Method for providing an adjustment notification to a self-traffic participant to be notified, method for ascertaining a group-estimation variable of a movement path, and method for ascertaining global group movement data

Abstract

A method for providing an adjustment notification, preferably a warning notification, to at least one self-notified traffic participant (10) via a warning device (12) for adjusting at least one driving function to adapt to at least one surrounding traffic participant (32) located in and / or moving in the surrounding area (2) of the self-notified traffic participant, and for predicting movement paths (P) for both the self-notified traffic participant and the at least one surrounding traffic participant (32). CL P R The warning device (12) provides and / or is provided to the warning device, along the predicted movement path of the traffic participant (10, 32), wherein the warning device (12) is based on the predicted movement path (P) CL P R The warning device (12) identifies at least one notification parameter characterizing the provision of the adjustment notification, preferably characterizing whether an adjustment notification (I1, A32) should be provided and / or at what time the adjustment notification (I1, A32) should be provided. According to the invention, the warning device (12) identifies the notification parameter based on at least one population prediction parameter, particularly statistical, used to determine the predicted motion path (P). CL P R The estimated quality of at least one of the following, wherein the at least one group estimated parameter is determined based on global group motion data of multiple traffic participants, which is generated in multiple traffic areas (2) that are different from each other.

Description

Technical Field

[0001] This invention relates to a method for providing adjustment notifications to at least one self-driving traffic participant (Ego-Verkehrsteilnehmer, sometimes referred to as a vehicular traffic participant), a method for determining at least one group prediction parameter for the estimated movement path, and a method for determining global group movement data. The invention further relates to a warning device for providing adjustment notifications to self-driving traffic participants, a group data determination device for traffic participants, a map generation device, and a machine-readable, and especially computer-implementable, motion model for predicting the movement paths of traffic participants. Background Technology

[0002] As is known from the prior art, vehicles equipped with communication devices can conduct local communication with other vehicles in the surrounding environment.

[0003] Therefore, the preferred technology used is known as so-called vehicle-to-vehicle communication (also abbreviated as C2C communication, which is German for “Auto-zu-Auto”-Kommunikation), or as V2V communication (an abbreviation of the Anglo-Saxon term Vehicle-to-Vehicle-Communication, German for “Fahrzeug-zu-Fahrzeug”-Kommunikation).

[0004] V2V communication is a special case of V2X communication (Vehicle-to-X-Kommunikation, i.e., "Vehicle-to-everything"-Kommunikation), which is also known under the concept of Car2x communication (for "Car to x"). This is generally understood as communication between a vehicle and its surrounding environment, i.e., neighboring vehicles or infrastructure devices.

[0005] Currently, vehicle-to-vehicle communication is used, for example, to warn each other of dangerous situations where, without warning, drivers might not react at all or react very late. For instance, this could warn a driver of approaching a congested area. Furthermore, this communication technology is advantageously used to warn drivers of emergency vehicles, such as ambulances (which often travel at higher speeds from behind).

[0006] Besides congestion, road intersections also present an increased risk of collision with other road users compared to lanes without intersections. The goal is to reduce accident risk in such traffic situations by warning drivers as early as possible. However, it is crucial to avoid issuing false warnings about collisions to drivers of other vehicles when no danger is present without prior warning. If, for example, another road user approaching the intersection arrives at the intersection with sufficient time (the other user's path intersects with the vehicle's at the intersection), allowing both users to cross without interference, then the driver of the other vehicle should ideally not be unnecessarily warned of a potential collision. This is a complex issue because such warnings depend on the movement of multiple road users, requiring anticipation of their movements.

[0007] Because the warning function must utilize a corresponding lead time to determine whether the situation has become critical, it is necessary to estimate the number of traffic participants involved within that timeframe. However, the longer the estimated timeframe, the lower the confidence level of the estimate. The goal in developing this function is to provide the probability of an actual encounter and / or the determined distance when passing the intersection of two traffic participants. A warning should only be displayed when the probability of a critical encounter at the intersection exceeds a specified value. Here, the more reliable and less erroneous the estimate, the earlier the warning can be generated, given the same false trigger ("FP") rate.

[0008] However, the precise parameterization of the function using the target probability depends on the function's design and must take customer preferences into account. The uncertainty of the forecast itself arises from the movement behavior of traffic participants at the intersection (especially through deceleration and acceleration processes) and depends on various factors, such as the intersection's structure, the presence of other traffic participants on site, and cultural differences.

[0009] According to CN 116129650 B, a traffic advance warning system is known. This system calculates the initial traffic flow of a road segment, the vehicle quantity change conditions at road intersections, and the vehicle change rate at the corresponding road intersections. Based on the vehicle change rate at the corresponding road intersections and the road capacity of each road segment, overloaded road segments and congestion times are determined for each road segment. Furthermore, based on traffic congestion information and the travel time required for a target vehicle to pass a target traffic light, advance warnings are issued to the target vehicle.

[0010] While the traffic advance warning system provides drivers with the possibility of avoiding overloaded road sections warned of by it, and can reduce the risk of accidents, it does not provide the feasibility of reducing the risk of collisions, for example, between vehicles whose paths meet at road intersections.

[0011] A method for predicting pedestrian crossing behavior at intersections is known from WO 2022 / 110611 A1. To this end, a neural network-based prediction model is proposed, trained using deep reinforcement learning. Here, the discrete behavioral states of pedestrians crossing the road, such as walking, running, and stopping, are predicted, and early warnings of danger are given to both crossing pedestrians and passing vehicles. As a reward, a modified time to collision (MTTC) is used, which considers the relative position, relative velocity, and relative acceleration between the vehicle and the pedestrian. This considers whether, when a vehicle encounters a crossing pedestrian at the intersection, it should brake to decelerate or accelerate to pass by.

[0012] Disadvantageously, this method can only be used at a selected, specific pedestrian-road intersection where corresponding sensors are installed at the road edge for this purpose, and the training method is performed for that particular pedestrian-road intersection. For example, if it is necessary to predict pedestrian crossing behavior in other pedestrian-road intersection areas, then correspondingly costly structures and new training methods are required. Summary of the Invention

[0013] The purpose of this invention is to overcome the disadvantages known from the prior art and to provide a method or device that enables warning traffic participants as early as possible while minimizing the need to warn them at intersections where there is no imminent danger of encounter.

[0014] A further object of the present invention is to provide a method or apparatus by means of which, by means of said method or apparatus, one can make the most accurate and effective conclusion possible regarding possible adjustments to the movement behavior of at least one of the traffic participants involved, by anticipating the movement processes of two traffic participants whose current movement behavior may cause potential dangerous traffic situations, in relation to their potential mutual influence in terms of their movement behavior. In particular, it should be possible to provide such conclusions as early as possible.

[0015] According to the invention, this objective is achieved by the subject matter of the independent claims. Advantageous embodiments and improvements of the invention are the subject matter of the dependent claims.

[0016] A method according to the invention (preferably computer-implemented) for providing adjustment notification (especially for output and / or transmission and / or provision) to at least one traffic participant (especially one's own vehicle) to be notified via a warning device, for adjusting at least one driving function and / or vehicle function to accommodate at least one surrounding traffic participant located in and / or moving within the surrounding area of ​​the traffic participant, wherein the warning device provides and / or will provide (respectively) predicted movement paths along which the movement priorities of the traffic participants are expected for both the traffic participant and, preferably, at least one surrounding traffic participant. The method may be computer-implemented. However, it is also contemplated that the method may partially include computer-implemented method steps, and further, for example, may include outputting the adjustment notification to an individual (e.g., a user / driver) by means of an output device.

[0017] In other words, the warning device provides and / or will provide a predicted movement path for its own traffic participants, along which their movement is expected. Preferably, the warning device also provides and / or will provide a predicted movement path for surrounding traffic participants, along which their movement is expected. The warning device provides an adjustment notification, which should at least be transmitted to and / or output to and / or provided to its own traffic participants.

[0018] For example, driving functions and / or vehicle functions can be understood as V2X perception functions and / or (V2X) safety functions.

[0019] "Adjusting at least one driving function" can also be understood as adjusting (the pedestrian's) movement, such as changing the speed of movement.

[0020] "Driving function" is particularly understood as a function that influences longitudinal and / or lateral movement, and preferably as longitudinal and / or lateral guidance (of traffic participants, such as vehicles), such as acceleration, braking, avoidance, and / or changing of driving route, and / or aborting and / or accelerating overtaking operations.

[0021] "Driving functions" can also be understood in particular as vehicle functions to be performed and / or performed during the movement of traffic participants, such as (acoustic and / or visual) warning functions to traffic participants in the surrounding area (e.g., operating the horn and / or flashing lights), or lighting functions for traffic participants, such as (road) lighting (while adjusting the lighting or illumination to suit surrounding traffic participants, e.g., to prevent glare).

[0022] Preferably, the "driving function" is (preferably V2X-based) an intersection assist function that warns vulnerable road users (VRUs), especially cyclists, on the collision path. Preferably, the driving function is a driver assistance function or mobility assistance function and / or a safety function of the vehicle and / or mobility assistance device.

[0023] Here, the adjustment of at least one driving function (and / or vehicle function) is preferably performed at the location of the traffic participant and / or surrounding traffic participants. It is also conceivable that by adjusting the at least one driving function and / or vehicle function, the relative motion behavior between the traffic participant and surrounding traffic participants is altered and / or adjusted.

[0024] Preferably, only the driving functions of the (to be notified) traffic participants themselves are adjusted.

[0025] Driving functions and / or vehicle functions can be driving functions or vehicle functions that can be performed at least partially automatically and / or fully automatically, especially when they are performed without any (manual) intervention by the driver or user (the driver or user) of other road users (the driver or user) without the need for intervention by other road users (the driver or user) of the vehicle or vehicle.

[0026] However, it is also conceivable that the adjustment of driving functions and / or vehicle functions may be (manually) made and / or triggered by the user or driver through intervention in vehicle guidance, or may have to be (manually) made and / or triggered by the user or driver through intervention in vehicle guidance.

[0027] "Providing adjustment notification to at least one traffic participant to be notified" can be understood as providing the adjustment notification (preferably via a warning device) to and / or transmitting it to the control device of the traffic participant or vehicle, so as to at least partially automatically, and preferably fully automatically, prompt the control device to trigger adjustments to driving functions and / or vehicle functions. Furthermore, preferably, this can also be understood as providing the adjustment notification (preferably via a warning device) for transmission to the control device of the traffic participant.

[0028] Additionally or alternatively, “providing adjustment notification to at least one self-regulating traffic participant” can be understood as providing adjustment notification (preferably via a warning device) to the user or driver of the self-regulating traffic participant. This type of output can be graphical (e.g., via a display device, such as a monitor and / or head-up display and / or an augmented reality output device of the vehicle, and / or via the user's or self-regulating traffic participant's (mobile) terminal device), and / or acoustic and / or tactile (especially via the output device of the self-regulating traffic participant and / or the vehicle, and / or via the user's or self-regulating traffic participant's (mobile) terminal device and / or mobile assistance device).

[0029] Here, the adjustment notification can represent the operation instructions for the user / driver of the traffic participant or vehicle, from which the user / driver can learn what kind of intervention should or should be performed in order to adjust at least one driving function and / or vehicle function.

[0030] Preferably, at least one driving function and / or vehicle function is adjusted to accommodate at least one surrounding traffic participant in order to prevent and / or mitigate potentially dangerous traffic situations (especially those involving traffic safety) involving the vehicle's own traffic participant and / or surrounding traffic participants. For example, a dangerous traffic situation might be a collision between the vehicle's own traffic participant and surrounding traffic participants.

[0031] However, it is also possible that only the driver and / or surrounding road users are involved in the potentially dangerous traffic situation. That is, for example, during an overtaking maneuver (where the driver ignores oncoming surrounding road users and overtakes another road user in the lane the driver is using as the overtaking lane), if the driver merges into the intended lane too early, the other road user may be at risk (while the oncoming surrounding road users are not involved in the dangerous situation).

[0032] Preferably, the adjustment notification characterizes a (potential) critical traffic situation and / or the cause and / or potential cause of such a (potential) critical traffic situation, such as a cyclist who is expected to cross the vehicle's path (e.g., a cyclist who is difficult for the driver to see or is obstructed).

[0033] "Adjustment to [at least one driving function and / or vehicle function]" can be understood as changing the operating parameters of the driving function, enabling and / or disabling the driving function and / or vehicle function.

[0034] "Traffic participants" should be understood in particular as vehicles, such as motor vehicles, and preferably as road vehicles, agricultural machinery, buses, trams, rescue vehicles, emergency vehicles, electric micromobility devices (such as, for example, electric-assisted bicycles (Pedelecs)), and / or as two-wheeled vehicles (such as bicycles) and / or motorized two-wheeled vehicles (such as motorcycles and / or scooters), and / or as electric scooters, and / or as road users (such as pedestrians), and / or users of other (pre-defined) lanes (general movement paths, such as, for example, bicycle lanes and / or sidewalks). In particular, "traffic participants" should be understood as motorized and / or non-motorized traffic participants.

[0035] For example, a road user can be a (motorized) vehicle, such as a passenger car (e.g., a car). A surrounding road user can be a so-called "VRU" (an abbreviation for "vulnerable road user," German: "ungeschützte Verkehrsteilnehmer"). A VRU specifically refers to a road user not surrounded by a (protective) (driver's) cabin, such as a pedestrian, cyclist, or motorcyclist.

[0036] Here, the invention can be particularly advantageously applied to combinations in which either the traffic participant itself or a surrounding traffic participant is a VRU (while another traffic participant is not a VRU, for example, a (motorized) vehicle with a passenger compartment).

[0037] It is also conceivable that traffic participants are individuals using mobility aids (such as walkers or wheels), or individuals carrying mobile devices (such as smartwatches and / or smartphones and / or fitness equipment), such as those engaging in physical activity (e.g., joggers). Due to their higher speed and intentional pursuit of personal running goals, joggers, as traffic participants in road traffic, may also face particular dangers. Here, the mobile device can preferably have a warning device.

[0038] Preferably, the adjustment notification is a warning notification. Preferably, the warning notification should be provided to at least one such traffic participant who is to be notified, i.e., the traffic participant who should be given priority to be warned or will be given priority to be warned.

[0039] Preferably, at least one surrounding traffic participant is a potential collision participant (within the surrounding area of ​​the traffic participant). Preferably, a warning notice is provided to at least one traffic participant to avoid a potential collision with the at least one (potentially colliding) surrounding traffic participant in the collision zone of the preferred road traffic surrounding area (especially the traffic area).

[0040] In particular, at this time, at least one driving function and / or vehicle function is adjusted to accommodate at least one surrounding traffic participant located in and / or moving within the surrounding area of ​​the traffic participant, in order to avoid a potential collision with said at least one surrounding traffic participant. Here, the adjustment of the driving function and / or vehicle function may be, for example, braking, acceleration, avoidance, and / or change of driving route, and such adjustment should be facilitated and / or triggered by providing a warning notification (e.g., automatic triggering, or triggered by manual intervention by the driver / user in guiding the vehicle or the movement of the traffic participant after a warning notification is output to the driver or user of the vehicle).

[0041] Preferably, the collision zone is an area where the travel lanes (e.g., driving lanes, bicycle lanes, safety barriers) or movement paths of traffic participants (preferably themselves and surrounding traffic participants) intersect. Preferably, the collision zone is (e.g., at an intersection) an area where the movement paths of their own traffic participants, as defined and / or pre-defined by traffic rules, can intersect with the movement paths of surrounding traffic participants, as defined and / or pre-defined by traffic rules (e.g., the bicycle lane intersects with the movement path defined for a left-turning vehicle).

[0042] It is conceivable that the collision zone (especially the point of collision) is a (traffic) area where the predicted movement paths of the two traffic participants intersect.

[0043] In addition to fixed collision zones or points (which are pre-defined by the structure of the intersection and / or traffic guidance at the intersection), collision zones can also be non-fixed. Such non-fixed collision zones can be, for example, collision zones or points that, for example, during overtaking, are generated by the combination of the overtaking traffic participant and an oncoming traffic participant in the same lane as the overtaking traffic participant. Here, as in turning operations, it is crucial to maintain a sufficient safe distance between the overtaking traffic participant and the oncoming traffic participant.

[0044] Warning devices, especially processor-based warning devices.

[0045] The warning device is preferably located in the surrounding area. Preferably, the warning device is a warning device for its own traffic participants (to be notified) and / or for surrounding traffic participants (e.g., those at risk of a potential collision). Preferably, the warning device is (fixedly and / or explicitly) attached to its own traffic participant or surrounding traffic participants. Preferably, the warning device is vehicle-attached, i.e., a component of the vehicle that cannot be removed without damage.

[0046] (Additionally or alternatively) It is also conceivable that the warning device is a warning device attached to a traffic participant, such as being a component (especially a fixed or non-destructively detachable) of the traffic participant's mobile terminal device (e.g., a smartwatch, smartphone, fitness equipment) and / or mobility aids. Preferably, the warning device moves with the traffic participant (either the participant or the surrounding area).

[0047] (Additionally or alternatively) It is also conceivable that the warning device is a fixed warning device, for example, as part of a traffic infrastructure device. For example, this type of warning device can be advantageously used in particularly dangerous traffic areas, where accidents or (serious) dangerous traffic situations occur frequently, and / or where there is a fear that an accident or (serious) dangerous traffic situation may occur.

[0048] Here, the “predicted movement path” of a traffic participant (either themselves or in the surrounding area) should be understood in particular as the (determined and / or invoked) future movement path of a traffic participant (either themselves or in the surrounding area), especially (preferably at least partially) in the surrounding area. The “predicted movement path” of a traffic participant (either themselves or in the surrounding area) should be understood in particular as the (future) movement path along which the movement or movement process of a traffic participant (either themselves or in the surrounding area) is foreseeable.

[0049] The warning device may retrieve and / or receive (e.g., from a fixed traffic infrastructure device) predicted movement paths of traffic participants (either itself or in the vicinity) from a preferably internal storage device. However, it is also conceivable that the warning device determines or calculates the predicted movement paths (in real time).

[0050] It is also conceivable that the warning device attached to itself can call upon its own navigation data and / or route guidance data to determine the predicted movement path.

[0051] Preferably, the warning device determines the predicted movement path of its own traffic participant and / or surrounding traffic participants based on at least one of the corresponding traffic participants, especially the current (detected and / or transmitted) vehicle location.

[0052] When a warning device is linked to its own traffic participant, the vehicle position of the traffic participant can be determined and / or determined based on GNSS data (Global Navigation Satellite System, which is a general term for positioning using existing and future global satellite systems).

[0053] When a warning device is attached to its own traffic participant, the vehicle positions of surrounding traffic participants can be determined based on environmental data detected by the traffic participant's own environmental detection device used to detect the surrounding environment of the traffic participant or the vehicle's surrounding environment.

[0054] Preferably, the surrounding environment detection device for detecting the vehicle's surrounding environment is selected from a group of sensors that includes: a (color) camera, a front-facing camera, a rear-facing camera, an infrared camera, a LiDAR sensor (Light Detection and Ranging or Light Imaging, Detection and Ranging), a radar sensor, an ultrasonic sensor, and combinations thereof. Preferably, the surrounding environment detection device generates spatially resolved (especially 2D and / or 3D) environmental data (of the corresponding vehicle's surrounding environment).

[0055] However, it is also conceivable that (especially fixed) traffic infrastructure devices detect environmental and / or location data (e.g., via V2X communication) related to the (especially current) location of surrounding traffic participants (and / or their own traffic participants) and transmit it to (e.g., their own traffic participants) warning devices (to determine the corresponding predicted movement paths). It is also conceivable that traffic infrastructure devices determine at least the predicted movement paths of their own traffic participants and / or surrounding traffic participants and transmit them to warning devices.

[0056] Alternatively or additionally, V2X and / or V2V communication data (sent by the traffic participant or by surrounding traffic participants) are used to determine the (current) location of the participant and / or surrounding traffic participants, and / or to determine the corresponding predicted movement path. This data preferably includes (current) location and / or movement data, and / or information about category / type, and / or information about geometry (length / width), and / or information about the role of the respective traffic participant.

[0057] According to the invention, the warning device (especially in the computer-implemented method steps) determines at least one notification parameter (preferably a warning parameter) representing the provision (and / or output) of an adjustment notification (preferably a warning notification) based on at least one predicted motion path, and preferably based on multiple predicted motion paths (based on the motion path predicted for itself as a traffic participant and based on the motion path predicted for surrounding traffic participants).

[0058] Preferably, the notification parameter (preferably a warning parameter) represents whether an adjustment notification (preferably a warning notification) should be provided, and / or at what time (a warning time point) the adjustment notification (preferably a warning notification) should be provided - for example, for outputting to the user and / or for triggering an adjustment to the at least one driving function and / or vehicle function.

[0059] According to the invention, the warning device (especially in the computer-implemented method step) determines (at least one) notification parameter (preferably a warning parameter) based on at least one, especially statistical, group prediction parameter, which is used to determine the prediction quality of the predicted motion path, and preferably to determine the prediction quality of at least one of the predicted motion paths.

[0060] Preferably, at least one group prediction parameter is determined based on global group motion data of multiple traffic participants generated in multiple distinct traffic zones. "Global group motion data" should be understood here specifically as motion data detected and / or recorded not only locally, i.e., for exactly one geographically defined, coherent traffic zone, but for several, preferably unrelated, coherent, and / or non-overlapping traffic zones. Preferably, these traffic zones are distributed across at least one country, preferably multiple countries, and particularly preferably across at least two continents. It is conceivable that these traffic zones are evenly distributed. Preferably, each traffic zone has at least one collision zone, and preferably multiple collision zones.

[0061] Preferably, the group prediction parameters involve at least one (preferably exactly one) driving action of the traffic participant and / or surrounding traffic participants. Thus, the group prediction parameters associated with the predicted movement path of the traffic participant could, for example, involve the traffic participant's (e.g., left-turn) turning action at an intersection. Similarly, the group prediction parameters associated with the predicted movement paths of surrounding traffic participants could involve the surrounding traffic participants crossing the intersection (e.g., in a straight line). Preferably, the multiple traffic zones each include at least one collision zone derived from at least one driving action of the traffic participant and at least one driving action of surrounding traffic participants.

[0062] Preferably, at least one group prediction parameter used to determine the prediction quality of the predicted motion paths for its own traffic participants represents global group motion data generated in multiple different traffic areas for multiple traffic participants (preferably of the same type as the traffic participants for whom motion paths were predicted).

[0063] Multiple distinct traffic zones should be understood herein to be at least 5, preferably at least 10, preferably at least 50, preferably at least 100, preferably at least 1000 (non-overlapping) traffic zones (preferably each having at least one collision zone).

[0064] Preferably, these traffic areas have a geometric dimension greater than 20m, preferably greater than 50m, preferably greater than 100m, preferably greater than 200m, preferably greater than 500m, and particularly preferably greater than 800m in at least one direction, preferably in at least two (preferably perpendicular to each other) directions.

[0065] Preferably, the multiple traffic zones, which are distinct from each other, each include at least one, and preferably multiple, collision zones and / or intersection zones (lane intersection zones). Preferably, each traffic zone includes at least one collision zone, preferably several collision zones, and particularly preferably at least five collision zones (wherein, these collision zones preferably relate to the category or type of the traffic participant involved, i.e., the category or type of the participant itself and the category or type of the surrounding traffic participants).

[0066] Multiple traffic participants (especially those of the same type or category as the traffic participants whose movement paths were predicted for them) should be understood here in particular to be at least 50 (especially different), preferably at least 100, preferably at least 500, preferably at least 1000, preferably at least 2000, and particularly preferably at least 10000 different traffic participants.

[0067] Preferably, the warning device determines the notification parameter based on a (especially statistical) group prediction parameter used to determine the quality of the predicted movement path for itself as a traffic participant, and based on a (especially statistical) group prediction parameter used to determine the quality of the predicted movement path for surrounding traffic participants.

[0068] The proposed method specifically involves the use of big data for early warning timing, particularly in V2X sensing capabilities, and especially for optimized applications, particularly big data analytics.

[0069] Compared to traditional ADAS (Advanced Driver Assistance Systems), the new V2X functionality enables significantly earlier notification of potential hazards to the driver (→early warning). One example is V2X-based intersection assist, which warns of a cyclist in the collision path. With an earlier warning time of >5 seconds, the driver can react calmly to the situation and, for example, mitigate the danger in time by braking earlier.

[0070] The proposed method offers the advantage of adapting the functionality to specific situations when using a sufficiently large dataset. For example, a dataset recorded by a camera drone depicting the behavior of cyclists and passenger vehicles at intersections.

[0071] Preferably, for multiple different categories or types of traffic participants, at least one statistical and / or probabilistic motion path parameter and / or at least one statistical group prediction parameter for determining the prediction quality of at least one predicted motion path are provided respectively. Here, preferably, each motion path parameter characterizes a motion path determined by global group motion data of the corresponding category or type of traffic participants. Preferably, the corresponding group prediction quality is determined based on global group motion data of multiple traffic participants generated in multiple, distinct traffic areas.

[0072] Preferably, the motion path parameters of the predefined category or type of traffic participants and / or the estimated parameters (used to determine the estimated parameters) are determined solely (or substantially solely) based on global group motion data of multiple traffic participants of the same type or category. Here, global group motion data may originate from traffic participant motion data determined in traffic situations where these participants performed their planned driving maneuvers without interference from other traffic participants (e.g., because no other traffic participants were present). For example, global group motion data may originate from the motion data of multiple such traffic participants who performed planned left turns without interaction (e.g., without concern about collisions with cyclists). Preferably, the same applies to surrounding traffic participants.

[0073] Preferably, the predicted motion path for the traffic participant involves driving maneuvers that the participant (in the future and / or expected) will perform and / or plan. Preferably, in the global group motion data of multiple traffic participants, only (or substantially only) such motion data of multiple traffic participants are considered, i.e., these motion data represent the same kind of driving maneuvers as the participant (in the future and / or expected) will perform and / or plan. Preferably, the same applies to surrounding traffic participants.

[0074] Preferably, the category or type of traffic participant is selected from a group of categories or types that include: vehicles, motor vehicles, passenger cars, public transport vehicles, private transport vehicles, the role of a transport vehicle, road vehicles, agricultural machinery, the purpose of a transport vehicle, (public) buses, trams, rescue and emergency vehicles, electric micromobility devices (such as, for example, electric-assisted bicycles (Pedelecs)), two-wheeled vehicles, bicycles and / or motorized two-wheeled vehicles (such as motorcycles and / or scooters), and / or electric scooters and / or road users (such as, for example, pedestrians, cyclists, motorcyclists, motorized and / or non-motorized traffic participants, VRUs, etc., and combinations thereof).

[0075] In a preferred approach, the surrounding area and the multiple traffic zones are distinct from each other in pairs. In other words, the quality of the population prediction determined based on global population motion data generated at the multiple traffic zones can be applied to (and is persuasive about) the surrounding area, which is distinct from the multiple traffic zones.

[0076] In a more preferred method, the warning device determines whether an adjustment notification should be provided based on a pre-set and / or pre-settable minimum prediction probability parameter and a population prediction parameter. This provides the advantage that an adjustment notification is provided only when the minimum prediction probability is reached. This ensures that a critical situation (e.g., a potential collision) is adequately identified, and that traffic participants are not unnecessarily distracted and / or warned.

[0077] In a more preferred method, a safety parameter is preset and / or pre-set to characterize the minimum safe distance between the warning device and surrounding traffic participants, preferably the minimum safe distance to be observed during its movement. Here, this safety parameter may be related to the type of planned driving maneuver and / or the movement parameters of the traffic participants and / or the category and / or geometry (e.g., length / width) of the traffic participants. However, it is also conceivable that the safety parameter may be based on individual presets and / or personal preferences, thereby taking into account, for example, the driver's sense of security.

[0078] Preferably, when the predicted movement paths of at least two traffic participants (especially the participant themselves and surrounding traffic participants) are at least temporarily less than one (or the aforementioned) particularly preset and / or presettable minimum safe distance, a warning notification may be provided / output and / or provided and / or (e.g., to the driver) to alert at least one traffic participant (preferably the participant themselves and / or surrounding traffic participants) to a potential danger arising from a potential collision with another traffic participant. This provides the advantage that the movement of traffic participants can be altered to prevent situations where the actual distance falls below the minimum safe distance.

[0079] Preferably, based on safety parameters, the warning device (in order to determine notification parameters) determines the probability that the traffic participant and surrounding traffic participants approach each other to a distance less than a preset and / or preset safe distance, based on the predicted movement path of the traffic participant and at least one group prediction parameter used to determine the predicted movement path quality of the traffic participant, and based on the predicted movement paths of surrounding traffic participants and at least one group prediction parameter used to determine the predicted movement path quality.

[0080] In a more preferred method, the (at least one) group prediction parameter represents the category and / or type and / or class of the traffic participant and / or surrounding traffic participants, and represents a topological parameter representing the topology of at least one movement trajectory of the traffic area or (the) surrounding area (preferably including the collision area), especially the structure of the intersection, and an extension parameter representing the geometric extension of the intersection area representing the preset movement trajectory of the traffic participants.

[0081] Preferably, at least one group prediction parameter is related to the category and / or type (and / or class) of its own traffic participant and / or surrounding traffic participants, and is related to the topological parameter of the topology of at least one motion trajectory characterizing the traffic area or (the surrounding area) (preferably including the collision area), especially the structure of the intersection, and is related to the extension parameter of the geometric extension range of the intersection area characterizing the preset motion trajectory of the traffic participants.

[0082] In a more preferred method (preferably within the scope of computer-implemented method steps), a surrounding situation determination device for the traffic participant and / or surrounding traffic participants (and / or especially fixed traffic infrastructure devices) (e.g., due to data access) collects and / or determines and / or receives surrounding situation data, which characterizes at least one influencing parameter, and preferably multiple influencing parameters.

[0083] Here, at least one influencing parameter, and preferably multiple influencing parameters, are selected from a group of influencing parameters, namely, the group of influencing parameters includes: topological parameters characterizing at least one motion trajectory in the collision area, especially the topology of the structure of the intersection (X-shaped intersection; T-shaped intersection; merging angle of the road / motion trajectory / lane merging into another lane or motion trajectory or lane; curvature change process of the lane and / or lane and / or motion trajectory; direction of movement along the lane, e.g., a two-way bicycle lane); parameters characterizing at least one other traffic participant, especially the presence in the collision area; traffic density; Speed ​​parameters representing at least one speed prescribed for at least one (involved) traffic participant in the collision area; traffic rule parameters representing preset traffic rules, especially right-of-way rules; range parameters representing the geometric extent of the collision area and / or intersection area of ​​the preset movement trajectory of traffic participants; traffic participant parameters representing the level and / or group of the traffic participants involved; cultural parameters, especially those representing the geographical area; environmental parameters, especially those representing at least one environmental condition that may affect visibility; time parameters representing the season and / or time of day; weather parameters representing the current weather; and combinations thereof.

[0084] Preferably, the group prediction parameter characterizes the quality of predictions of the group movement behavior of multiple traffic participants under various surrounding conditions, particularly statistical conditions, which characterize at least one influencing parameter (and preferably multiple influencing parameters). Here, preferably, at least some of the various surrounding conditions involve traffic areas that are different from each other.

[0085] Preferably, at least one (and preferably each) group prediction parameter is determined based on at least one influencing parameter and preferably based on multiple influencing parameters.

[0086] Preferably, the predicted movement paths of the traffic participants themselves and / or surrounding traffic participants are determined based on at least one influencing parameter and preferably based on multiple influencing parameters.

[0087] Preferably, the plurality of influencing parameters are at least two, preferably at least three, preferably at least four, preferably at least five, preferably at least seven (selected from the above group) influencing parameters.

[0088] In a more preferred method, a surrounding situation identification device for the user and / or surrounding traffic participants (and / or especially fixed traffic infrastructure devices) collects surrounding situation data, which represents at least two, preferably at least three, preferably at least four, preferably at least five, preferably at least eight, preferably at least ten, preferably at least fifteen influencing parameters selected from the above-mentioned group of influencing parameters.

[0089] The surrounding conditions determination device can be (e.g., as described above) a surrounding environment detection device. It is also conceivable that the surrounding conditions determination device can access environmental data generated by the surrounding environment detection device, or can invoke such data or data derived therefrom. However, it is also conceivable that the surrounding conditions determination device invokes and / or receives and / or determines map data and / or navigation data and / or temperature data, etc., thereby enabling it to (e.g., by a warning device) determine at least one value of a corresponding influencing parameter.

[0090] Preferably, the at least one group prediction parameter is determined based on the aforementioned parameters, namely, based on the category and / or type (and / or) class of the traffic participant and / or surrounding traffic participants, and based on the topological parameters of the surrounding area (e.g., determined by the warning device and / or traffic participants and / or by the surrounding situation determination device), and based on a characteristic parameter such as the extension range of the intersection area (preferably determined by the surrounding situation determination device).

[0091] For example, topological parameters characterizing at least one motion trajectory in the collision area, especially the structure of an intersection, can be understood as: the number of arms of lanes and / or traffic lanes merging into the intersection or node, the number of traffic lanes, the geometry of the intersection (length of the widening section, radius), the form of the intersection (e.g., X-shaped intersection, T-shaped intersection), the merging angle of the road / motion trajectory / lane merging into another lane or motion trajectory, the curvature change process of the lanes and / or traffic trajectories, the connection relationship between at least two and preferably several traffic lanes merging into the node or intersection (that is, from which traffic lane entering the arm to which traffic lanes leaving the arm), the permitted direction of movement along the lane (e.g., a bicycle lane that allows two-way traffic), and combinations thereof.

[0092] Preferably, the warning device pre-sets and / or provides a driving operation planned and / or (future and / or anticipated) to be performed by the vehicle itself or by surrounding traffic participants. Preferably, the predicted movement path characterizes the planned and / or (future and / or anticipated) driving operation to be performed. This driving operation may, for example, be selected from the group of driving operations including left turns, right turns, (straight) crossing of intersections, overtaking operations, etc., and combinations thereof.

[0093] Preferably, at least one preset influence parameter is assigned to each driving operation (selected from the driving operation group), and preferably multiple preset influence parameters are assigned.

[0094] Preferably, the surrounding situation assessment device invokes the influence parameters assigned to the planned and / or (future or expected) driving operation, and collects and / or determines data about these influence parameters.

[0095] Preferably, the warning device determines at least one population prediction parameter and / or data characterizing the predicted motion data based on these influencing parameters.

[0096] It is conceivable that the group motion data is at least partly vehicle-to-everything (V2X) communication data, preferably vehicle-to-vehicle (V2V) communication data. Preferably, the group motion data is generated and / or determined at least partially (preferably only) based on V2X communication data, and more preferably based on V2V communication data.

[0097] In a more preferred method, the warning device determines the predicted movement path of at least one (especially potentially colliding) surrounding traffic participant based on vehicle-to-everything (V2X) communication data and preferably vehicle-to-vehicle (V2V) communication data acquired by the communication device and sent by at least one (especially potentially colliding) surrounding traffic participant.

[0098] Preferably, the predicted motion path of at least one (especially a potential collision) surrounding traffic participant is determined based on group data of multiple traffic participants generated in multiple traffic zones (collision zones) that are different from each other.

[0099] In a more preferred method, for at least one category and / or at least one type of (surrounding) traffic participant, preferably for multiple categories or types of traffic participants, a motion model containing multiple motion parameters is provided to the warning device (especially separately). Preferably, the motion model, or the data representing it, and / or the multiple motion parameters are stored in a storage device within the traffic participant and / or within the warning device. Preferably, a (corresponding) predicted motion path is determined based on the motion model.

[0100] The advantage this provides is that traffic participants can determine their respective predicted movement paths in real time based on the motion model without time loss, and particularly preferably based on the positions of surrounding traffic participants (without, for example, communicating with a backend server).

[0101] Here we can imagine a motion model that assumes constant speed or constant acceleration.

[0102] Here, the motion model can be associated with a driving maneuver that the respective traffic participant plans and / or (future or anticipated) to perform. That is, for example, a cyclist who wants to cross the intersection in a straight line can select a motion model with a constant speed (via a warning device), while a cyclist who wants to turn right at the intersection can select a motion model with a constant acceleration.

[0103] Preferably, the motion model is associated with at least one influencing parameter, and more preferably with multiple influencing parameters. Preferably, the motion model assigned to and / or selected based on a driving operation planned and / or (future or anticipated) by the corresponding traffic participant is associated with at least one, and more preferably with multiple influencing parameters assigned to (the driving operation planned and / or future or anticipated) by the corresponding traffic participant.

[0104] It is conceivable to detect at least one, and preferably multiple, data characterizing motion parameters (e.g., speed and / or location data of traffic participants), and based on, for example, changed motion parameters and / or location data, to update and / or recalculate motion paths predicted according to the motion model (repeatedly and / or at regular time intervals).

[0105] Preferably, the population prediction parameters used to determine the predicted quality give a confidence level of the prediction (predicted motion path).

[0106] In a more preferred method, at least one predicted time point for arrival at the collision zone, particularly the collision point, is determined for the traffic participant and / or for surrounding traffic participants. Preferably, the probability that the traffic participant and surrounding traffic participants will arrive at the collision zone (and / or be in the collision zone) within the predicted time point is determined based on a preset time period (“gap”).

[0107] Preferably, the (potential or estimated) collision point and / or collision area are determined based on the predicted movement paths of the user and / or surrounding traffic participants. In particular, their locations are determined. However, it is also conceivable that, for example, based on map data and / or navigation data, at least one potential collision point and / or collision area (in particular, its location) has already been pre-defined based on the positions of the user and / or surrounding traffic participants. Therefore, for example, for a pre-defined intersection that the user is approaching, the location of potential collision points with other surrounding traffic participants may already be pre-defined.

[0108] Preferably, the estimated time to reach the collision point or collision area is determined based on the location of the collision point or collision area and the predicted movement path of the traffic participant (marked in the figure as the so-called "time to path intersection point", abbreviated as TTPIP, i.e., the TTPIP of the traffic participant in this case).

[0109] Preferably, based on the location of the collision point or collision area and the predicted movement paths of surrounding traffic participants, the estimated time to reach the collision point or collision area is determined (labeled in the figure as the so-called estimated "time to path intersection point", abbreviated as TTPIP). pred That is, the TTPIP of surrounding traffic participants. pred ).

[0110] Preferably, the predicted motion paths (respectively) and collision points are based on a one-dimensional perspective (therefore, the motion of traffic participants is modeled as linear or path-like motion, and thus there are no two-dimensional collision regions, but rather point-like collision points). This provides the advantage of relatively low computational cost (e.g., for informing parameters).

[0111] Preferably, at least one group prediction quality is provided and / or (especially determined by the warning device), the group prediction quality giving a probability density function of the prediction error distribution (especially for the corresponding type or category of the corresponding traffic participant and / or preferably for the driving operation or predicted movement path), or characterizing the probability density function. Here, the prediction error distribution can be for different TTPIPs respectively. pred Provide or determine, and characterize the calculated value of TTPIP at the path intersection point (PIP). pred With TTPIP for actual duration true The difference between them, wherein the actual duration is preferably determined by global group motion data. For example, it can be based on preset (preferably multiple) TTPIPs. pred With TTPIP pr–d - TTPIP true Given the difference, the statistical relationship of the probability P is presupposed (and / or can be determined based on the population's predicted quality) (see also the statistical relationship of the probability P). Figure 12 (Explanation).

[0112] Preferably, the prediction error distribution can be based on the probability density function of the prediction error distribution of its own traffic participants (especially for multiple TTPIPs of its own traffic participants). pred), and based on the probability density function of the prediction error distribution of its own traffic participants, and preferably based on a preset duration (indicated in the figures as "gap" [T gap,1 , T gap,2 ] or "T gap (This is a measure of the probability that both the user and surrounding users must reach the point of impact (PIP) within the specified time period to trigger the provision of adjustment notifications (or warning notifications)).

[0113] This probability metric can be determined by integrating the probability density function of the prediction error distribution of its own traffic participants and the probability density function of the prediction error distribution of surrounding traffic participants. For example, this probability metric can be determined based on the probability density function of the prediction error distribution of its own traffic participants and the probability density function of the prediction error distribution of surrounding traffic participants. Figure 13 The accompanying drawings are provided within the scope of the description. P X (car, VRU, ∆TTPIP pred , T gap,1 , T gap,2 The formula for determining the prediction error distribution is given here using a car (as a traffic participant) and a VRU (Vehicle Refurbished Unit) as an example. Generally, the probability density function of the prediction error distribution of the car should be replaced with the probability density function of the prediction error distribution of the car itself, and the probability density function of the prediction error distribution of the VRU should be replaced with the probability density function of the prediction error distribution of the surrounding traffic participants.

[0114] The notification parameter can be determined based on the measured probability and a preset minimum predicted probability parameter. Therefore, for example, the measured probability can be compared to a minimum probability. An adjustment notification can only be provided and / or output if, for example, the measured probability exceeds the preset minimum predicted probability.

[0115] In a more preferred method, the warning device determines at least one local collision parameter based on the determined (detected) environmental data and / or map data related to the surrounding area (especially the collision area), and determines the predicted movement path of at least one (especially a potential collision) surrounding traffic participant based on the local collision parameter.

[0116] In a more preferred method, the warning device determines at least one local collision zone parameter based on the determined (detected) environmental data and / or based on (navigation) map data related to the collision zone stored on the internal storage of the traffic participants and / or the warning device, and determines the predicted movement path of at least one potential surrounding traffic participant based on the local collision zone parameter.

[0117] In a more preferred approach, the warning device is provided with multiple influence parameters determined based on global group motion data. These influence parameters are, in particular, independently updatable via an external server, especially a backend server (e.g., during the update process, particularly through a communication connection with an external communication partner). This provides the advantage that the influence parameters used to determine group prediction parameters can be updated as the global group motion data is expanded.

[0118] The present invention also relates to a method for determining, particularly at least local group movement data and preferably (at least partial segments) global group movement data for generating a global map of at least partial segments of group movement data (preferably predicted movement paths) characterizing multiple movement paths of multiple different traffic participants. This method is performed by means of at least one group data identification device for at least one traffic participant, particularly a vehicle, that determines the group data. The global map of at least partial segments of the group movement data should be understood in particular as including group movement data at multiple different traffic areas. Preferably, multiple different traffic areas can be selected as described within the scope of the method for providing adjustment notifications.

[0119] According to the present invention, the crowd data identification device determines local crowd movement data based on vehicle-to-vehicle communication data received by the communication device of at least one traffic participant determining the crowd data and transmitted by at least one traffic participant transmitting vehicle-to-vehicle communication data, and provides the local crowd movement data, particularly to the communication device of the traffic participant determining the crowd data, for transmission to an external storage device, preferably to an external backend server. Preferably, the communication device transmits the local crowd movement data to the external storage device.

[0120] Preferably, multiple crowd data identification devices are provided to determine the traffic participants' group data. These devices each determine local group movement data and provide the local group movement data for transmission to an external storage device. The method steps described below are based on only one crowd data identification device for determining the traffic participants' group data, but are similarly applicable to multiple crowd data identification devices. Preferably, the multiple crowd data identification devices move in and / or are located in different traffic areas from each other.

[0121] Local group movement data can be vehicle-to-vehicle communication data of multiple (different) traffic participants, which includes the location data of the respective traffic participants and the time point at which the location data of the respective traffic participants is confirmed and / or detected, or includes parameters characterizing and / or derived parameters.

[0122] Preferably, the local group movement data includes local group movement data of categories of traffic participants used to send vehicle-to-vehicle communication data.

[0123] Preferably, the localized group motion data involves motion data of multiple traffic participants (preferably vehicles) identified within a preset time period, such as within 1 minute and / or within 10 minutes and / or within 1 hour, which is recorded and / or generated in a traffic area having a geometric dimension of at least 25m, preferably 50m, preferably 100m, preferably 200m, preferably 500m, and particularly preferably at least 800m.

[0124] Preferably, when determining local group movement data, the vehicle-to-vehicle communication data received by the traffic participants determining the group data (within a preset and / or preset time period) from all traffic participants who sent vehicle-to-vehicle communication data is considered.

[0125] Preferably, at least one traffic participant sending vehicle-to-vehicle communication data is located in the surrounding environment of the traffic participant determining the group data (especially within the communication range of the communication device of the traffic participant determining the group data, within which data exchange is performed, especially by sending and / or receiving vehicle-to-vehicle communication data).

[0126] For example, datasets of cyclists and passenger vehicles at intersections, recorded by camera drones, are extremely rare globally. As a result, it is very difficult to parameterize, for example, the predicted movement paths (which represent the movement behavior of multiple traffic participants of the same category or type) (as opposed to the aforementioned influencing parameters and other factors), and it is only possible to attempt to formulate universally effective functional parameterizations using individual random samples.

[0127] To determine the warning timing by considering various possible factors, it is proposed to identify multiple, as diverse as possible, data points by identifying traffic participants in the group data. For this purpose, it is preferable for fleet vehicles (VE test vehicles, ideally customer vehicles) to record anonymized Car2X data, for example, as they approach an intersection, thereby reproducing, for example, the trajectories of cyclists and vehicles in the intersection area. Thanks to Car2X communication, the recording vehicle receives high-quality data once it enters the effective radio range. Here, the recording vehicle or the traffic participant collecting the group data does not necessarily have to be a vehicle located at the same intersection as the cyclists – simply passing by or being within the effective reception range is sufficient.

[0128] Now, preferably, this data (local group movement data) is enriched at the backend with other data (such as map data), and preferably, influencing factors for the warning time points are identified. First, the recorded trajectories of passenger vehicles and cyclists can be clustered based on the direction of movement and the actions performed at the intersection. Subsequently, for example, it can be studied whether the cyclist's direction of travel (along the regular direction of travel on the bike lane / in the opposite direction), the presence of passenger vehicles at the intersection, and the impact on movement behavior, and thus also on prediction, can be measured.

[0129] Now, the influencing factors calculated using this "big data" approach are fed back to the vehicle as a set of parameters. For example, at intersection A in a rural area (where bicycles always travel at high speed), the vehicle can display a warning 7 seconds before passing the intersection, while at intersection B in an urban area, the warning is only displayed 5 seconds in advance. For the parameterized configuration of the vehicle, both intersection-specific parameters and standardized parameters (e.g., for intersection type / size / ...) can be envisioned.

[0130] For example, vehicle-to-vehicle communication data can be communication data sent, especially according to the communication technology ETSI ITS G5.

[0131] For example, this type of vehicle-to-vehicle communication data can be sensor data (or parameters derived therefrom) recorded and / or collected and / or generated by the surrounding environment detection devices (onboard sensors) of traffic participants that send vehicle-to-vehicle communication data.

[0132] For example, so-called "Collective Perception Messages" (CPM) can be used as this type of vehicle-to-vehicle communication data. In particular, this type of CPM message can contain information about the motion dynamics, position, and other attributes of objects (e.g., vehicles, pedestrians, animals, and other objects), detected by (especially by) one or more surrounding environment detection devices (such as radar, lidar, and cameras) of the traffic participant sending the vehicle-to-vehicle communication data, and accessible to the transmitting traffic participant. CPM, in particular, enables interoperable exchange of basic information about (the factors necessary to interpret the transmitted data), about the transmitting traffic participant, their sensor capabilities, sensed objects, and / or road-related sensing areas.

[0133] CPMs are generated, especially in a quasi-periodic manner, such as by CPM generation events. This offers the advantage of keeping the amount of data transmitted small by CPM transmissions triggered based on the perception of new objects and / or new attributes.

[0134] The advantage of this type of vehicle-to-vehicle communication data is, for example, the ability to acquire information (especially about motion behavior or location and / or time) related to traffic participants who do not themselves have the ability to send (or receive) vehicle-to-vehicle communication data, and thus it can be taken into account in local group data (and consequently in global group motion data).

[0135] Preferably, the vehicle-to-vehicle communication data is so-called "Cooperative Awareness Messages" (CAM). This type of vehicle-to-vehicle communication data is used to notify nearby communication partners (such as other road users) of its presence cyclically, periodically, or at regular time intervals (by sending CAM messages).

[0136] The advantage of using CAM messages is that, for example, a new CAM message is sent when the distance from the location where the previous CAM message was sent exceeds a preset distance (e.g., 4m). This allows for very intensive data collection.

[0137] Furthermore, the advantage of using CAM messages is that when a traffic participant's route changes—that is, when the orientation in the 1984 World Geodetic System (WGS84) changes by more than 4° compared to the last transmitted CAM—a new CAM is sent. This provides, for example, very accurate acquisition of the traffic participant's movement path during turning maneuvers.

[0138] Furthermore, the advantage of using CAM messages is that when a traffic participant's current speed exceeds the speed value sent in the last CAM by more than 0.5 m / s, a new CAM is sent, thus reflecting large acceleration values ​​well in the group data.

[0139] Preferably, the transmitted vehicle-to-vehicle communication data includes information related to the time of generation of the vehicle-to-vehicle communication data and the location of the traffic participants.

[0140] Preferably, the transmitted vehicle-to-vehicle communication data additionally or alternatively includes information related to at least one geometric and / or motion parameter of the traffic participant, which is selected from a group of parameters including: direction of travel, orientation of the traffic participant, direction of the traffic participant, speed, length, width of the traffic participant, longitudinal acceleration, turning radius, yaw rate, acceleration control, lane position, lateral acceleration, steering wheel angle, vertical acceleration, power rating, toll area, etc., and combinations thereof.

[0141] Preferably, the transmitted vehicle-to-vehicle communication data additionally or alternatively includes information related to the role of the traffic participant (e.g., special vehicle, ordinary passenger vehicle, etc.). This provides the advantage that data relating only to the behavior of the special vehicle or the behavior of other traffic participants in the vicinity of the special vehicle can be collected immediately.

[0142] Preferably, the transmitted vehicle-to-vehicle communication data additionally or alternatively includes information related to path history. This provides the advantage of, for example, more accurate determination of the location of traffic participants, and also the advantage of being able to collect group data even at times when the traffic participants sending the vehicle-to-vehicle communication data were not yet within the effective range of the traffic participants collecting group data.

[0143] In a preferred method, the vehicle-to-vehicle communication data transmitted at the transmission time point by at least one traffic participant transmitting vehicle-to-vehicle communication data in a common data packet includes current motion data of the traffic participant transmitting the communication data related to the transmission time point, and includes historical motion data of traffic participants transmitting communication data at multiple detection time points that are different from each other and earlier than the transmission time point.

[0144] In a more preferred method, the group data identification device identifies local group data related to multiple different traffic participants sending vehicle-to-vehicle communication data within a preset and / or preset time period, and provides this local group data, especially in the form of common data packets, to the communication device for (especially simultaneously) transmission to an external storage device.

[0145] In a more preferred method, the crowd data identification device analyzes redundant motion data received by the traffic participant transmitting the communication data, particularly multiple vehicle-to-vehicle communication data or data derived therefrom, at substantially the same detection time point, and removes redundant motion data to identify local crowd data. In other words, the crowd data identification device can remove redundant data from the vehicle-to-vehicle communication data it receives. This provides the advantage of keeping the amount of data that should be transmitted to external storage devices as small as possible.

[0146] In a more preferred method, the crowd data identification device determines local crowd data at regular time intervals and provides the local crowd data to a communication device for transmission to an external storage device. Preferably, the communication device (especially the one determining the traffic participants in the crowd data) transmits the local crowd data to the external storage device.

[0147] Preferably, local group data is integrated into global group motion data in an external storage device. Preferably, multiple movement paths for each traffic participant are determined based on the group motion data. Preferably, for each type or category of traffic participant, and preferably for the driving or movement operation performed (e.g., a left turn), an average or statistical movement path is generated, preferably which can be used as a predicted movement path.

[0148] As already mentioned, the advantage of recording convoy data is that data can be obtained from almost any imaginable situation as a convoy of vehicles passes by. It might now be argued that this can be achieved even without Car2X equipment, namely by recording the trajectories of convoy vehicles (e.g., at intersections). This is indeed the case for passenger cars. However, as a passenger car manufacturer employing this approach, it lacks access to data on other road users, i.e., data on cyclists (which are only available to bicycle manufacturers). Similarly, interactions between road users can only be observed under very limited conditions, namely, only if other road users are detected by sensors throughout the operation.

[0149] Preferably, a global group movement data map is generated based on a dataset of multiple transmitted local group movement data. In this map, the corresponding movement data (especially anonymized) is explicitly assigned to each detected traffic participant, and preferably stored according to the corresponding location data of the traffic participants. In particular, this achieves a map representation that provides movement data (and / or movement paths derived therefrom) assigned to multiple traffic participants based on their position or location.

[0150] Preferably, for each detected traffic participant (for which motion data has been transmitted to an external storage device), the movement path of the traffic participant can be determined (based on the motion data explicitly assigned to the traffic participant).

[0151] The present invention also relates to a (especially computer-implemented) method for determining at least one, especially statistical, group prediction parameter of the predicted movement path of a preset category of traffic participants located in a preset traffic area based on a plurality of determined (measured and especially driven) movement paths, the determined movement paths being obtained based on multiple group movement data of multiple traffic participants.

[0152] According to the present invention, multiple group movement data are multiple global group movement data generated and / or collected in multiple traffic areas that are different from each other.

[0153] In other words, at least one, especially a statistical, population prediction parameter is determined based on multiple defined movement paths, which are based on multiple global population movement data generated and / or collected from multiple traffic participants in multiple different traffic areas.

[0154] Preferably, the multiple determined (measured and especially traveled) motion paths can be determined using the methods described above for determining global group motion data.

[0155] Here, multiple determined (measured) movement paths in multiple different traffic areas can be captured from an overhead perspective (bird's-eye view) of the respective traffic areas, for example, using drone photography. Preferably, the object trajectories of moving objects (traffic participants) in the sequence are extracted or determined from the drone photographs, especially drone photograph sequences. Preferably, drone photographs or drone photograph sequences from multiple different traffic areas are used.

[0156] Preferably, multiple different traffic zones individually or in combination share the preferred features and / or characteristics described above within the scope of the method for providing adjustment notifications.

[0157] Preferably, multiple determined (measured) movement paths in multiple different traffic areas are generated or determined using the methods described above for determining local group movement data and / or global group movement data. Here, the local group movement data is preferably generated and / or determined using vehicle-to-vehicle communication data (as described above).

[0158] Preferably, at least one, especially a statistical, group prediction parameter is determined based on multiple (measured) movement paths, which are determined and / or generated by means of multiple different traffic participants of a preset category.

[0159] The proposed method offers the advantage of adapting the functionality to specific situations when using a sufficiently large dataset. For example, a dataset recorded by a camera drone depicting the behavior of cyclists and passenger vehicles at intersections.

[0160] Now, preferably, these data (local group movement data) are enriched in the backend with other data (such as map data), and preferably, influencing factors for estimating parameters and / or warning time points are determined.

[0161] Preferably, when determining at least one estimated parameter, the (measured) movement paths of traffic participants moving at speeds below a preset and / or preset minimum speed (e.g., 2.0 m / s) are not taken into account, and / or are removed from the movement data map. The aim here is to exclude vehicles that stop while turning. Alternatively, these vehicles can be clustered, and specifically grouped into their own clusters (rather than excluding them entirely from the data).

[0162] In addition, when determining the estimated parameters, the (measured) determined motion paths shorter than the preset and / or preset minimum distance (e.g., 5m) can be disregarded, or they can be removed from the (local) motion data dataset or map.

[0163] First, the recorded trajectories of passenger vehicles and cyclists can be clustered based on their direction of movement and actions performed at intersections. Then, for example, it can be investigated whether the cyclist's direction of travel (along the regular direction of travel on the bike lane / in the opposite direction), the presence of passenger vehicles at intersections, and consequently, their impact on movement behavior and prediction can be measured.

[0164] Preferably, the influencing factors calculated using this "big data" method are fed back to the vehicle as a set of parameters. That is, for example, at intersection A in a rural area (where bicycles always travel at high speed), the vehicle can display a warning 7 seconds before passing the intersection, while at intersection B in an urban area, the warning is only displayed 5 seconds in advance. For the parameterized configuration of the vehicle, both intersection-specific parameters and standardized parameters (e.g., for intersection type / size / ...) can be envisioned.

[0165] At least one group prediction parameter can characterize (in particular, traffic participants of a predefined category and / or traffic areas) the predicted movement path.

[0166] In a preferred method, at least one group prediction parameter characterizes the statistical prediction quality of a pre-defined predicted motion path. Here, the group prediction parameter can be the group prediction parameter used to determine the prediction quality, as described above within the scope of the method for providing adjustment notifications. In particular, the group prediction parameter can be a probability density function of the prediction error (preferably for a specific category or type of traffic participant and / or driving operation and / or especially the predicted path) (as detailed above or in the accompanying drawings).

[0167] A significant drawback of methods that create predictive models from drone data or other captured datasets is that the current dataset only shows small fragments, such as an intersection. Therefore, this data is all locally biased, and identifying influencing factors from different datasets requires considerable effort and is largely a matter of luck. Furthermore, much of the dataset is quite small or contains relatively few trajectories, making it difficult to draw statistically significant conclusions.

[0168] The disadvantage of using drone data to determine the dataset is that it requires a large investment (operating drones or survey stations, and labeling the data).

[0169] In a more preferred method, to determine the group prediction parameters, multiple defined movement paths are clustered according to different types of driving maneuvers. This clustering can be based on deceleration and / or acceleration and / or steering angle parameters, the structure and topology of the traffic area being driven (e.g., determined by map data), and / or the degree of danger of traffic conditions detected by traffic participants (e.g., during overtaking maneuvers). This provides the advantage that, for example, other influencing factors can be studied for similar driving maneuvers to determine the predicted movement path as accurately as possible (e.g., based on the type of planned driving maneuver). Furthermore, in application scenarios, the planned driving maneuver can be derived relatively easily from navigation data or map data or prescribed and / or preset route guidance.

[0170] In a more preferred method, in order to determine the group prediction parameters, multiple determined movement paths of multiple traffic participants are statistically evaluated based on at least one, and preferably multiple, influencing parameters.

[0171] Here, one or more influencing parameters are selected from a group of influencing parameters that includes: topological parameters characterizing at least one trajectory of movement in the collision area, particularly the topology of the intersection structure; parameters characterizing the presence of at least one other traffic participant, particularly in the collision area; traffic density; speed parameters characterizing at least one speed prescribed for at least one (involved) traffic participant in the collision area; traffic rule parameters characterizing pre-set traffic rules, particularly right-of-way rules; range parameters characterizing the geometric extent of the collision area and / or the intersection area of ​​the pre-set trajectory of the traffic participant; traffic participant parameters characterizing the level and / or group of the involved traffic participant; cultural parameters characterizing the geographical area, particularly environmental parameters characterizing at least one environmental condition, particularly those affecting visibility; time parameters characterizing the season and / or time of day; weather parameters characterizing the current weather; and combinations thereof.

[0172] Preferably, statistical and / or average and / or typical motion paths are determined based on multiple (measured) motion paths, particularly preset driving operations for traffic participants. It is also conceivable that a preset parameterized configuration of the preset motion path is preset, and the parameters of the parameterized motion path are determined based on multiple (measured) motion paths (e.g., using an optimization algorithm).

[0173] It is conceivable that, when assessing whether a preset influence parameter affects the statistical motion behavior of traffic participants, the standard deviation (or variance) of the measured motion path associated with the statistical motion path is determined, and the influence parameter is evaluated based on the magnitude of the standard deviation.

[0174] Preferably, a pattern recognition method (e.g., KI-based) is used, which identifies the existence patterns of specific influencing parameters (e.g., the aforementioned group of influencing parameters) and specific (determined and / or pre-defined) changes in statistical and / or average motion paths.

[0175] Preferably, regression analysis is used to describe and analyze the relationship between these influencing parameters and (global) group motion data and / or multiple motion paths.

[0176] Preferably, the predicted motion path of the preset traffic participants (e.g., for preset driving operations) is parameterized based on the confirmed influence parameters.

[0177] Preferably, based on the parameterized configuration of the predicted motion path, a motion model for estimating the motion paths of traffic participants is generated according to these influencing parameters.

[0178] The present invention also relates to a machine-readable and, in particular, computer-implementable motion model for predicting the motion path of a traffic participant (preset category, preferably determined by the traffic participant and / or category to be determined) based on the location of the traffic participant, preferably detected by a traffic participant's surrounding environment detection device, wherein the motion model is input with parameters characterizing the location (e.g., location data) and at least one parameter characterizing one (or said) category of the traffic participant as input parameters, wherein the motion model maps the input parameters to output parameters characterizing the motion path to be predicted based on a plurality of model parameters.

[0179] Preferably, multiple model parameters are determined based on global group motion data of multiple traffic participants generated in multiple different traffic areas.

[0180] Preferably, multiple different traffic zones individually or in combination share the preferred features and / or characteristics described above within the scope of the method for providing adjustment notifications.

[0181] Preferably, at least one and more model parameters characterize multiple statistical movement paths in multiple different traffic areas.

[0182] Preferably, the motion model can be generated by the method described above for determining at least one population prediction parameter for predicting motion paths according to a preferred embodiment.

[0183] Preferably, at least one planned and / or predicted driving action can be input as an input parameter to the motion model.

[0184] Preferably, at least one influencing parameter can be input to the motion model, and more preferably, multiple influencing parameters can be input as input parameters. Preferably, multiple influencing parameters (as described above) can be assigned to the motion variables based on the planned and / or predicted driving operation.

[0185] For example, model parameters can be influencing parameters and / or motion data that characterizes the (current) motion state of traffic participants for whom the predicted motion path should be determined.

[0186] Preferably, the motion model can be input with environmental data detected by the traffic participant's surrounding environment detection device and / or surrounding situation data generated by the traffic participant's surrounding situation ascertainment device (as raw data or data derived therefrom), preferably, the motion model determines the values ​​characterizing the influencing parameters from these data.

[0187] The present invention also relates to a warning device for a vehicle, particularly for a traffic participant, for providing an adjustment notification, preferably a warning notification, to at least one traffic participant to be notified, for adjusting at least one driving function to accommodate at least one surrounding traffic participant located in and / or moving in the area surrounding the traffic participant.

[0188] Here, the warning device is provided with and / or will be provided with the expected movement paths of the traffic participants, both for itself and for at least one surrounding traffic participant.

[0189] Here, the warning device is adapted and configured to determine, based on the predicted motion path, at least one notification parameter characterizing the provision of an adjustment notification, preferably characterizing whether and / or at what time an adjustment notification should be provided.

[0190] According to the invention, the warning device is adapted and configured to determine a notification parameter based on at least one, particularly statistical, crowd prediction parameter, which is used to determine the prediction quality of at least one of the predicted movement paths. Preferably, the at least one crowd prediction parameter is determined based on global crowd movement data of multiple traffic participants generated in multiple, distinct traffic areas.

[0191] Preferably, the warning device is adapted, suited, and / or configured to perform the methods described above for providing adjustment notifications, and performs, individually or in combination, all the method steps described above in conjunction with the method. Conversely, the method may be equipped, individually or in combination, with all the features described within the scope of the warning device.

[0192] The present invention also relates to a vehicle, particularly a motor vehicle, comprising the warning device for a vehicle according to one embodiment described above. The vehicle may, in particular, be a (motorized) road vehicle.

[0193] The vehicle can be a motor vehicle, particularly a driver-controlled motor vehicle (“Driveronally”), a semi-autonomous motor vehicle, an autonomous motor vehicle (e.g., autonomy level 3, 4, or 5 (standard SAE J3016)), or an automatically driven motor vehicle. Here, autonomy level 5 indicates a fully autonomous vehicle. Similarly, the vehicle can be an unmanned transportation system. Here, the vehicle can be controlled by a driver or drive autonomously. Furthermore, in addition to road vehicles, the vehicle can also be an air taxi, an airplane, and other mobile vehicles or other vehicle types, such as air, water, or rail vehicles.

[0194] The present invention also relates to a crowd data identification device for traffic participants, particularly for vehicles, especially motor vehicles, for determining, particularly at least local global crowd motion data, and preferably determining (at least partial segments) global crowd motion data, for generating crowd motion data characterizing multiple movement paths of multiple different traffic participants and / or predicting at least partial segments of movement paths of global maps.

[0195] According to the present invention, the crowd data identification device is adapted and configured to determine local crowd movement data based on vehicle-to-vehicle communication data received by the communication device of the crowd data identification device from at least one traffic participant transmitting vehicle-to-vehicle communication data, and to provide the local crowd movement data to the communication device of the crowd data identification device for transmission to an external storage device, preferably to an external back-end server.

[0196] Preferably, the crowd data identification device is adapted, suitable for, and / or configured to perform the method described above for determining local and global crowd motion data, and preferably global crowd motion data, and to perform, individually or in combination, all the method steps described above in conjunction with the method. Conversely, the method may be equipped, individually or in combination, with all the features described within the scope of the crowd data identification device.

[0197] The present invention also relates to a map generation apparatus (and / or a motion path parameter determination apparatus and / or a prediction parameter determination apparatus), which is used to determine at least one, particularly statistical, group prediction parameter of the predicted motion path of a preset category of traffic participants located in a preset traffic area based on multiple determined motion paths of multiple traffic participants.

[0198] According to the present invention, the multiple determined movement paths are multiple global group movement data generated and / or collected in multiple traffic areas that are different from each other.

[0199] Preferably, the map generation device (and / or the motion path parameter determination device and / or the prediction parameter determination device) determines at least one, especially statistical, group prediction parameter based on multiple motion paths, which are determined and / or generated by means of multiple different traffic participants of a preset category.

[0200] Preferably, multiple group prediction parameters (e.g., predicted movement paths and / or group prediction parameters) are determined for multiple traffic areas, and these are associated with their location data as a map representation.

[0201] Preferably, the map generation apparatus (and / or motion path parameter determination apparatus and / or prediction parameter determination apparatus) is adapted for, suitable for, and / or configured to perform the method described above for determining at least one group of prediction parameters for predicting motion paths, and performs, individually or in combination, all the method steps described above in conjunction with the method. Conversely, the method may be equipped, individually or in combination, with all the features described within the scope of the map generation apparatus (and / or motion path parameter determination apparatus and / or prediction parameter determination apparatus).

[0202] The present invention also relates to a computer program or computer program product comprising program means, in particular program code, which represents or encodes at least some and preferably all of the method steps of one of the methods according to the present invention, and preferably one of the preferred embodiments described herein, and is designed to be executed by a processor device.

[0203] The present invention also relates to a data storage device that stores at least one embodiment or a preferred embodiment of a computer program according to the present invention. Attached Figure Description

[0204] Other advantages and implementation methods are illustrated in the accompanying drawings: in: Figure 1 An illustration of the collision zone from road traffic is provided to illustrate the application areas of the preferred embodiments of the invention; Figure 2 Another example of a traffic situation with a collision zone in road traffic is shown to illustrate the advantageous application, particularly according to a preferred embodiment of a warning device for traffic participants according to the invention; Figure 3 This illustrates the process for generating a warning notification to be output to the driver of the vehicle. Figure 4 Three exemplary display contents are shown, which are used to output warning notifications to the driver in order to increase the driver's awareness or attention to cyclists crossing the road; Figure 5 This illustrates another traffic situation involving a vehicle equipped with a warning device according to a preferred embodiment of the invention; Figures 6a to 6c A schematic diagram illustrating the estimated probability is shown; Figures 7a to 7c This demonstrates the impact of choosing a specific prediction model for predicting the movement paths of traffic participants when compared with statistically collected data from the corresponding traffic participants. Figure 8 , Figure 9 An exemplary collision zone in road traffic is shown, in which the movement paths of different traffic participants (here, motor vehicles and cyclists) may intersect; Figure 10 A diagram illustrating the predicted motion path for explaining or visualizing a left-turn scenario is shown. Figure 11 An exemplary histogram showing the time difference between the estimated duration and the actual duration for a (motorized) vehicle until reaching the (preset) collision point; Figure 12 This illustrates another exemplary histogram or kernel density estimation with a Gaussian kernel for the difference between the predicted or estimated TTPIP and the actual TTPIP for a (motor) vehicle (automobile). Figure 13 An exemplary visualization showing the probability of a vehicle intersecting with the movement path of a VRU within a preset time period; Figure 14 Box plots are shown to illustrate the prediction errors in determining the predicted duration of the (motorized) vehicle until it reaches the preset collision point, or in the estimated duration of the (motorized) vehicle until it reaches the preset collision point, in comparison of two different motion models used for vehicle motion. Figure 15 TTPIP for (motorized) vehicle and bicycle riders is shown. pred Instructions to arrive at the same time (i.e., TTPIP) pred,car = TTPIP pred,cyc The probability obtained by (motorized) vehicles and cyclists reaching the PIP within a defined time period ("gap"); Figure 16 Showing TTPIP pred,car Different combinations of ∆TTPIP during the time interval [-4 s, 2 s] (“gap”) T gap A heatmap showing the probability of intersections between the movement paths of motor vehicles and cyclists; Figure 17 The illustration shows a method according to a preferred embodiment of the invention, which is used to determine crowd movement data for generating a global map of crowd movement data and / or movement paths of multiple traffic participants; and Figure 18 A schematic diagram of a vehicle for collecting group motion data is shown. Detailed Implementation

[0205] Figure 1 The diagram illustrates a collision zone 2 (in this intersection area) in road traffic, in which different traffic participants (here referred to as the PKW (passenger car) of the vehicle indicated by reference numeral 10, which approaches the intersection in lane 3 to turn right into lane 5, and other traffic participants 22, 24, 30) meet at the intersection.

[0206] exist Figure 1 The intersection shown is a four-way intersection (four-way intersection) where two straight lanes, 3 and 5 (each with two lanes for opposing directions of travel), converge. One of the intersecting lanes (lane 5) has a zebra crossing on both sides for pedestrians 24 to cross it for traffic management. Reference numeral 7 indicates a zebra crossing located on lane 5, into which vehicles 10 turn right at the intersection.

[0207] exist Figure 1 In the current traffic situation depicted, cyclist 22 is crossing zebra crossing 7, while pedestrian 24 is on the sidewalk in front of zebra crossing 7, on the opposite side of lane 5 relative to vehicle 10. Cyclist 30 is slightly further away from zebra crossing 7, but is riding towards zebra crossing 7 from the same side as the pedestrian.

[0208] For traffic management purposes, a traffic signal facility 20 is provided, which is designed as a networked infrastructure device. This infrastructure device can, for example, exchange V2X messages with networked vehicles 10 and transmit warnings 13 and / or provide ADAS data 13 (ADAS is an abbreviation for "Advanced Driver Assistance System") to the communication device 14 of vehicle 10, which the vehicle's assistance system can process.

[0209] For example, traffic signal facility 20 may be capable of detecting the surrounding environment of a zebra crossing area 7, which is essentially a rectangular loop. Figure 1 This includes zebra crossings, sidewalks, and adjacent areas such as adjacent bicycle lanes. Traffic signal facilities can use object recognition to assess detected surrounding environmental data and preferably report the identified objects (here, cyclist 22 and pedestrian 24) along with their locations to traffic participants with V2X capabilities, such as vehicle 10 (and report to cyclist 30).

[0210] Utilizing Figure 1 The bounding boxes drawn around cyclist 22 and pedestrian 24 indicate that infrastructure device 20 has identified these two traffic participants in the zebra crossing area. Conversely, cyclist 30, who is still outside the detection range of infrastructure device 20, cannot be detected (and identified) by infrastructure device 20.

[0211] The warning device of vehicle 10, particularly processor-based devices, may use object data received by the communication device 14 of vehicle 10 to supplement the surrounding environment data obtained by the onboard sensors of vehicle 1, and perform driver assistance functions based on the surrounding environment data. For example, the driver assistance function may be an "awareness" function, in which, for example, when a (potential) dangerous situation is predicted by the driver assistance function, the driver or user's attention is increased, and preferably the attention is shifted to the (potential) danger point. Figure 1 The “awareness” function presented in the document is in particular a VRU awareness function, where “VRU” is an abbreviation for “Vulnerable Road User” (German: “gefährdeter Verkehrsteilnehmer”).

[0212] "VRU" refers specifically to road users who are at particularly high risk. For example, "VRU" includes pedestrians, cyclists, and users of electric micro-vehicles (such as electric scooters). VRUs are not surrounded by a protective driver's cabin like those in cars or transport vehicles, and therefore typically represent vulnerable road users in road traffic.

[0213] The VRU awareness function (performed by warning device 12) determines, based on the received object data, whether the user or driver of vehicle 10 should be warned by outputting (here graphically represented as a symbolic diagram) warning notification 16, or whether the user or driver should stop with extra caution because the VRU (may) stop along the predicted path of vehicle movement.

[0214] Here, in a preferred embodiment, the warning device 12 according to the invention, when determining whether and / or when to issue a warning notification 16 to the vehicle user or driver, considers the probability of a collision between the vehicle and another traffic participant, or the probability of a dangerous situation arising from insufficient safe distance between the vehicle and another traffic participant. To this end, the warning device 12 can, based on the predicted movement paths determined for its own vehicle 10 and for the other traffic participant, determine whether the two predicted movement paths intersect and / or whether there is sufficient (safe) distance between them, thereby targeting the safe and / or harmless movement of both traffic participants (along the originally planned movement paths).

[0215] exist Figure 1 In the traffic conditions described, the cyclist 20 (identified or detected by infrastructure device 20) is already on the zebra crossing 7. At normal driving speeds, at the time when vehicle 10 turns right into lane 5, the cyclist 20 will have already left the zebra crossing on the opposite side.

[0216] Therefore, if the warning device 12 of vehicle 10 compares the predicted path of movement for vehicle 10 with the predicted path of movement for cyclist 22, it concludes that the two predicted paths will not intersect and that a sufficient (preset) safe distance will be maintained between the two traffic participants 10 and 22. Thus, in this case, the warning parameter determined by the warning device 12 can characterize the assessment result determined by the warning device, i.e., there is no need to output a warning notification, or there is no indication to provide a warning notification (for output to the user).

[0217] The situation may be different for pedestrian 24 (also detected or identified by infrastructure device 20). From the perspective of vehicle 10, the pedestrian is moving in the opposite direction of movement, is on the sidewalk on the side of the roadway opposite vehicle 10, and should step onto zebra crossing 7 to cross the roadway 5.

[0218] Although vehicle 10 is farther from pedestrian 24 than pedestrian 7, it typically moves faster and therefore may meet at the zebra crossing at approximately the same time. Thus, when comparing two corresponding predicted paths, warning device 12 can obtain a warning parameter as an evaluation result, indicating that a warning notification should be output to alert the driver or user of vehicle 10.

[0219] The situation is likely similar for cyclist 30. Although in Figure 1In the snapshot of the traffic situation presented, the cyclist 30 is farther from the potential intersection than pedestrian 24. However, since the cyclist 30 typically travels at a higher speed than the pedestrian, the assessment of the predicted movement paths of the vehicle and the cyclist also indicates (without further intervention) a collision is foreseeable. Therefore, the warning device 12 of the vehicle 10 can also identify a warning parameter that indicates the user or driver of the vehicle 10 should be warned.

[0220] exist Figure 1 In the example given, cyclist 30 was not detected by infrastructure device 20. Given the limited field of vision of the driver and / or onboard sensors or the vehicle-attached ambient detection device of vehicle 10 regarding traffic participants 24 and / or 30, while infrastructure device 20 can transmit object data about pedestrian 24 to the vehicle for ADAS data provision, it cannot (yet) transmit object data about cyclist 30. Therefore, relying solely on ambient data about the VRU on the zebra crossing area 7 provided by infrastructure device 20 may pose a risk that vehicle 10 does not obtain ambient data about cyclist 30 early enough, and therefore cannot, for example, warn the driver or user of vehicle 10 early enough by issuing a warning notification.

[0221] In a preferred embodiment, such as the traffic situation presented herein (in which cyclist 30 is a networked cyclist who transmits V2X communication data, and particularly preferably V2V communication data, to his or her surroundings), the communication device 14 of vehicle 10 may also receive V2X communication data, preferably V2V communication data, from networked traffic participants.

[0222] In this preferred embodiment, the warning device 12 is adapted and configured to evaluate V2X communication data (preferably V2V communication data) received by the communication device 14, and to take into account the V2X communication data (preferably V2V communication data) when determining warning parameters, and / or when determining and / or recalling the predicted movement paths of traffic participants who sent V2X communication data and / or sent V2V communication data.

[0223] When two traffic participants (i.e., here a vehicle and, for example, a pedestrian 24 or a cyclist 30) approach each other too closely (especially a potential collision event) with a (sufficiently high) preset probability (without further intervention), the driver is warned in advance only by issuing a warning notification to the driver, thus achieving the maximum possible benefit to the driver or user of the potentially warned vehicle 10.

[0224] According to a preferred embodiment, the warning device according to the invention is suitable and defined for, when determining warning parameters, especially the warning time point, taking into account the estimated quality of the (corresponding) predicted movement path derived by means of group data (from the convoy).

[0225] A warning notification is provided (to the driver) only when there is a sufficiently high probability that the traffic participants involved are close to each other at the intersection or meet at the intersection.

[0226] In determining such probabilities, special consideration should be given to the corresponding estimated quality or probability of stay of the predicted movement paths involved for the traffic participants.

[0227] Preferably, (especially via a warning device) the warning time point is determined based on the probability of path intersection.

[0228] Figure 2 Another example illustrating a traffic situation with a collision zone in road traffic is provided to illustrate, in particular, the warning device 12 for traffic participants 10 according to the invention. Figure 2 Advantageous application of preferred embodiments (not shown in the text).

[0229] from Figure 2 As can be seen, the direct line of sight of the cyclist 30 from the user of vehicle 10 is blocked by building 4. The cyclist is riding on a bike path that intersects with the vehicle 10's lane and is moving toward the intersection formed by the vehicle's lane and the bike path. In other words, the area where the cyclist 30 is currently located, from the vehicle 10's starting point, is not visible, especially not by the vehicle 10's vehicle-attached ambient environment detection device.

[0230] and Figure 1 The example shown is different; cyclist 30 is not a networked traffic participant.

[0231] Since this spatial design often leads to accident hotspots, an infrastructure device 20 can be installed, arranged such that the VRU (Vehicle Refurbisher Unit) lane, in this case, the bicycle lane intersecting the roadway, can be detected. Reference numerals 20a and 20b here denote the boundaries of the spatial detection range of the infrastructure device 20.

[0232] (Also) cyclists 30 that are not detectable by vehicle 10 are detected by infrastructure device 20, and regarding cyclists 30, infrastructure device 20 sends V2X communication data, which is received by vehicle 10, particularly its communication device, and the V2X communication data (or data derived therefrom) is provided to the vehicle's warning device 12 to determine warning parameters and / or warning notifications. Wireless V2X communication between infrastructure device 20 and vehicle 10... Figure 2 The reference numeral S is used to denote it.

[0233] Figure 3 This diagram illustrates the process for generating a warning notification S4 to be output to the driver of the vehicle. In step S1, based on map matching and motion path prediction, relevant Car2X objects or V2X objects are selected, that is, objects (or traffic participants) I1 and I2 detected via V2X communication.

[0234] Reference numeral I3 indicates an icon illustrating the vehicle bus's use of its own data (such as speed). This leads to the conclusion that the driver is concerned with the cyclist's "cycling awareness" or increased attention (in...). Figure 3 (Referring to S3), which can be considered in further vehicle guidance (S4). S4 indicates the output of awareness notification or warning notification of the graphic here on the central display (shown here), which displays the real-time 3D surrounding environment and is transmitted via smart light in method step S2.

[0235] exist Figure 4 The image shows three exemplary displays that are used to output warning notifications to the driver in order to increase the driver's awareness or attention to cyclists crossing the road.

[0236] In the above diagram, reference numeral 18 indicates the lane being used by the user's vehicle, which the driver can see, for example, through the vehicle's windshield. In a display (e.g., a central display or a (front) display and / or head-up display within the driver's field of vision), a symbolically presented cyclist can be presented as a graphic warning symbol I1 along with a warning triangle IW. Preferably, a direction is output using another graphic symbol R, which can be expected from the driver's perspective. For example, the direction symbol R can be output as an arrow. Additionally or alternatively, a text notification IT can preferably be output to the driver for a more detailed explanation, such as, for example, "Attention, cyclist."

[0237] Figure 4The lower left portion of the diagram illustrates a 3D surrounding environment illustration, which is preferably output to the driver. Reference numeral 17 indicates a display showing a 3D surrounding environment illustration of the vehicle's surroundings in front of the vehicle in the direction of travel.

[0238] Reference numeral A8 indicates the lane currently being traveled by the driver's vehicle. The driver's vehicle itself is also symbolically represented in the 3D surrounding environment illustration, enabling rapid orientation for the driver. Furthermore, the bicycle lane A6 intersecting the current lane A8 and the cyclist A32 on that lane are also shown in the 3D surrounding environment illustration. The cyclist A32 is specifically a potential traffic participant who may collide with the vehicle. Preferably, the 3D surrounding environment illustration only outputs objects necessary to convey to the driver the potential collision zone, the relative position of the vehicle with respect to its current position, and the relative position of potential traffic participants. In this way, all necessary information can be obtained by the driver as quickly as possible.

[0239] The lower right portion of the diagram illustrates another preferred possibility for issuing a warning notification R to the driver / user regarding a potential collision with a cyclist towards the vehicle, using an AR HUD (short for "Augmented Reality-Head-Up-Display"). Reference numeral 8 shows the currently traveling lane as visible to the driver through the windshield. Reference numeral AVL indicates a traffic sign for warning cyclists as visible to the driver / user of the vehicle through the windshield. This portion of the diagram also shows the cyclist not yet visible to the driver. Using the ARHUD, three right-pointing arrow tips R gradually appear as navigation cues for the driver. Preferably, the color of the navigation cues serving as warning notifications to the driver is changed (e.g., changing blue arrow tips to orange (or red) arrow tips), and / or the shape of the navigation cues is changed (e.g., changing the last line to a vertical bar) to further emphasize the danger.

[0240] For example, it is conceivable that multiple warning notifications would be displayed above the vehicle's (current) speed display 19 and / or navigation instructions 19.

[0241] Figure 5 This illustrates another traffic situation in which a warning notification can be output to a vehicle 10 equipped with a warning device 12 according to a preferred embodiment of the invention to warn of a potential collision or falling below a safe distance with a cyclist 32 crossing the lane of the vehicle 10.

[0242] Therefore, the warning device 12 (which, for example, obtains position information of all five cyclists 32 located in the collision zone 2) identifies cyclists 32 whose predicted motion paths intersect with the predicted motion paths of (their own) vehicle 10. Figure 5 In the traffic situation shown, the cyclist is cyclist 32, indicated by reference numeral M.

[0243] Figures 6a to 6c A schematic diagram is shown to illustrate the estimated probability.

[0244] Figure 6a This diagram illustrates traffic conditions from a road with a collision zone, specifically at a T-junction. Reference numeral M8 denotes the lane markings used to define the driving lanes for motorized traffic. Reference numeral M6 indicates the boundary lines of the bicycle lanes. Reference numeral M7 marks the median strips between two opposing lanes of traffic.

[0245] Reference numeral 10 again indicates a vehicle (in this case, a car) acting as a traffic participant. This vehicle has a warning device 12 attached to it. Vehicle 10 is moving in the driving lane towards the T-junction in the direction of arrow FR. Figure 6a In the traffic situation shown, the lane where vehicle 10 is located merges into the straight lane that extends here.

[0246] On the lane facing vehicle 10, there is a bicycle lane, bounded by the bicycle lane boundary M6 on one side and the lane boundary M8 on the other. A cyclist is present on the bicycle lane, schematically represented here as circle 32. The cyclist 32 moves in the direction indicated by the given arrow, i.e., from right to left in the plan view.

[0247] Therefore, the cyclist 32 moves toward the intersection or the merging area of ​​the driving lane, and the vehicle 10 is in that driving lane.

[0248] Reference numeral P10 indicates the predicted movement path of vehicle 10 when it intends to turn left from lane MF (which merges into straight lane GF) into the left lane section of straight lane GF as seen from vehicle 10.

[0249] The ellipses (drawn in dashed lines) indicated by reference numerals 34 and 36 represent the estimated probability that cyclist 32 will remain at two different (future) points in time.

[0250] Both ellipses 34 and 36 intersect with the predicted motion path P10 of vehicle 10, indicating that, given the corresponding motion behaviors of vehicle 10 and cyclist 32, there is a concern or potential concern that vehicle 10 may collide with cyclist 32.

[0251] As in Figure 6a As explained regarding cyclist 32, the medium ellipses 34 and 36 are... Figure 6b It is presented again in China, not only Figure 6a The prediction of the future location of vehicle 10 and cyclist 32 is related to a certain degree of inaccuracy in the forecast.

[0252] Instead of the dwell probability in two-dimensional space (e.g.) Figure 6a As illustrated by the ellipse in the diagram, this is now based on a one-dimensional observation approach. Thus, the probability of dwelling along the predicted line (the movement path of traffic participants) is used, i.e., determined in one-dimensional space. This offers the advantage that determining the one-dimensional dwelling probability is less computationally intensive compared to two-dimensional dwelling probabilities, and still provides a sufficiently accurate model for predicting the relative positions of two traffic participants.

[0253] Regarding cyclist 32, this cyclist has a definite position (indicated by a circle) at the initial time point of observation (see arrow). At later time points, the cyclist's approximate stopping position can be approximated by a Gaussian function GRO. At even later time points (e.g., time t), the cyclist's approximate stopping position can again be described by a stopping probability function GRT, which also essentially follows a Gaussian curve but has a larger variance than the stopping probability function GRO at the earlier time points.

[0254] G10 represents the probability function of vehicle 10's dwell time at time t. This dwell probability function, due to the higher vehicle speed, has a larger variance or bias than the dwell probability function GRT for a cyclist at time t.

[0255] Based on these two probability functions G10 and GRT, for example, it is possible to calculate the probability that vehicle 10 and cyclist 32 will meet at time t (within the collision area).

[0256] This is Figure 6c The following is an explanation. An exemplary curve showing the change in probability is shown here, which gives the relationship between the probability of vehicle 10 meeting cyclist 32 and the time t measured from the starting time point.

[0257] This is based on the following: the two dwell probability functions, namely the dwell probability functions of vehicle 10 and cyclist 32, at time point t.K It has a maximum value at the intersection of its two movement path lines. Here, a movement path line can be understood as a line along which the movement path of a traffic participant extends.

[0258] In other words, for two traffic participants (here, vehicle 10 and the cyclist), they at time t K The probability of both reaching the intersection (at the same time) is the highest.

[0259] Regarding the question of whether a warning notification should be issued to one of the traffic participants, further consideration should be given to not only the precise meeting at the (geometric) intersection of their (geometric) paths, but also the arrival at the intersection within a preset very short time interval, or within a shorter than preset time period (e.g., in -t). s and +t S The two traffic participants have already reached the intersection before (between). Therefore, for example, a sufficient safe distance can be achieved between them when passing through the collision zone (especially by taking the time interval T into account). G (Probability change curve within).

[0260] Figures 7a-7c This demonstrates the impact of choosing a specific prediction model for predicting the movement paths of traffic participants when compared with statistically collected data from the corresponding traffic participants.

[0261] Here, in Figures 7a to 7c The box plots presented show the estimated time to reach the (preset) intersection (TTPIP) along the horizontal axis, based on the motion path prediction model selected for each traffic participant (averaged by intervals (bins)). pred .

[0262] Therefore, the horizontal axis represents the duration of TTPIP (averaged over intervals). pred (in seconds), that is, the duration is determined by the selected path prediction model within such a time period that the traffic participant arrives at the (preset) intersection ("PIP", or "path intersection point") (after calculating the path prediction model).

[0263] Along the vertical axis (y-axis), for the corresponding estimated duration (x-axis), the difference (in seconds) between the estimated duration until reaching PIP and the actual measured duration is presented, i.e., TTPIP. pred -TTPIP true .

[0264] The centerline t in the box M This displays the median of the data. Half of the data is above this value, and the rest is below it.

[0265] The upper distribution end t of the box oP and lower distribution end t uP Display the 75th and 25th quantiles or percentiles respectively.

[0266] The lines extending from the housing are called "whiskers" (also known as "antennas") and represent the expected changes in data. The whiskers extend beyond the housing at 1.5 times the interquartile range, that is, 1.5 times the upper and lower distribution ends of the housing. When the data does not extend to the end of the whisker, the whisker reaches the minimum and maximum data values.

[0267] Here Figures 7a to 7c In the diagram, point PA represents a so-called outlier, which is located above or below the end of the whisker.

[0268] Figure 7a This illustrates such an evaluation of the cornering PKW under the assumption that the speed remains constant, namely: in: Time to reach the intersection : Distance from the path intersection point (distance from the PIP "Path Intersection Point", i.e., Entfernung zum Pfadkreuzungspunkt) Initial velocity (Anfangsgeschwindigkeit) In comparison, Figure 7b It shows the assumption of acceleration The same diagram for a cornering PKW vehicle under constant conditions: In comparison Figure 7a and Figure 7b It is evident from the following situation that, for turning vehicles, when using a constant speed model for calculation, deceleration causes the vehicle to actually arrive at the intersection before the estimated time point (TTPIP). pred With the measured TTPIP true Compared to a negative value, that is, E TTPIP = TTPIP pred – TTPIP true ).

[0269] On the contrary, Figure 7bIn the case shown, using a model with the assumption of constant acceleration (which is negative in this case), the variances are similar, but the median t... M Or percentile t oP t uP It is located near the zero line.

[0270] However, for the cyclist's driving maneuver (crossing the intersection straight without turning, typically without braking), the relatively constant speed along the straight path allows for good prediction of the cyclist's performance using appropriate models, such as... Figure 7c As presented in the text.

[0271] Figure 8 and Figure 9 This illustrates an exemplary collision zone in road traffic, where the movement paths of different traffic participants (here, motor vehicles and cyclists) may intersect.

[0272] Here, the two collision areas are shown in a bird's-eye view, i.e., from an overhead perspective, for example, by a drone preferably taking (camera) photographs from this perspective to detect the collision areas and / or the ground area surrounding them. This type of (camera) photograph is preferably used to acquire and / or (especially statistically) determine the movement paths of different traffic participants. From the acquired or measured movement paths of the different traffic participants, (especially statistically) uncertainty parameters regarding prediction accuracy can be determined.

[0273] exist Figure 8 In the middle, refer to the reference numeral P. R It represents the (geometric) path of a cyclist's movement, along which the (measured, statistical, and / or average) path of movement extends; that is, along which the cyclist moves.

[0274] Reference number P CL A (geometric) path of motion for a left-turning (motorized) vehicle (e.g., a car), along which the (measured, statistical, and / or average) path of motion extends, i.e., along which the left-turning vehicle moves before and / or while and / or after turning left.

[0275] Reference number P C This indicates the (expected and / or foreseeable) point of collision, which in this case is determined by the path of motion P of the left-turning vehicle. CL The intersection point with the path line PR of the cyclist crossing the intersection is given.

[0276] exist Figure 9 In the middle, reference numeral P R P r1P r2 P r3 P r4 P r5 This represents the (geometric) path of a cyclist who (starting from different lanes and lane sides) crosses the intersection (P) in a straight line. r5 Import P r2 ;P r3 ;P r1 Import P R ), or turn left at the intersection (P r4 Import P r2 Here, the corresponding motion path is shown as the direction (at least in a portion) of the collision area (or surrounding area), along which the cyclist's motion path (statistically and / or on average and / or usually) extends, or along which the cyclist moves.

[0277] Reference number P CR1 P CL2 P C3 P C6 Accordingly, this represents the (geometric) path of a (motorized) vehicle (e.g., a car) moving (in a traffic lane), the (motorized) vehicle in its different operations (i.e., for example, in straight travel (P) C3 ;P C6 ), turn right (P CL2 ) and turn left (P CL2 The movement path and ruts of the vehicle extend along the movement path line.

[0278] In the traffic guidance pre-set at the intersection shown here, there are multiple collision points between (motor) vehicles and cyclists (or between cyclists, or between (motor) vehicles).

[0279] Figure 10 This illustrates or visualizes the motion paths P of multiple mappings of (motorized) vehicles (e.g., passenger cars or automobiles) in a left-turn lane, for example. G The predicted and / or preset motion path P for the left turn scenario P The motion path shown in the diagram is mapped or matched to the (parameterized) motion path.

[0280] In the parameterization (selected here), the travel path P of the (motorized) vehicle. G With predicted motion path P P The maximum deviation is 5m.

[0281] Figure 11An exemplary histogram is shown, which illustrates the time difference (TTPIP) between the predicted or estimated duration until reaching the (preset) collision point for a (motor) vehicle (or automobile) and the actual required duration. pred -TTPIP true (Based on the preset predicted duration until reaching the collision point). Here, the H-axis indicates the frequency of the estimated time error, and the horizontal axis indicates the time difference (TTPIP). pred - TTPIP true (Measured in seconds).

[0282] Figure 12 This shows the difference between the predicted or estimated Time to Path Intersection Point (TTPIP) for the (motor) vehicle (automobile) 3.0 seconds before the estimated path intersection point ("Zeitdauer biszumPfadschnittpunkt") and the actual TTPIP (TTPIP is an abbreviation for "Time to Path Intersection Point," i.e., "Zeitdauer biszumPfadschnittpunkt" in German). pred - TTPIP true Another exemplary histogram, or kernel density estimation with a Gaussian kernel (the solid line indicated by the reference numeral y), is shown here. The probability density P is plotted on the ordinate.

[0283] Here, a histogram visualizes the prediction error or estimation error. Preferably, Gaussian kernel density estimation y can identify and / or estimate the probability density function of the prediction error distribution y or estimation error distribution y for traffic participant type and path. f pred It is abbreviated as "PDF" (Probability Density Function).

[0284] Figure 13 An exemplary visualization showing the probability of a vehicle intersecting the motion path of a VRU during a time interval of [-0.5s, 0.5s] (“Gap”).

[0285] Plot the probability density P on the left vertical axis.

[0286] Plot the parameter " on the horizontal axis" E pred + ∆TTPIP pred ",in E pred = TTPIP pred - TTPIP true and ∆TTPIP pred = TTPIP pred,car - TTPIP pred,cyc .

[0287] Therefore, ∆TTPIP pred Give the difference between the estimated time to reach the PIP (Plan of Instantaneous Exploration) for a (motorized) vehicle (e.g., a car) and the estimated time for a bicycle (or VRU). If these two estimated times are TTPIP... pred,car and TTPIP pred,cyc They are exactly the same size, i.e., ∆TTPIP pred =0, which means that the two traffic participants arrive at PIP at the same time.

[0288] “y1: Car” represents the probability density of (motorized) vehicles or automobiles. “y1: VRU (∆TTPIP)” pred =0 s)” and “y1: VRU (∆TTPIP pred = 3 s)” represents the probability density distribution for VRU (the cyclist in this case). P X .

[0289] The probability density function of the prediction error distribution for (motor) vehicles f pred,car (“Probability Density Function”, abbreviated as “PDF”), in Figure 13 The curve "y1: Car" illustrates the probability density function of the prediction error distribution for cyclists. f pred,VRU ,exist Figure 13 The curve “y1: VRU (∆TTPIP)” passes through the curve. pred = 0 s)” indicates that, starting from this point, the probability of a vehicle intersecting with the VRU's movement path during the time interval [-0.5s, 0.5s] (“Gap”) can be determined using the following relationship: Plot on the right-hand vertical axis P X .

[0290] shaded area dt Displays the differential for (motor) vehicles or automobiles, and uses " t±T gap, 1,2 The area indicated by "" shows the integration region within the display, with integration boundaries showing the probability density for VRUs (cyclists). t+T gap,1 and T +T gap,2 The region represented indicates that the external integral contains a differential. dt The factors of the function to be integrated. The curve y2 is shown... t For ∆TTPIP pred = 0 s integral result.

[0291] In order to target other ∆TTPIP pred Determining the probability of a non-zero value (i.e., two traffic participants arriving at the PIP at different times based on their estimated duration until arrival) requires, for example, shifting the PDF for cyclists by the difference in their estimated arrival times at the PIP, such as... Figure 13 The finer line in the middle is "y1: VRU (∆TTPIP)". pred As explained in "= 3 s)".

[0292] For each moved PDF, it can be done as described above. Px (car), VRU, ∆TTPIP pred , T gap,1 , T gap,2 As given, the probability metric is calculated to obtain the probability for different TTPIPs. pred A probability set of combinations.

[0293] The resulting probability is achieved by estimating the positive predictive value (PPV), also known as Relevanz (correlation) or positiverprädiktiver Wert (positive predictive value), based on the number of correct positive (nTP) and false positive (nFP) awareness notifications between (motor) vehicles (e.g., cars) and VRUs for that specific scenario. This includes the estimation method used, the specific and expected "gap" (time period) that must be below the gap to trigger an awareness notification. This enables the determination of whether the expected "consciousness cue" (attention stimulus) or total attention arousal occurs at a preset time point (TTPIP). pred,car And the preset "gap" (time period) ∆TTPIP pred This is important for the driver.

[0294] Furthermore, through information regarding specific TTPIPs pred,car ∆TTPIP pred Calculate the value P X The average value can be achieved, as estimated below for... T gap Application of parameterized positive predictive values: .

[0295] Figure 14 The TTPIP (Traffic-Temperature-Induced Probability) model, used to illustrate the predicted or estimated duration of a (left-turning) (motorized) vehicle until it reaches a predetermined collision point, is presented in comparison to two different motion models for vehicle motion. pred (TTPIP is an abbreviation for "Time to Path Intersection Point", which is "Zeitdauerbis zum Pfadschnittpunkt" in German.) Box plot of the estimated error.

[0296] The reference numeral "CV" indicates a box plot for the selected motion model used to determine the predicted or estimated motion path, along which the (motorized) vehicle moves at a constant speed.

[0297] Reference numeral "CA" represents a box plot (shown as a thick dashed line) for the selected motion model used to determine the predicted or estimated motion path, on which the (motorized) vehicle moves with constant acceleration.

[0298] Plot the estimated duration TTPIP along the horizontal axis until it reaches PIP. pred, car .

[0299] Draw TTPIP along the vertical axis err, car = TTPIP pred, car – TTPIP true, car (In seconds) The difference between the predicted or estimated TTPIP and the actual (measured) TTPIP. Here, the measured TTPIP is determined based on the motion path, which is determined by drone photographs of the collision area (at this intersection).

[0300] TTPIP err, car A positive value corresponds to the following situation, in which the estimated time TTPIP until reaching PIP ("Path Intersection Point"). pred, carGreater than the measured time until PIP (TTPIP) true, car Therefore, (motor) vehicles (e.g., cars) arrive at the PIP earlier than expected. Conversely, for negative values, (motor) vehicles (e.g., cars) arrive at the PIP later than expected.

[0301] The chart shows the uncertainty of the prediction / estimate, visualized by boxes representing the 0.25 and 0.75 percentiles, and increasing with TTPIP. pred, car It grows as its size increases.

[0302] Furthermore, the chart shows that for constant velocity motion models (CV), the absolute value of the median prediction error increases with the prediction duration, up to TTPIP. pred, car The time was initially 5 seconds, but increased to approximately 1.8 seconds. In the drone data set, 50% of the vehicles arrived at the intersection point 0.56 to 2.66 seconds later than predicted.

[0303] In comparison, for the constant acceleration motion model CA, the absolute median value of the prediction error or forecast error is significantly lower, especially for the 5s TTPIP. pred,car The value is -0.46s. Since the driver brakes while turning (which is ignored in the constant speed acceleration model), the constant acceleration model is more accurate for predicting this scenario.

[0304] For cyclists traveling in a straight line across the collision zone, no significant difference was found between the two models (not presented here).

[0305] For the two figures below, the constant acceleration model is applied not only to left-turning car drivers but also to cyclists crossing the intersection in a straight line (see...). Figure 8 (The traffic situation or collision area shown in the image).

[0306] Figure 15 TTPIP is provided for (motorized) vehicle (or car) and bicycle riders. pred Instructions to arrive at the same time (i.e., TTPIP) pred,car = TTPIP pred,cyc In the case of a (motorized) vehicle (or car) and a cyclist, the probability of reaching the PIP (path intersection point), i.e., the point of collision, within a defined time period ("gap") is measured by the supermodel. For example, this probability can be measured using the following formula: in, f pred,car andf pred,VRU It is the probability density function (PDF) corresponding to the prediction error distribution for traffic participants and routes (see also...) Figure 12 ).

[0307] P X Alternatively, the y-axis can provide information about (motorized) vehicles (or cars) over a specific time period. T gap The probability that the path intersects with the cyclist's path.

[0308] Plot the estimated duration (TTPIP) up to the PIP along the horizontal axis. pred, car .

[0309] The chart shows two points: the probability of obtaining TTPIP varies. pred The increase decreases, and the critical time gap is used for attentional stimulation or awareness cue. T gap The longer the TTPIP, the higher the probability for a given TTPIP.

[0310] For example, for a time interval of -2s to 2s (“gap”), the probability of two traffic participants crossing the PIP (collision point) within the defined “gap” or time interval, assuming they arrive at the PIP simultaneously, is calculated in TTPIP. pred The value is 0.76 at 5 seconds.

[0311] For an asymmetric time interval of -4s to 2s (which takes into account an acceptable 4s interval but allows for a 2s passage after the cyclist), the probability is 0.93.

[0312] Figure 16 The probabilities obtained for time intervals of [-4s, 2s] (“gap”) are also visualized, but different combinations of TTPIPs for (motorized) vehicles (or cars) and cyclists are also considered.

[0313] Here, Figure 16 A diagram for TTPIP is shown. pred,car A heatmap showing the probability of different combinations of ∆TTPIP intersecting the movement paths of a (motorized) vehicle (or car) and a cyclist within a defined time period ("gap"). The grayscale axis shows the (motorized) vehicle within the time period [-4s, 2s] ("gap"). Tgap The probability that the path of the cyclist intersects with the path of the cyclist.

[0314] TTPIP for vehicles pred For shorter periods, the probability drops sharply at the defined time interval (“gap”). In TTPIP... pred Over a longer period of time, the probability decreases more slowly.

[0315] This result allows for the following conclusion: the probability of (motorized) vehicles and cyclists crossing within a defined time period ("gap") is correlated with the predicted or estimated time (TTPIP) until the (motorized) vehicle reaches the PIP. pred,car It is related to the difference between the cyclist's predicted arrival time and the actual arrival time.

[0316] For example, when TTPIP pred,car = 5s and ∆TTPIPpred is 1s (the cyclist is expected to arrive at the PIP before the (motorized) vehicle or car), the probability obtained is 0.84, and when ∆TTPIP pred At -1s (the expected arrival time of the cyclist at the point of impact), the probability is 0.94. For 2s and -2s, the probabilities are 0.68 and 0.81, respectively. If the cyclist arrives at the point of impact (PIP) 4s later than the vehicle (or car), the probability drops to only 0.32. Averaging the probabilities over the time interval from -4s to 2s yields a positive predictive value (PPV) of 0.77 for this parameterization.

[0317] Of course, the probability is lower when the expected (motorized) vehicle and cyclist arrive at the PIP at the same time. When the time of the attention cue is 0, the cyclist might still arrive at most 4 seconds later than predicted, and is still within the defined period for a successful attention cue. If it starts with, for example, a 3-second offset, then a false positive result is obtained when the cyclist arrives 1 second later than predicted. Therefore, a lower overall probability is obtained.

[0318] Figure 17 The present invention describes a method according to a preferred embodiment, which is used to identify group motion data for generating a global map of group motion data and / or motion paths characterizing the motions of multiple (different) traffic participants.

[0319] Figure 17 A segment of the road network in the traffic area is shown, with multiple roads or lanes that are interconnected here, exemplarily indicated by reference numerals S1, S2, S3, and S4.

[0320] Reference numeral 1 indicates a traffic participant whose (group movement) data was determined, and who is moving straight along road S1, i.e., without turning, in the traffic situation presented here, from... Figure 17 The road section S1 in the lower left of the middle is in Figure 17 It moves in the direction of the upper right of the middle road S1.

[0321] Reference numeral 110 indicates another traffic participant, a motor vehicle that starts from road S4 and moves along the road in the direction of road S1, turns left into road S1 at the junction of road S4 and then continues along road S1.

[0322] Reference numeral 112 indicates another traffic participant, in this motor vehicle, which starts from road S3 and turns into road S2, then turns right from road S2 into road S4, and moves in the direction of road S1.

[0323] Two traffic participants 110 and 112 transmit V2V communication data during their travel, and vehicle 1 can receive the V2V communication data by means of communication device 114. Figure 18 A schematic structure of a vehicle 1 with a communication device 114 is shown, by means of which V2V communication data or V2X communication data can be received.

[0324] The communication connections generated by sending and receiving V2V communication data are illustrated by lines indicated by reference numerals "V2V-110" and "V2V-112".

[0325] Here, different traffic participants are represented by different symbols. The same symbol indicates the location of the corresponding traffic participant at different points in time. Here, the symbol gives the corresponding location of vehicle 1 and other traffic participants 110 and 112, as can be determined by ascertaining its own location data (vehicle 1) and by means of (especially) V2V communication data received by vehicle 1, which in particular characterizes the current location of the traffic participant that sent the V2V communication data.

[0326] V2V communication data is transmitted at specific points in time. Therefore, for each of these points in time, as long as the person or vehicle 1 is within the receiving range of the corresponding V2V communication, the person or vehicle 1 specifically obtains information about the position, speed, acceleration, heading, etc. of the traffic participants 110, 112 that sent the V2V communication data.

[0327] The arrow marked "V2V-110" indicates that vehicle 1 can still receive V2V communication data from vehicle 110 at the following time points: at this time point, due to the distance between vehicle 1 and 110 and their different driving directions, vehicle 1 can no longer be detected by the vehicle itself through onboard dedicated sensing devices (such as cameras and / or lidar sensors and / or radar sensors).

[0328] The situation is similar for vehicle 112, which sends V2V communication data. Figure 17 The exemplary communication connection “V2V-112” illustrates that vehicle 1 has received V2V communication data from traffic participant 112 at a given time. This means that at that time, traffic participant 112 is outside the driver's field of vision, outside the detection range of vehicle 1's own ambient environment detection device (i.e., the vehicle-mounted dedicated sensor), and not even on road S4 merging into road S1. Furthermore, besides the communication connection represented by “V2V-112”, multiple structures, such as houses, may exist along the direct connection between vehicle 1 and vehicle 112.

[0329] like Figure 18 As illustrated in the schematic structure, vehicle 1 further includes a group data identification device 116, which identifies local group (motion) data based on vehicle-to-vehicle communication data (V2V communication data) received by communication device 114, and provides this local group (motion) data, in particular, to communication device 115 of vehicle 1 (illustratively shown in the schematic structure). Figure 18 (as shown in the reference numeral 118) for transmission to an external (especially cloud-based) storage device 120, preferably to an external back-end server and / or a cloud-based server.

[0330] Subsequently, the map generation device 130 can call up the group data or group movement data stored on the storage device 120, and use the group data or group movement data to generate a global map about the movement data or movement path, especially according to the type of traffic participants.

[0331] Reference numeral 100 denotes a system for generating a global map of group motion data and / or group motion paths, and preferably predicted motion paths. Here, system 100 includes, in particular, at least one vehicle 1 that collects motion data and / or receives vehicle-to-vehicle communication data (V2V communication data), and preferably includes multiple vehicles of this type. Preferably, these vehicles each have a group data identification device that transmits the group motion data to a storage device 120, in particular, externally, based on the received V2V communication data (by means of a communication device). Preferably, the system also includes an external storage device 120, and particularly preferably includes a map generation device.

[0332] The applicant reserves the right to claim that all features disclosed in the application documents are essential to the present invention, provided that such features, individually or in combination, are novel relative to the prior art. Furthermore, it is noted that features that may be advantageous in themselves are also depicted in the various figures. Those skilled in the art will readily recognize that a particular feature described in a figure may be advantageous even without employing other features from that figure. Moreover, those skilled in the art will recognize that advantages also arise from combining multiple features shown in a single figure or different figures.

[0333] Reference number list 1 vehicle 2. Collision Zone Lanes 3 and 5 4. Buildings 7. Zebra crossing area with zebra crossing 8. The current driving lane 10 vehicles 12 Warning devices 13 Warning Messages / ADAS Data Provision 14 communication devices 16. Warning notification, this VRU is aware of 17. Display screen shows 18 The current driving lane 19. Speed ​​specifications and / or navigation specifications 20. (Networked) infrastructure installations, such as traffic signal facilities. 22 Identified cyclists 24. Pedestrians identified 30 cyclists 32 cyclists 34 and 36: Illustrations of the estimated probability of future stay areas S V2X-communication Steps S1-S4 R direction symbol I1, IW warning symbols IT text notification IR graphic direction symbols A10 displays its own vehicle The driving lane displayed by A8 A6 Crossed bicycle belts A32 Cyclist Traffic signs displayed by AVL M mark M8 lane edge lines M7 center band M6 bicycle with boundary lines P10 Predicted motion path of vehicle 10 GF straight driving lane MF merges into the straight driving lane GF. G10, GRT, and GRO Dwell Probability Distribution t oP t uP percentile t M median t oA t uA Whiskers P A outliers P R The path of a straight-line cyclist P CL Path of a car turning left P C Collision area, especially the point of impact P r1 P r2 P r3 P r4 P r5 Cyclist's movement path P CR1 P CL2 P C3 P C6 (Motor) vehicle's path of motion P P Visualization of predicted motion paths P G Visualization of the driving motion path H frequency Y Prediction Error Distribution Traffic participants 110 and 112 sending V2V communication data 114 Communication devices 115 Communication devices 116 Group Data Identification Device S1-S4 roads 130 Map Generating Device 120 External storage devices

Claims

1. A method for providing an adjustment notification, preferably a warning notification, to at least one self-notified traffic participant (10) via a warning device (12) for adjusting at least one driving function to adapt to at least one surrounding traffic participant (32) located in and / or moving in the surrounding area (2) of the self-notified traffic participant, and for predicting movement paths (P) for the self-notified traffic participant and for the at least one surrounding traffic participant (32). CL P R The warning device provides and / or is provided with the warning device, and the movement of the traffic participants (10, 32) can be anticipated along the movement path, wherein, The warning device (12) is based on the predicted motion path (P) CL P R Identify at least one notification parameter characterizing the provision of the adjustment notification, wherein the notification parameter preferably characterizes whether an adjustment notification (I1, A32) should be provided and / or at what time the adjustment notification (I1, A32) should be provided. Its features are, The warning device (12) identifies the notification parameter based on at least one population prediction parameter, which is statistically significant, and is used to determine the predicted motion path (P). CL P R The estimated quality of at least one of the following, wherein the at least one group estimated parameter is determined based on global group motion data of multiple traffic participants, which is generated in multiple traffic areas (2) that are different from each other.

2. The method according to claim 1, characterized in that, The warning device (12) determines whether an adjustment notification should be provided based on the population prediction parameter according to a preset and / or preset minimum estimated probability parameter.

3. The method according to at least one of the preceding claims, characterized in that, The warning device (12) presets and / or can preset safety parameters characterizing the minimum safe distance between itself and the surrounding traffic participants, and based on the safety parameters, the warning device (12) determines, according to the predicted movement path of itself and at least one group prediction parameter for determining the estimated quality of the predicted movement path of itself and the predicted movement path of the surrounding traffic participants, and according to the predicted movement path of the surrounding traffic participants and at least one group prediction parameter for determining the estimated quality of the predicted movement path, how likely it is that the itself and the surrounding traffic participants will approach each other to a distance closer than the preset and / or preset distance.

4. The method according to at least one of the preceding claims, characterized in that, Determine at least one predicted time point for the traffic participant and / or for the surrounding traffic participants to reach the collision area, especially the collision point, and determine the probability that not only the traffic participant but also the surrounding traffic participants will reach the collision area within the preset time period before or after the predicted time point, based on a preset time period.

5. The method according to at least one of the preceding claims, characterized in that, The surrounding situation determination device of the traffic participant and / or the surrounding traffic participants collects and / or determines the surrounding situation data, and the surrounding situation data represents at least one influencing parameter. The influencing parameters are selected from a group of influencing parameters, namely, the group of influencing parameters includes: topological parameters characterizing at least one trajectory of movement in the collision area, especially the topology of the intersection structure; parameters characterizing the presence of at least one other traffic participant, especially in the collision area; traffic density; speed parameters characterizing at least one speed prescribed for at least one (involved) traffic participant in the collision area; traffic rule parameters characterizing preset traffic rules, especially right-of-way rules; range parameters characterizing the geometric extent of the collision area and / or the intersection area of ​​the preset trajectory of the traffic participant; traffic participant parameters characterizing the level and / or group of the involved traffic participant; cultural parameters characterizing the geographical area, especially environmental parameters characterizing at least one environmental condition that can affect visibility; temporal parameters characterizing the season and / or time of day; weather parameters characterizing the current weather, etc.; and combinations thereof. The at least one group prediction parameter is determined based on the at least one influence parameter.

6. The method according to at least one of the preceding claims, characterized in that, The warning device (12) determines the predicted movement path of the at least one surrounding traffic participant (32) based on vehicle-to-everything communication data (V2X communication data) and preferably vehicle-to-vehicle communication data (V2V communication data) obtained by the communication device (14) and sent by the at least one surrounding traffic participant (32).

7. A method for determining at least local group movement data and preferably global group movement data of at least a portion of a segment by means of at least one group data determining device (116) of at least one traffic participant, particularly a vehicle (1), for generating a global map of the group movement data and preferably predicted movement paths, representing multiple movement paths of multiple different traffic participants, Its features are, The group data identification device (116) identifies local group movement data based on vehicle-to-vehicle communication data (V2V-110, V2V-112) sent by at least one traffic participant (110, 112) that is received by the communication device (114) of the at least one traffic participant (1) who identified the group data, and provides the local group movement data to the communication device (115) of the traffic participant (1) who identified the group data, in particular, for transmission to an external storage device (120), preferably to an external back-end server.

8. The method according to the preceding claims, characterized in that, The vehicle-to-vehicle communication data transmitted by at least one traffic participant in a common data packet at the transmission time point includes the current motion data of the traffic participant transmitting the communication data with respect to the transmission time point, and the historical motion data of the traffic participants transmitting the communication data at multiple different detection time points with respect to the earlier transmission time point.

9. The method according to at least one of the preceding two claims, characterized in that, The group data identification device identifies local group data of multiple different traffic participants sending vehicle-to-vehicle communication data within a preset and / or preset time period, and provides the local group data, especially in the form of common data packets, to the communication device for transmission to the external storage device.

10. The method according to at least one of the preceding three claims, characterized in that, The group data identification device analyzes the redundant motion data of the traffic participants who sent the communication data at substantially the same detection time point. This includes data received by the traffic participants, especially data sent by just one traffic participant who sent the communication data, or data derived therefrom, and removes the redundant motion data to identify the local group data.

11. The method according to at least one of the preceding four claims, characterized in that, The group data identification device identifies local group data at regular time intervals and provides the local group data to the communication device for transmission to the external storage device.

12. A method for determining, particularly statistical, at least one group prediction parameter, based on multiple identified movement paths of traffic participants of a predetermined category located in a predetermined traffic area (2), wherein the multiple identified movement paths are obtained based on multiple group movement data of multiple traffic participants. Its features are, The multiple group motion data are multiple global group motion data generated and / or collected in multiple traffic areas (2) that are different from each other.

13. The method according to the preceding claims, characterized in that, The group prediction parameters characterize the prediction quality of the preset predicted motion path.

14. The method according to the preceding claims, characterized in that, To determine the group's estimated parameters, the identified motion paths are clustered according to different types of driving operations.

15. The method according to any one of the preceding three claims, characterized in that, To determine the group prediction parameters, multiple identified movement paths of the multiple traffic participants are statistically evaluated based on at least one, and preferably multiple, influencing parameters, wherein the influencing parameters or the multiple influencing parameters are selected from a group of influencing parameters that includes: topological parameters characterizing at least one movement trajectory in the collision area, particularly the topology of the intersection structure; parameters characterizing the presence of at least one other traffic participant, particularly in the collision area; traffic density; speed parameters characterizing at least one speed prescribed for at least one (involved) traffic participant in the collision area; traffic rule parameters characterizing preset traffic rules, particularly right-of-way rules; range parameters characterizing the geometric extent of the collision area and / or the intersection area of ​​the preset movement trajectory of the traffic participants; traffic participant parameters characterizing the level and / or group of the involved traffic participants; cultural parameters characterizing the geographical area, particularly environmental parameters characterizing at least one environmental condition, particularly those affecting visibility; temporal parameters characterizing the season and / or time of day; weather parameters characterizing the current weather; and combinations thereof.

16. A machine-readable and, in particular, computer-implementable motion model, said motion model being used to predict, based on the location of a traffic participant, a preset, preferably ascertained by the traffic participant and / or of a category to be ascertained, preferably detected by a surrounding environment detection device of the traffic participant, wherein, The motion model is supplied with at least one parameter representing the location and the category of the traffic participant as input parameters, wherein the motion model maps the input parameters to output parameters representing the motion path to be predicted based on multiple model parameters, characterized in that the multiple model parameters are determined based on global group motion data of multiple traffic participants generated in multiple different traffic areas.

17. A warning device (12) for a traffic participant (10), particularly for a vehicle, the warning device being used to provide an adjustment notification, preferably a warning notification, to at least one traffic participant (10) to be notified, for adjusting at least one driving function to accommodate at least one surrounding traffic participant (32) located in and / or moving in the surrounding area (2) of the traffic participant, wherein, The predicted movement paths (P) for the traffic participant itself and for the at least one surrounding traffic participant (32) are respectively. CL P R The warning device (12) is provided and / or provided to the warning device, which anticipates the movement of the traffic participant (10, 32) along the movement path, wherein the warning device (12) is adapted and configured to, based on the predicted movement path (P) CL P R The process involves identifying at least one notification parameter characterizing the provision of the adjustment notification, wherein the notification parameter preferably characterizes whether an adjustment notification (I1, A32) should be provided and / or at what time the adjustment notification (I1, A32) should be provided. Its features are, The warning device (12) is adapted and configured to identify the notification parameter based on at least one population prediction parameter, particularly statistical, used to determine the predicted motion path (P). CL P R The prediction quality of at least one of the following, wherein the at least one group prediction parameter is determined based on global group motion data of multiple traffic participants, which is generated in multiple traffic areas (2) that are different from each other.

18. A vehicle (10), particularly a motor vehicle, comprising a warning device (12) according to the preceding claims.

19. A group data identification device (116) for traffic participants, particularly for vehicles, especially motor vehicles, said group data identification device for identifying, particularly at least (partially) local and global group movement data, and preferably global group movement data, for generating a global map of at least partial sections of the group movement data and / or predicted movement paths, representing multiple movement paths of multiple different traffic participants, characterized in that, The group data identification device (116) is adapted and configured to identify local group movement data based on vehicle-to-vehicle communication data (V2V-110, V2V-112) sent by at least one traffic participant (110, 112) who sends vehicle-to-vehicle communication data, received by the communication device (114) of the group data identification device (116), and to provide the local group movement data to the communication device (115) of the group data identification device (116) for transmission to an external storage device (120), preferably to an external back-end server.

20. A map generation apparatus (130) for determining, in particular statistically, at least one group prediction parameter based on multiple identified movement paths of traffic participants of a predetermined category located in a predetermined traffic area (2), the identified movement paths being obtained based on multiple group movement data of multiple traffic participants. Its features are, The multiple group motion data are global group motion data generated and / or collected in multiple traffic areas (2) that are different from each other.

Images