047] The turn probability can be represented by a turn sum. The turn sum can comprise two or more addends that influence or indicate the turn probability.

[0048] The addends can each form a contribution to the sum, which is defined as the product of the value and the influence. In particular, an influence or weight can be defined and included for each addend. The influence or weight can be specified as a scalar. For example, the influence or weight of the turn signal can be 0.3. The value of the "turn signal" can be 0 (turn signal not set) or 1 (turn signal not set), wherein the weighted value in the sum is then 0.3 (1 · 0.3).

[0049] In an exemplary embodiment, the turn sum can be defined as follows:

[0050] Swerve out + turn signal + lateral speed + relative speed + object X speed + heading (+ swerve out)

[0051] A scale and optional influence can be set for the kick-out contribution. Figure 2 is a graph (0-1 function) showing the turn-out value as a function of the turn-out probability. In this example, starting from a probability of 50%, the contribution of the turn-out to the total turn-out (scale) increases. The maximum value is 1, and the influence can be set. For example, if the influence is 0.4, the maximum contribution of the turn-out to the total turn-out is 1·0.4=0.4.

[0052] If the value of the knock-out probability drops below a certain value, for example 0.9, debounce can be performed, i.e. the change is only recorded after the debounce time.

[0053] The contribution of the turn signal is a yes / no decision and can be set to an influence of, for example, 0.3. Optionally, debounce can be performed, for example with a debounce time of 2 seconds.

[0054] A scale and optional influence can be set for the contribution of the transverse velocity. Figure 3 is a graph (0-1 function) showing values ​​as a function of lateral velocity. In this example, starting at a lateral velocity of 0.5 m / s, the lateral velocity contributes increasingly to the steering sum (scaling). For example, this value can be increased to 1.5 m / s and then remain constant. The maximum value is 1, and the influence can be set. For example, if the influence is 0.3, the maximum contribution of the lateral velocity to the steering sum is 1·0.3 = 0.3.

[0055] The relative speed is the relative speed between the vehicle and the vehicle ahead. A suitable influence can be set for the relative speed contribution. For example, the relative speed can be a filtered relative speed with time constants for uphill and downhill slopes.

[0056] The object X speed (the longitudinal speed of the vehicle ahead v x )'s contribution sets the scale and optional impact. Figure 4 is a graph (1-0 function) showing the contribution as a function of the longitudinal velocity v x In this example, the object's X velocity contributes most to the steering sum at low speeds and decreases (scaled) at high speeds. The maximum value is 1, and the influence can be set. For example, if the influence is 0.4, the object's X velocity contributes at most 1·0.4=0.4 to the steering sum.

[0057] Calculates the heading or heading angle of the driving path. The driving path moves roughly within the area of ​​a circular path and is defined by the curvature value. Figure 5 It is expressed in the following relationship:

[0058]

[0059] μ is the heading, is the heading travel path, y is the lateral distance of the object, and R is the radius (=1 / curvature).

[0060] For larger heading angles, the contribution to the steering sum increases, e.g. Figure 6 The maximum value is 1, and the influence can be set. For example, if the influence is 0.4, the maximum contribution of the heading to the steering sum is 1 0.4 = 0.4.

[0061] Steering scale

[0062] In addition to the turn probability, a turn scale multiplied by the turn probability can also be used to enable turn recognition. The product of the turn probability and the turn scale can be a maximum of 1. Specifically, the turn probability and the turn scale can be specified so that their product can be a maximum of 1, which corresponds to enabling turn recognition. Specifically, the greater the product of the turn probability and the turn scale, the greater the probability that the preceding vehicle is turning.

[0063] Specifically, the turn scale depicts the impact of lateral position and lateral velocity relative to the travel path on turn recognition. Lateral velocity can be factored into turn recognition, allowing it to be validated only when the potential turn is at a higher lateral velocity. This reduces errors in turn recognition.

[0064] The turn scale may include two or more factors that influence or indicate turn identification. In an exemplary embodiment, the two or more factors may be selected from the group consisting of: turn distance, turn curvature, longitudinal turn speed, lateral turn speed, distance to roundabout, distance to T-junction, lateral turn distance, and lateral turn acceleration.

[0065] The turning distance is used to limit the turn recognition to a maximum valid range based on the simplified driving path calculation (1-0 function). In particular, large object distances can be removed, such as Figure 7 As shown in .

[0066] The steering curvature prevents erroneous triggering of the steering recognition by limiting the curvature value of the driving path. Figure 8 . Optionally, debounce can be set for certain curvature values, such as > 0.009. Turn curvature is particularly useful for certain curve scenarios where turns should not be recognized, such as on roads with frequent changes in curvature, when following a leading vehicle on a curve, and when the host vehicle cuts in from behind a leading vehicle.

[0067] The longitudinal steering speed ensures that the steering recognition is only activated when the vehicle ahead is at a low longitudinal speed. This can be implemented using a 1-0 function, such as Figure 9 It is shown as an example in FIG.

[0068] A lateral steering speed is used in order to activate the steering recognition only when the leading vehicle is at a high lateral speed. As a result, a reduction in false positives in steering recognition can be achieved. The lateral steering speed can be implemented with a 0-1 function, as shown in Figure 10 .

[0069] The lateral steering speed is particularly advantageous for certain scenarios where steering should not be recognized, for example when driving past (i.e. the ego vehicle and the leading vehicle are on different lanes) and / or when the leading vehicle is slowly turning off.

[0070] The distance from a roundabout avoids false positives in front of and in the roundabout. Similarly, the distance from a T-junction avoids false positives related to T-junctions. Both can be implemented with a 0-1 function, as shown in Figure 11 .

[0071] The lateral steering distance from the driving path can be expressed as a 0-1 function, as shown in Figure 12 . With reference to Figure 5 , the lateral steering distance can be defined as follows:

[0072] Lateral distance = A - R,

[0073] wherein A is the maximum distance from the driving path and R is the radius of the driving path.

[0074] The lateral steering acceleration ensures that steering is only actively allowed from a certain lateral acceleration. The lateral steering acceleration can be shown as a 0-1 function, as shown in Figure 13 .

[0075] The road is clear

[0076] Road clearness can be implemented by looking at the road profile. By looking at the road profile in front of the ego vehicle, false positives will be prevented from triggering steering recognition.

[0077] To this end, the road profile can be divided into several segments in which there are different radii (curvatures). Depending on the situation, different road radii can be allowed in order to allow steering recognition.

[0078] With reference to Figure 14 , three segments X i , X i+1 , X i+2 , the ego vehicle E and the leading vehicle ZO are shown.

[0079] If the radius 1 is smaller than a predetermined value and if there is a potential steering in segment 1, no steering recognition will take place. At a radius 2 smaller than a predetermined value and a potential steering close to x i+1No turn recognition is performed in case the road segment distance 1 is small and the radius difference is larger than a predetermined value. No turn recognition is performed in case the potential turn is in road segment 2 and the radius 2 is smaller than a predetermined value. No turn recognition is performed in case the radius 3 is smaller than a predetermined value and the sum of road segment distance 1 and road segment distance 2 is smaller than a predetermined value. No turn recognition is performed in case the potential turn is in road segment 3 and the radius 3 is smaller than a predetermined value.

[0080] Other possible conditions for turn identification

[0081] One or more of the following aspects can optionally be employed in turn recognition.

[0082] • Initial record delay (cut-in): The turn object should be the target object at least a certain time, e.g. 3 seconds, before recognition.

[0083] • Curvature de-bounce: A certain time, e.g. 4 seconds, from a certain curvature value. This avoids false triggers in a curve when the driving path passes the object.

[0084] • Exclusion of certain object types: e.g. trucks, buses, bicycles, pedestrians, motorcycles. Other object types cannot be excluded here.

[0085] • Hysteresis: This value remains at 1 until it drops below a certain value, e.g. 0.7.

[0086] According to the invention, it is recognized whether a vehicle driving ahead is turning. This is done by means of a turning probability. In determining the turning probability or recognizing the turning object, a number of criteria can be used which are designed to increase or optimize the probability that the vehicle driving ahead is actually turning.

[0087] Although the present application has been illustrated and described in more detail by means of preferred embodiments, the application is not restricted by the disclosed examples but can be implemented in other ways. It is therefore clear that many variations are possible. It is also clear that the exemplary embodiments presented are by no means exhaustive, and that these examples should not be understood as a limitation on the scope of the application, the applicability or configuration thereof. Rather, the preceding description and the drawings are to enable the person skilled in the art to carry out the exemplary embodiments, wherein the person skilled in the art, upon perceiving the disclosed inventive concept, can make various changes, for example with respect to the function or arrangement of the individual components mentioned in the exemplary embodiments, without departing from the scope of protection as defined by the claims and their legal equivalents, for example as further explained in the description.

Claims

1. A method (100) for identifying a turning object, comprising: Detecting (110) a vehicle traveling ahead through an environment sensing mechanism of the own vehicle; determining (120) a turn probability of the preceding vehicle based on at least one first criterion, wherein the at least one first criterion is selected from the group consisting of: turning out, turn signal, lateral speed, relative speed, object speed, and heading; multiplying the turn probability by a turn scale based on at least one second criterion to scale the turn probability, wherein the at least one second criterion is selected from the group consisting of: turn distance, turn curvature, longitudinal turn speed, lateral turn speed, distance to a roundabout, distance to a T-intersection, lateral turn distance, and lateral turn acceleration; as well as When the scaled turning probability is equal to or greater than a predetermined threshold, the leading vehicle is marked (130) as a turning object.

2. The method (100) according to claim 1, further comprising: An impact value is applied to at least one of the at least one first criterion and the at least one second criterion.

3. The method (100) according to claim 1 or 2, further comprising: At least one of the at least one first criterion and the at least one second criterion is de-dithered.

4. The method (100) according to claim 1 or 2, wherein: The preceding vehicle is marked as a turning object only if turn detection is permitted based on the road course.

5. A storage medium comprising a software program designed to be executed on one or more processors and thereby to perform the method (100) according to any one of claims 1 to 4.

6. A system (200) for identifying a turning object, comprising: A detection module (210) is designed to detect a vehicle traveling ahead through an environment sensing mechanism of the vehicle itself; and At least one processor unit (220) designed to: determining a turn probability of the preceding vehicle based on at least one first criterion, wherein the at least one first criterion is selected from the group consisting of: a turn-out, a turn signal, a lateral speed, a relative speed, an object speed, and a heading; multiplying the turn probability by a turn scale based on at least one second criterion to scale the turn probability, wherein the at least one second criterion is selected from the group consisting of: turn distance, turn curvature, longitudinal turn speed, lateral turn speed, distance to a roundabout, distance to a T-intersection, lateral turn distance, and lateral turn acceleration; as well as When the scaled turning probability is equal to or greater than a predetermined threshold, the leading vehicle is marked as a turning object.

7. A vehicle comprising the system (200) according to claim 6.

8. The vehicle of claim 7, wherein the vehicle is a motor vehicle.

Images