Methods for determining the degree of automation of moving objects, methods for controlling an assisted or automated vehicle and vehicle

Abstract

The invention relates to a method for detecting the degree of automation of moving objects (PMOs) in the vicinity of an ego vehicle (1), wherein the moving object (PMO) is detected by means of at least one environmental sensor (2) of the ego vehicle (1), wherein several successive positions of the moving object (PMO) are detected and stored as a point cloud in a coordinate system, after which a motion path of the moving object (PMO) is determined from the point cloud and a dispersion value of the motion path is calculated, after which the calculated dispersion value is compared with a predetermined threshold value, wherein in the case that the dispersion value lies within a predetermined range below the threshold value, it is assumed that the moving object (PMO) is driving automatically, and after which this information is output.An inventive method for controlling assisted or automated driving of an ego-vehicle (1) uses this method for detecting the degree of automation in order to output the determined information to a driving control unit of the ego-vehicle (1). The invention also relates to a vehicle (1) with a sensor system which is configured to carry out the method(s).

Description

[0001] The invention relates to a method for detecting the degree of automation of moving objects in the environment of an ego vehicle. Furthermore, the invention relates to a method for controlling assisted or automated driving of an ego vehicle, in which such a method for detecting the degree of automation of moving objects is implemented. The invention also relates to a vehicle with a sensor system.

[0002] Vehicles with sensor systems to enable assisted or automated driving are generally known from the prior art. Such sensor systems typically comprise at least one sensor, generally several sensors, whose data is combined. The sensors can include cameras or stereo cameras, as well as radar sensors, LiDAR sensors, or similar. Particularly preferred is the data from different types of sensors, for example, camera images and sensor data from a LiDAR sensor, to be combined to form a common "image" of the surroundings. This is then referred to as sensor fusion. Vehicles with such sensors are now capable of assisting a person driving the vehicle by taking over certain driving tasks or, in the event of safety-critical situations, even intervening in the driving behavior.In addition, a partially automated or automated journey of a vehicle with such sensors is also possible, i.e. a journey in which the vehicle drives more or less on its own and only needs to be monitored to a greater or lesser extent by the person driving the vehicle.

[0003] In principle, it is now of interest both for a person driving a vehicle independently and for partially or fully automated vehicles to know whether moving objects, especially vehicles, in the surrounding area are also driving in an automated or non-automated manner. In many cases, this simplifies the assessment of potentially upcoming driving maneuvers and can thus enable more efficient and safer driving, for example, by changing safety distances, adjusting speeds, or similar measures.

[0004] The current state of the art only includes vehicles that communicate externally whether they are driving in an automated or non-automated manner, for example through corresponding light signals that inform other road users. An example in this context is DE 10 2016 113 913 A1, which shows such a device under the title "communication device." Light signals that can be recognized from outside the vehicle are preferably used for this purpose. In principle, these signals are of course not only recognizable to a person driving a non-automated vehicle or to other road users such as pedestrians or cyclists, but also to the vehicle's environmental sensors. The disadvantage, however, is that the degree of automation can only be recognized by vehicles that communicate this to the outside via display elements.However, it would be desirable to know the level of automation of all vehicles in the vicinity of an ego vehicle without having to explicitly communicate with them, and in particular if all or some of the vehicles in the vicinity do not have a communication device within the meaning of the above-mentioned DE document.

[0005] DE 10 2014 202 453 A1 discloses a method in which a movement trajectory is determined to identify a vehicle's operating mode. Depending on the determined movement trajectory, an autonomy value is determined that is representative of whether the vehicle is operating autonomously or non-autonomously.

[0006] DE 10 2016 205 569 A1 discloses a method for a vehicle with which a determination is made as to whether another vehicle located in the vicinity of the vehicle is operating at least partially automatically. For this purpose, the other vehicle is detected using an environmental sensor system, and a driving trajectory is derived from its position. The detected driving trajectory is analyzed for a course and patterns in driving behavior. If constant areas are detected in speed or acceleration profiles along the driving trajectory, it is concluded that the other vehicle is operating in an automated manner.

[0007] The object of the present invention is therefore to provide a further method for detecting the degree of automation of moving objects in general. A further object may then be to provide a method for controlling assisted or automated driving based on this detected data.

[0008] According to the invention, this object is achieved by a method for detecting the degree of automation of moving objects with the features of claim 1. Advantageous embodiments and further developments emerge from the dependent subclaims. The object is also achieved by a method for controlling assisted or automated driving of an ego vehicle, in which the method for detecting the degree of automation is carried out, wherein the determined information is further used according to the features of claim 8. Advantageous further developments of this method for controlling assisted or automated driving also emerge from the dependent subclaims. The object is also achieved by a vehicle according to claim 11, which has a sensor system and at least one computing unit, so that it is suitable for carrying out the method for detecting the degree of automation.In the case of this vehicle according to the invention, an advantageous further development also results from the dependent claim.

[0009] The method is used to detect the degree of automation of moving objects, also referred to in this context as potentially moving objects or, in English, as Potential Moving Objects (PMO). The method according to the invention uses at least one environmental sensor of an ego vehicle. For this purpose, one or more cameras, but also radar sensors, LiDAR sensors, and the like, can be used in particular and, if necessary, merged after data has been determined with regard to their detected environmental data. The detected moving object is then tracked via the environmental sensors so that several consecutive positions of this moving object can be detected and stored as a point cloud in a coordinate system. This can be done for each detected object in the environment of the ego vehicle.From the point cloud, a movement path of the respective moving object is then determined, i.e. its movement trajectory.

[0010] The method is based on the realization that the movement trajectory of an automated vehicle exhibits a much smaller variance than that of a vehicle driven manually by a person. Accordingly, after determining the movement path, a variance value for this movement path is calculated. The calculated variance value is then compared with a predefined threshold value. If the variance value lies within a predetermined range on this side of the threshold value, an automated moving object is assumed; otherwise, it is not. This information can then be output.

[0011] Whether the scatter value is greater or smaller than the corresponding threshold depends on the type of calculation, so that, depending on this, a scatter value can, in principle, be either larger or smaller to indicate automated driving. Since the type of calculation is known in the method, it can then be decided whether a less scattered trajectory is present in one or the other case.

[0012] The information thus obtained can then be used within the ego vehicle.

[0013] The generation of information as to whether a moving object, in particular a vehicle, is being driven in an automated or non-automated manner can be achieved using the method according to the invention without the involvement of the moving object. No communication connection to the moving object whose degree of automation is to be detected is necessary. Furthermore, no outward-facing displays or indications from this moving object are necessary. Since the environmental sensors of the ego vehicle operate reliably under various conditions, detecting the degree of automation is also largely independent of the prevailing lighting situation, weather conditions, or the like.

[0014] The existing conflict of differentiating between assisted and unassisted driving can thus be resolved very simply and efficiently based solely on data that the ego vehicle itself can easily capture in the majority of environmental scenarios that occur. Since only the trajectory of the moving object needs to be captured, it is sufficient if a single point on this object can be captured and tracked over a certain period of time or distance, for example, a headlight, a taillight, or the like. This facilitates the use of the method even in poor visibility conditions, such as at night, in rain, or similar situations.

[0015] Since the prediction of potential movement trajectories of moving objects is already carried out today when programming control systems for assisted or automated driving on the basis of an existing section of a previous trajectory, with the acquisition of this section of the existing trajectory, all data is already available in many control systems, so that the method according to the invention can be implemented extremely simply and efficiently even with existing hardware.

[0016] Another key advantage of the method is that the inventive approach enables a significant increase in the added value of the so-called navigation display. This display combines the display of navigation data and assistance systems to generate added value, particularly for a vehicle that is at least partially manually driven. All of this ultimately contributes to driving safety, as it allows for differentiation between the moving objects in the surrounding area with regard to automated and non-automated driving, which enables improved reactions to these moving objects both in manual driving mode and in assisted or automated driving mode.

[0017] According to the invention, the calculation of the scatter value of the movement path of detected moving objects is suspended as long as disruptive environmental influences have been detected by sensors of the ego vehicle. Such disruptive environmental influences can, in particular, be influences that cause the movement path to assume a correspondingly high scatter even in the case of automated driving. These disruptive environmental influences could, for example, be poor road conditions, ruts, or the like. In particular, however, a crosswind, which has a strong influence on the driving trajectory or its scatter value, can also be detected as a disruptive environmental influence.

[0018] A very advantageous embodiment of the method according to the invention can provide for the information to be output to a control system for the ego vehicle and / or visualized for a person using the ego vehicle. The information can therefore be displayed, for example, to a person who is manually driving the ego vehicle. Alternatively, or preferably in addition to this, if a control system for assisted or automated driving is present in the ego vehicle, the information can also be output to this control system. Here, the information can also be used for various purposes, as will be explained in more detail later with reference to the method according to the invention for controlling assisted or automated driving of an ego vehicle.

[0019] According to a very advantageous refinement of the method, the motion path or trajectory can be determined using a polynomial. Preferably, the polynomial can be determined from the existing point cloud using so-called principle polynomial analysis.

[0020] Another very advantageous embodiment of the method according to the invention can also provide for the predetermined threshold value to be determined from values ​​recorded by vehicles driving in automated or assisted mode. The threshold value can thus be determined, for example, using corresponding experimental data recorded by vehicles known to have been driven in automated or assisted mode.

[0021] Alternatively or in addition, the threshold can also be determined using a scatter value from a movement path determined during manual driving. For this purpose, the reverse approach is essentially chosen, and the recorded data thus represents the scatter that reliably indicates a manually moved object. In this case, a safety margin can preferably be added to the raw value generated by manual driving in order to achieve a threshold that is a safe distance from the scatter of the movement path during manual driving, so that values ​​on this side of the threshold indicate an automated vehicle with a very high degree of certainty.

[0022] According to a highly advantageous refinement of this embodiment of the method according to the invention, the disruptive environmental influences to which the method is exposed can now be detected based on steering corrections of the ego vehicle. If steering corrections are continually necessary in the ego vehicle, for example due to crosswind influences and / or ruts, this can be detected accordingly via the control system of the ego vehicle. This is then an indication that moving objects, especially moving objects traveling in the same direction and / or on the same road, could be subject to similar disruptive environmental influences.Potentially, their movement paths are also subject to greater scatter, so that the method according to the invention can no longer reliably determine whether the scatter value lies within a range in which automated driving can be reliably distinguished from non-automated driving. Accordingly, the method can then be temporarily suspended until the disruptive environmental influences no longer occur.

[0023] The method according to the invention for controlling assisted or automated driving of an ego vehicle uses precisely this method according to one of the exemplary embodiments described above to obtain information about the degree of automation of moving objects. This information is then always output to a driving control system of the ego vehicle. As already explained above, an additional output can be provided to a person using the vehicle, for example, a graphical output on a central display.

[0024] In a very advantageous refinement of this method, the control method according to the invention then uses this information to adapt its own driving behavior depending on the degree of automation of moving objects. If, for example, the detected moving object is not driven automatically or manually according to the generated information, a greater safety distance can be maintained, for example.

[0025] A further very advantageous embodiment of the control method further provides that the driving control of the ego vehicle requests a data connection to the moving objects driving as automated and, once this data connection has been established, activates cooperative driving. In the event that the ego vehicle itself is moved automatically or at least with assistance, a vehicle-to-vehicle connection can now be requested and established in order to initiate cooperative driving. Such cooperative driving can, for example, include an adaptation of the driving behavior according to the embodiment described above, which adapts the own driving behavior based on planned driving trajectories that the moving object communicates via the data connection.Furthermore, an exchange of sensor data is possible to realize expanded environmental monitoring, for example, by utilizing the additional range of the sensors of another vehicle driving in front of the ego vehicle to enable a greater sensor range in the direction of travel. The activation of so-called platooning operation can also be implemented here. When several vehicles are traveling in convoy communicating with each other via the data connection, a very close distance is maintained between the vehicles, with only the front vehicle being driven manually or automatically, while the other vehicles in the convoy follow at a short distance in order to save space and energy.

[0026] Other measures for cooperative driving, which are generally known from the field of vehicle-to-vehicle connections, are of course also conceivable.

[0027] The vehicle according to the invention comprises a sensor system that includes at least one camera and is configured to recognize and detect moving objects in the environment as such. It also comprises a computing unit configured to carry out the above-described method for detecting the degree of automation of moving objects in one or all of the described embodiments. Such a vehicle can then generate reliable information about the degree of automation of moving objects in its environment without having to interact with them in any way or requiring data provided by them regarding their degree of automation.A particularly advantageous embodiment of the vehicle according to the invention can also provide for it to include a driving control system, wherein the existing computing unit or another computing unit within the vehicle is configured to carry out the described method for controlling assisted or automated driving. Ideally, the vehicle therefore has the capability of both carrying out the method for detecting the degree of automation of moving objects in its environment and using this information to adapt its own driving accordingly via a driving control system.

[0028] Further advantageous embodiments of the two methods according to the invention and of the vehicle also emerge from the exemplary embodiment, which is described in detail below with reference to the figures.

[0029] Showing: Fig. 1 a schematically indicated ego vehicle with a camera as part of a sensor system; Fig. 2 a flowchart for explaining the method according to the invention; Fig. 3 a schematic representation of detected potentially moving objects; Fig. 4 a schematic representation of the recorded values ​​of a potentially moving object; Fig. 5 a point cloud generated from this of the positions of this potentially moving object; and Fig. 6 an exemplary representation of the calculation of the scatter based on the extracted polynomial.

[0030] In the presentation of the Fig. Figure 1 shows a highly schematic representation of a vehicle, which will be referred to below as ego-vehicle 1, and which has a sensor system by means of which it records the environment in order to detect potentially moving objects, so-called PMOs. As an example of the sensor system, a camera 2 is schematically indicated within the ego-vehicle 1. This camera is connected to at least one computing unit 3, wherein this computing unit 3 implements the subsequent process sequences via a suitable program code. In addition to the computing unit 3, a communication connection (not shown here) can also be established to a server external to the vehicle or a cloud, which can take over parts of the necessary computing operations, so that computing operations can be outsourced from the ego-vehicle 1 to external sources.

[0031] However, all this plays a subordinate role in its exact structure for the present invention, so that now based on the process flow of the Fig. 2, the method for sensor-based detection of the degree of automation of PMOs will be discussed. The flow chart comprises seven individual process steps S1 to S7, which will be discussed in detail. Process step S1 comprises the determination of the object position of various PMOs. In the schematic representation of the Fig. 3, the position of the ego vehicle 1 is shown as an example via a cross. In addition, in the exemplary embodiment shown here, the object positions of three potentially moving objects, typically other road users, in particular vehicles, can be recognized via three points, which are designated PMO1, PMO2 and PMO3, respectively. As indicated by the arrow labeled 4 in the flow chart on the left, the system then jumps to step S2, in which a new position of the same object is detected. As indicated by the arrow 5 on the right side of the flow chart, these steps S1 and S2 are repeated for a defined route section or a predetermined time horizon, for example a route section of 100 meters or a time horizon of 5 seconds. In the representation of the Fig. 4 this is indicated accordingly for the object PMO1, so that several consecutive points are recorded, each of which shows the position of the object PMO1 at a different time.

[0032] In step S3, the point cloud resulting from steps S1 and S2 is modeled for the respective PMO. This modeling of the accumulated point cloud can preferably be performed using the so-called Principal Polynomial Analysis. This point cloud in a coordinate system that plots the x-position of the PMO over its z-position is shown in the Fig. 5 recognizable.

[0033] In step S4, a scatter value is calculated. For this purpose, an adjustment, a so-called fit, is performed on the basis of the point cloud. The polynomial generated by the fitting is shown in Fig. 5 and marked with the reference symbol P.

[0034] It is assumed that in an assisted scenario, the so-called planner, i.e. the control software of the PMO, calculates its trajectory or movement path based on sensor data and / or high-resolution map data. The determined trajectory will therefore follow a fixed course. A scatter or variance calculated based on the polynomial should then be very small. In this first case, it can therefore be assumed that the PMO is driving automatically. In the other case, a manual drive results in significantly greater scatter in the movement path, caused, for example, by distractions of the person driving the vehicle, external adversities such as wind, uneven road surfaces, an erratic driving style, fatigue, etc.

[0035] In the presentation of the Fig. 6, the polynomial P is indicated by the calculated scatter of the data points along the polynomial. The polynomial P from step S3 is shown centrally with a thick line and is analogous to the representation in Fig. 5 with the reference symbol P. The individual data points along the polynomial P are marked with a dot for the example of the automated procedure over several circles, while the data points of a non-automated circle are marked with a cross. In the case of a polynomial P recorded in the real application of the procedure, unlike in the representation of the Fig.6 of these data points, only one or the other is present. From these existing data points, the orthogonal distance of each data point to the polynomial P is determined, and these distances are summed. The calculated sum is then divided by the number of data points to obtain a dispersion value. On average, the distances of the data points in the case of non-automated driving deviate significantly more from the polynomial P than those of automated driving for the route section shown here.

[0036] In step S5, this dispersion value is then compared with a predefined threshold value, and on this basis, the degree of automation of the detected PMO is determined in step S6. This determination is typically binary, so that the only information generated is whether the vehicle is driving automatically or not. In step S7, this information is then forwarded, for example, to a display device for visualization and / or to a driving control system of the ego vehicle 1, which reacts accordingly to the fact that the detected PMO is driving automatically or not, for example, by adjusting safety distances and / or attempting to contact the detected PMO in vehicle-to-vehicle communication, or the like. All of this has already been explained in detail above.

[0037] The decision in step S6 can now be made, as indicated by the arrow labeled 6 on the right in the flow chart, after multiple execution of the method steps S1 to S5, in order to verify the statement by an evaluation over several route sections or time periods and thus make the result more reliable.

Claims

[1] Method for detecting the degree of automation of moving objects (PMO) in the environment of an ego vehicle (1), for which purpose the moving object (PMO) is detected by means of at least one environmental sensor (2) of the ego vehicle (1), wherein several successive positions of the moving object (PMO) are detected and stored as a point cloud in a coordinate system, after which a movement path of the moving object (PMO) is determined from the point cloud and a scatter value of the movement path is calculated, after which the calculated scatter value is compared with a predetermined threshold value, and wherein, in the event that the scatter value lies in a predetermined range on this side of the threshold value, an automated moving object (PMO) is assumed, and after which this information is output, characterized bythat the calculation of the scatter value of the movement path is suspended as long as disturbing environmental influences are detected by sensors (2) of the ego vehicle (1). [2] Method according to claim 1, characterized by that the information is output to a control system for the ego vehicle (1) and / or visualized for a person using the ego vehicle (1). [3] Method according to claim 1 or 2, characterized by that the motion path is determined using a polynomial (P). [4] Method according to claim 3, characterized by that the polynomial (P) is determined using principal polynomial analysis. [5] Method according to one of claims 1 to 4, characterized by that the specified threshold value is determined from recorded values ​​of automated or assisted vehicles. [6] Method according to one of claims 1 to 5, characterized bythat the predetermined threshold value is determined by means of a scatter value from a movement path determined during manual driving. [7] Method according to claim 6, characterized by that steering corrections of the ego vehicle (1) are evaluated to detect disturbing environmental influences. [8] Method for controlling assisted or automated driving of an ego vehicle (1), in which a method according to one of claims 1 to 7 is carried out, wherein the determined information on the degree of automation of moving objects (PMO) in the environment of the ego vehicle (1) is output at least to a driving control of the ego vehicle (1). [9] Method according to claim 8, characterized by that the driving control of the ego vehicle (1) adapts the driving behavior depending on the degree of automation of moving objects (PMO) output to it. [10] Method according to claim 9, characterized bythat the driving control of the ego vehicle (1) requests a data connection to the automated moving objects (PMO) and activates a cooperative driving operation after the setup has been completed. [11] Vehicle (1) with a sensor system which comprises at least one camera and which is designed to recognise and detect moving objects (PMO) in the environment as such, and with a computing unit (3) which is designed to carry out the method according to one of claims 1 to 7. [12] Vehicle (1) according to claim 11, characterized by a driving control, wherein the computing unit (3) or a further computing unit (3) is configured to carry out the method according to one of claims 9 to 10.

Images