Automatic warning in relation to a vehicle sidewall collision zone

Abstract

The invention discloses a method comprising the following, carried out by a computer unit of a vehicle: receiving sensor data from sensors recording a traffic environment of the vehicle; running a continuously updated environment model, wherein the environment model comprises, for a reference sidewall of the vehicle: a warning zone located behind the vehicle, and a collision zone extending, in a forward direction of the vehicle, from a rear collision boundary located to the side of the reference sidewall to a front collision boundary located to the side of the reference sidewall; in response to detection of a traffic object within the traffic environment, generating a model object describing the traffic object in the environment model; performing a warning routine comprising: in response to detection that the model object overlaps the warning zone, obtaining a probable trajectory of the model object in the environment model based on the sensor data; checking whether a predefined warning condition is satisfied, wherein the satisfied warning condition comprises an overlap between the model object following the probable trajectory and the collision zone; performing a warning response if the warning condition is satisfied.

Description

2024PF01922 1 AUTOMATIC WARNING REGARDING A VEHICLE SIDE WALL COLLISION ZONE AREA OF TECHNOLOGY

[0001] The invention relates to a computer-based warning system for a potentially imminent collision between a vehicle and a traffic object approaching from behind in the area of ​​a side wall of the vehicle. STATE OF THE ART

[0002] Vehicles, especially motor vehicles such as cars, trucks, and transport vehicles, often have areas of the surrounding traffic environment that are difficult to see. These areas can be obscured by body parts, trailer attachments, or similar objects. If such an area cannot be seen even with technical aids, it is referred to as a vehicle's blind spot. If a traffic object approaches a vehicle from a blind spot, a dangerous situation is more likely to arise. However, traffic objects approaching from outside a blind spot can also pose an increased risk of collision with a vehicle, for example, if the attention of the driver, another person in the vehicle, or a computer unit controlling the vehicle (other than the vehicle's computer unit) is not directed towards the approaching traffic object at that moment.

[0003] Another source of danger is rigid or movable side panel elements of a vehicle that protrude from the vehicle's profile, such as exterior mirrors, doors, or fuel filler flaps. These can increase the likelihood of collision with passing traffic. In particular, the movement of such elements can be unpredictable for other road users, so extra caution may be required when operating movable elements on the exterior of a vehicle. A well-known example is a so-called "dooring" accident, in which a cyclist, who approaches a stationary vehicle from behind unnoticed (e.g., due to the absence of engine noise), collides with a vehicle door that is being opened just as the cyclist reaches the vehicle, without any chance to avoid the collision.

[0004] Modern motor vehicles are increasingly equipped with sensors such as cameras, radar, lidar and / or ultrasonic sensors, which enable the detection of 2024PF01922 2 The traffic situation in the vehicle's vicinity is being monitored. Road infrastructure is also increasingly being equipped with sensors that enable vehicle-independent monitoring of traffic conditions. Based on the sensor data generated in this way, vehicle collision warning systems can be developed, and their use is also gradually increasing. To generate warnings in the specific situation of a road user approaching a vehicle from behind, initial systems are under development. These systems are based, for example, on a sensor-monitored zone diagonally behind the vehicle and determine the risk of a collision with a side panel of the vehicle based on the sensor data recorded in the monitored zone. SUMMARY

[0005] It is an object of the invention to provide an improved method that enables computer-based warning of a potentially imminent collision between a vehicle and a traffic object approaching from behind in the area of ​​a side wall of the vehicle, as well as a correspondingly configured computer unit and a corresponding computer program product. The objects underlying the invention are achieved by the features of the independent claims.

[0006] One aspect relates to a procedure, exhibiting, by a computer unit of a vehicle: Receiving sensor data from sensors that detect the vehicle's traffic environment; Operating an environment model that is continuously updated based on sensor data, where the environment model has a reference side wall of the vehicle: a predefined warning zone located behind the vehicle and a predefined collision zone extending in a predefined forward direction of the vehicle from a rear collision boundary located laterally to the reference side wall to a front collision boundary located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating a model object that describes the traffic object in the environment model; Performing a warning routine that includes: In response to a detection that the model object overlaps with the warning zone, obtaining a predicted trajectory of the model object in the environment model based on sensor data; 2024PF01922 3 Check whether a predefined warning condition is met, where the met warning condition shows an overlap of the model object following the expected trajectory with the collision zone; Perform a predefined warning response when the warning condition is met.

[0007] Another aspect relates to a computer unit for a vehicle, wherein the computer unit has a processor and a memory functionally connected to the processor, the memory storing program instructions which, when executed by the processor, cause the computer unit to perform a procedure which has: Receiving sensor data from sensors that detect the vehicle's traffic environment; Operating an environment model that is continuously updated based on sensor data, where the environment model has a reference side wall of the vehicle: a predefined warning zone located behind the vehicle, and a predefined collision zone extending in a predefined forward direction of the vehicle from a rear collision boundary located laterally to the reference side wall to a front collision boundary located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating a model object that describes the traffic object in the environment model; Performing a warning routine that includes: In response to a detection that the model object overlaps with the warning zone, obtaining a predicted trajectory of the model object in the environment model based on the sensor data; Check whether a predefined warning condition is met, where the met warning condition shows an overlap of the model object following the expected trajectory with the collision zone; Perform a predefined warning response when the warning condition is met.

[0008] Another aspect relates to a computer program product, in particular a computer-readable medium, wherein the computer program product carries computer-executable code for execution by a processor of a computer unit of a vehicle, wherein the execution of the instructions causes the processor to perform a procedure which has: 2024PF01922 4 Receiving sensor data from sensors that detect the vehicle's traffic environment; Operating an environment model that is continuously updated based on sensor data, where the environment model has a reference side wall of the vehicle: a predefined warning zone located behind the vehicle and a predefined collision zone extending in a predefined forward direction of the vehicle from a rear collision boundary located laterally to the reference side wall to a front collision boundary located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating a model object that describes the traffic object in the environment model; Performing a warning routine that includes: In response to a detection that the model object overlaps with the warning zone, obtaining a predicted trajectory of the model object in the environment model based on the sensor data; Check whether a predefined warning condition is met, where the met warning condition shows an overlap of the model object following the expected trajectory with the collision zone; Perform a predefined warning response when the warning condition is met.

[0009] It is understood that one or more of the embodiments disclosed herein may be combined with one another, as long as the embodiments do not exclude each other. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The following examples are explained in more detail using the drawings. They show: Fig. 1 shows a block diagram of a vehicle equipped with a computer unit; Fig. 2 shows a flowchart with steps of a procedure that enables computer-based warning of a potentially imminent collision between a vehicle and a traffic object approaching from behind in the area of ​​a side wall of the vehicle; Fig. 3 shows a diagram of an environment model describing a vehicle with warning zones and collision zones; and 2024PF01922 5 Figs. 4-7 each show a diagram of an environment model describing a traffic object with a likely trajectory leading into the collision zone. DETAILED DESCRIPTION

[0011] The use of sensor data for computer-based detection of traffic situations where a collision could occur between a road object approaching a vehicle from behind and a protruding or moving element of the vehicle's side panel holds great potential. By skillfully utilizing the available sensor data, it appears possible to achieve a low level of true-negative and false-positive detection rates for this type of collision, enabling, for example, widespread market launch.

[0012] The method described herein is intended for execution by a computer unit of a vehicle. For example, the vehicle could be a road vehicle, in particular an automobile such as a passenger car, a van, a truck, a bus, commercial vehicle, etc. Specifically, vehicles with structural elements that project from a side wall (e.g., exterior mirrors) or can be moved into a position projecting from a side wall (e.g., doors or flaps on fuel filler necks or charging ports) could be considered. Regardless of the examples mentioned, the method could also be applicable to other vehicles, vehicle types, etc.

[0013] A computer unit could be any unit installed on board the vehicle capable of (e.g., digital) data processing. Typically, a computer unit might include a processor and memory, which could be integrated into the processor and / or implemented as a separate component. Steps of the process could be in the form of program instructions that the processor can execute. Such program instructions could, for example, be stored in a non-volatile portion of memory and, if necessary, loaded into a non-volatile portion of memory before the process is executed. Alternatively or additionally, steps of the process could be implemented by a fixed logic circuit (e.g., on a semiconductor / circuit board).

[0014] The computer unit could also have one or more interfaces through which it can receive information provided by sensors. Such sensors could be located, for example, on board the vehicle and / or outside the vehicle. One of the computer unit's interfaces could, for example, be an input through which the computer unit directly receives a sensor signal provided by a sensor unit, encoding sensor data. Another possible interface could, for example, be a wired or wireless interface through which the computer unit receives sensor data decoded from a sensor signal, which is then processed, for example, by appropriate processing logic and / or a 2024PF01922 6 further computer unit. In particular, sensor data from onboard sensors could be transmitted directly to the computer unit and / or from sensors located outside the vehicle (e.g., onboard other vehicles and / or as part of traffic monitoring systems) could be transmitted to the computer system via further computer systems (e.g., a traffic information system or vehicle computer units configured for vehicle-to-vehicle communication).

[0015] Furthermore, it can be assumed that the vehicle is located in a traffic environment containing static and / or moving traffic objects (e.g., streetlights, walls, other road users, especially people and / or vehicles). Among the aforementioned sensors could be those configured to acquire information describing the traffic environment by receiving signals from it. Without limiting the above, the sensors could include: optical sensors, such as those operating in the visual spectral range and / or infrared sensors, especially cameras, lidar; ultrasonic sensors; radar sensors; etc. The computer unit could receive sensor data-encoding signals from at least one sensor configured to detect the traffic environment via an interface as described above.

[0016] The computer unit executing the procedure could operate an environmental model (hereinafter also referred to as "the model") that describes traffic objects within the traffic environment. These traffic objects could, for example, be recognized based on sensor data generated by the sensors that detect the traffic environment. For this purpose, the sensors and / or associated evaluation logic could, for example, detect positions, colors, shapes, geometries, velocity components, and / or other properties of the traffic environment accessible to the respective measurement methods of the sensors. This data would then be fed into a known or future object recognition method to identify individual traffic objects within the traffic environment.

[0017] In the environment model, these detected traffic objects could then be represented by corresponding model objects. The model could, for example, describe geometric and / or mechanical relationships between the vehicle and the detected traffic objects at a specific time, including properties such as positions, orientations, velocity components, etc., of the model objects relative to the vehicle. The model is preferably two-dimensional to enable the procedure described here, as well as any other routines implemented by the computer unit, to be executed with less data processing effort compared to higher-dimensional models. In particular, the model could be a coordinate model that assigns one or more coordinate tuples to a detected traffic object; an object model that represents a detected model by a geometric shape; or a vector model that assigns a 2024PF01922 7 The model assigns one or more vectors to a detected traffic object, which may extend beyond a coordinate representation; or a combination of such models. The information assigned to a detected traffic object in the environment model is referred to herein as the model object. The model object could be supplemented by further information that could be detected from subsequent and / or other sensor data and / or queried from other information sources, such as computer units of other road users or a traffic information system.

[0018] The environmental model could be continuously updated based on sensor data to track changes such as position or speed changes of traffic objects relative to the vehicle and / or other variable properties of the traffic objects. In particular, the environmental model could be updated cyclically, for example, based on sensor measurement cycles with a predefined cycle duration. Further information could be derived from repeated detections of traffic objects, such as confidence information like the number of repeated detections of an object or the probability that the detected traffic object corresponds to a real traffic object.

[0019] It can further be assumed that the vehicle has one or more (e.g., two) side walls. Vehicle side walls typically run essentially parallel to a longitudinal axis of the vehicle or parallel to the forward direction assumed at a steering angle of 0° and typically have side wall elements that project from the side wall (e.g., Exterior mirrors) or can be moved into a configuration projecting from the side wall (e.g., doors or flaps to cover a fuel filler neck or charging port). One purpose of the method described herein could be to detect potential collisions between traffic objects in the traffic environment (e.g., other vehicles, pedestrians, obstacles) and side wall elements that project from the side wall or could be moved into a configuration projecting from the side wall as early as possible and to warn of such a detected potential collision.

[0020] For this purpose, the environment model could have a warning zone and a collision zone for one or more of the side walls, which could be defined specifically for the associated side wall, referred to herein as the reference side wall. For example, if the vehicle has two side walls, the first of the side walls could be the reference side wall for a first warning zone and a first collision zone, which could be positioned and dimensioned in the model specifically for the first side wall. Similarly, the second side wall of the vehicle could be the reference side wall for a second warning zone and a second collision zone, which could be positioned and dimensioned in the model specifically for the second side wall. However, the warning zone could also be non-side-wall specific, i.e., the 2024PF01922 8 The model could provide a uniform, identical warning zone for all side walls of the vehicle. Preferably, all warning zones and collision zones are located outside the vehicle. Conceptually, the collision zone could be an area within which, in the presence of a traffic object, a collision could occur between the traffic object and a side wall element that protrudes from the side wall or can be configured to protrude from the side wall. The corresponding warning zone could accordingly be designed as a sensor-detected area within which a traffic object approaching the collision zone is detected based on sensor data relating to the traffic object located within the warning zone.

[0021] The warning zone could generally be located behind the vehicle. As understood here, behind the vehicle is any point that has a coordinate on a forward-pointing coordinate axis ("the longitudinal axis") that is located in the negative forward direction away from a coordinate that a rear corner point of the vehicle, situated between a side wall and a rear of the vehicle, has on the longitudinal axis. Points located behind the vehicle therefore also include points that are spaced from the vehicle in a transverse direction perpendicular to the longitudinal axis. The collision zone could be located laterally outside the reference side wall, thus extending, viewed in the forward direction, from a rear collision boundary located laterally to the reference side wall to a front collision boundary located laterally to the reference side wall.

[0022] Unless otherwise specified, the following description is limited to the warning zone and the collision zone with reference to a specific reference side panel of the vehicle. Unless otherwise specified, analogous explanations also apply to all other warning and collision zones with reference to other respective reference side panels of the vehicle. This includes the case where only a single warning zone is defined for the entire vehicle, which serves to trigger warnings for all side panels of the vehicle.

[0023] The procedure could include a warning routine that uses the conditions described above to trigger a warning response in the event of a collision risk emanating from a traffic object located in the warning zone. The warning routine could involve checking the model (e.g., regularly) to see if any of the detected model objects overlap with one of the warning zones defined in the model. This could occur, for example, if at least one point associated with the traffic object lies on an edge of the warning zone or within the warning zone. If a traffic object registered in the environment model overlaps with the warning zone, the warning routine could proceed to check whether this traffic object (hereinafter referred to as "the traffic object") could pose a collision risk within the warning zone.

[0024] For this purpose, detecting the overlap of the traffic object with the warning zone could trigger the generation of a trajectory that the traffic object is expected to follow relative to the vehicle. Determining such a trajectory could be based on a (further) evaluation of sensor data related to the traffic object. This could involve, among other things, geometric and / or mechanical parameters that could be determined or derived from the sensor data. Such parameters could be directly related to the traffic object (e.g., Information regarding position, spatial orientation, speed (component), acceleration (component), etc.) or indirectly related to the traffic object (e.g., the course of a road marking / a lane; position, orientation, geometry, speed, acceleration and / or other properties of other traffic objects with which the traffic object could interact).

[0025] Obtaining the predicted trajectory could involve determining the predicted trajectory by the computer unit itself, or receiving the predicted trajectory from another unit specializing in determining predicted trajectories based on sensor data.Similarly, the determination of the expected trajectory by the computer unit itself could be implemented by program instructions that are part of the program instructions implementing the procedure; or the program instructions implementing the procedure could involve requesting and receiving the expected trajectory from a preceding or concurrent process that the computer unit executes independently of the present procedure in order to provide expected trajectories of detected traffic objects for the environment model; or retrieving the expected trajectory from a memory in which said preceding or concurrent process has stored the expected trajectory in order to make it available to other processes executed by the computer unit, such as the program instructions implementing the procedure.

[0026] Whether a traffic object could pose a future collision hazard within the collision zone could be determined by the computer unit by evaluating a predefined warning condition. In addition to other optional supplementary or partial conditions, as described herein in a non-exhaustive and non-restrictive manner, the check to determine whether the warning condition is met could at least include a check to see if the traffic object enters the warning zone if it follows the previously determined predicted trajectory.

[0027] For this purpose, for example, the model object, a copy, or a part of the model object could be moved along the expected trajectory. For example, a reference point of the traffic object could be placed on support points located on the expected trajectory. In this hypothetical 2024PF01922 10 The displacement (extrapolation) of the model object could optionally also include an anticipated change in the orientation of the model object within the environment model, resulting from tracking the trajectory. If the displaced model object overlaps with the collision zone at one of the control points (i.e., there is at least one point of the model object that lies on an edge of the collision zone or within the collision zone), this could be considered a fulfilled warning condition or at least a fulfilled sub-condition of the warning condition. In the latter case, the warning condition could be formulated, for example, as a logical combination (e.g., using AND, OR, and / or NOT operators) of the individual sub-conditions.

[0028] Additional or alternative methods could also be used to check for overlap by shifting the model object along the expected trajectory. For example, one could attach the expected trajectory to a point on the traffic object and check whether the attached trajectory intersects an edge of the collision zone.

[0029] If the warning condition is met, the warning routine could initiate one or more predefined warning responses. A warning response could be designed to be perceptible to an entity located in the vehicle and / or the traffic environment (e.g., a person inside the vehicle and / or a person in the traffic environment; a computer unit other than the computer unit controlling the vehicle's movement; or a computer unit controlling the movement of the traffic object). Many possible warning responses are conceivable, including, but not limited to, those mentioned herein. The type of warning response(s) executed could be independent of, or specific to, the particular traffic situation detected by the environmental model or other properties, sizes, parameters, etc., determined from the sensor data.For example, the type of warning response triggered could depend on the determined time to collision within the warning zone (hereinafter abbreviated as TTC for "time to collision"), the type, position, speed, etc. of the traffic object relative to the vehicle, the presence (type, number, etc.) of other traffic objects in the traffic environment, the available space for evasive maneuvers, the presence or number of people in the vehicle, the time of day, weather conditions in the traffic environment, etc. In particular, but not limited to the above, the warning response could be carried out by means of a function of the vehicle, e.g., by means of a human-machine interface, especially a screen and / or a loudspeaker; a signaling device such as a turn signal (e.g., a hazard warning light).turn signal), a brake light, a rear fog light, a horn of the vehicle and / or another traffic object that has its own 2024PF01922 11. It has a computer unit that is communicatively connected to the vehicle's computer unit.

[0030] The method described herein possesses several features that, individually or in combination, could offer advantages. For example, the method involves operating an environmental model within which geometric relationships between the collision zone and a traffic object detected in the warning zone could be analyzed. In particular, the environmental model could allow for the prediction of a future state of the traffic environment and the representation of various traffic objects by model objects with individual geometries that approximate the actual shapes of the traffic objects. This could enable a more realistic detection of a potential future overlap of the traffic object with the warning zone.

[0031] A further advantage could be that the method distinguishes between a warning zone, in which a potentially collision-prone traffic object can be detected by sensors, and a collision zone, in which the potential collision between the detected traffic object and a side wall element protruding from the reference side wall could occur. The separately defined collision zone could allow for a more flexible and precise definition of the area adjacent to the vehicle, where collisions can actually occur, than would be possible, for example, with an analysis of the direction of relative movement between the vehicle and the traffic object. Furthermore, the collision zone could enable a more precise geometric overlay with the model object following the expected trajectory, thus allowing a more reliable prediction of a potential collision.

[0032] Using a prediction of the traffic object's likely trajectory could also be advantageous, as it would allow for a more flexible assessment of the traffic object's future behavior. For example, additional indicators beyond the traffic object's direction of movement relative to the vehicle, such as one or more acceleration components, could be used to more accurately estimate the geometric shape of the likely trajectory. Furthermore, the likely trajectory could be determined to match a recognized traffic situation by, for example, detecting and taking into account lane following, lane changes, or obstacle avoidance by the traffic object.

[0033] Overall, the method could therefore enable a more adaptable and reliable detection of whether a collision is imminent between a traffic object behind the vehicle and a side wall element protruding from the reference side wall. 2024PF01922 12

[0034] In one example, the fulfilled warning condition also has at least one of the following fulfilled additional conditions: - A traffic object reaches or exceeds a predetermined minimum speed. This could prevent excessively frequent triggering of the warning response by filtering out traffic objects that, although within the warning zone, are traveling below the minimum speed and are therefore so slow that they are likely to avoid a potential collision or be noticed in time by a person on board the vehicle even without the warning response. This would ensure that people who perceive a triggered warning response take it seriously and do not assume it is another false alarm due to an overly frequent warning in the past. For example, the speed of the traffic object could be...A speed component of the traffic object in a current longitudinal direction of the traffic object or in the direction of the vehicle, or a current total speed of the traffic object, may be included in the warning condition. - The expected time until a collision between the traffic object and the vehicle within the collision zone (remaining collision time, time to collision, TTC) reaches or falls below a predefined maximum duration. This could prevent the warning response from being triggered too early, thus giving the traffic object and / or the vehicle more time to change its potential collision course with the warning zone. Similar to the above, this method could also prevent overly frequent warnings for slow-moving traffic objects. - The probability of a traffic object's existence reaches or exceeds a predefined minimum probability. This would allow traffic objects whose existence cannot be proven with insufficient certainty based on sensor data to be filtered out, as these could potentially be false positives that do not pose a collision risk. In this way, the probability of false alarms could be reduced. - A certain amount of the vehicle's forward speed reaches or falls below a predetermined maximum speed, preferably while the vehicle is stationary. This could suppress the warning response in cases where the vehicle is traveling at such a speed that a change in the configuration of a moving side wall element into a collision-hazardous position within the collision zone (for example, the opening of a vehicle door) can be considered unlikely. Furthermore, this could take into account that when the vehicle is traveling forward, traffic objects have a lower relative speed to the vehicle. The vehicle might have such a long collision time that it could be sufficient to perform an evasive maneuver in such cases.

[0035] In one example, the warning routine is only executed if one of the additional conditions is true. This could reduce the data processing capacity of the computer unit and potentially other data processing components, as computationally intensive processes such as obtaining the expected trajectory and checking the warning condition could be eliminated in these cases.

[0036] In one example, traffic object detection involves recognizing the object type of the traffic object, with the model object being created with a geometric shape predefined for that specific object type. In this way, the environment model could provide a more realistic representation of the actual geometry of the detected traffic objects through the corresponding model objects. This could lead to more reliable collision prediction with a lower false positive rate and / or a lower true negative rate. For example, different object types could be a car, a bus, a two-wheeler, a truck, a vehicle with a trailer, etc., which could be represented in the model by model objects of different shapes characteristic of each object type. These could then be scaled, for example, to the dimensions of the individual traffic object determined from the sensor data.

[0037] In one example, checking whether the model object following the expected trajectory overlaps with the collision zone involves determining a reduced set of reference points representing the model object and is limited to checking whether any of the reference points enter the collision zone when the model object follows the expected trajectory. In this way, the complexity of overlap detection could be reduced from a two-dimensional problem (comparing the positions of points on two surfaces) or a one-dimensional problem (comparing the positions of points on the boundaries of two surfaces) to a lower-dimensional, preferably zero-dimensional, problem (comparing the positions of individual reference points with the position of the collision zone). This could reduce the data processing capacity required for the check.

[0038] In one example, at least one of the following is true: - In the event that the specified object type is a vehicle, the predefined geometric shape is a polygon, specifically a rectangle, and the reference points are selected from the outer vertices of the polygon. Using a polygon could have the advantage that the check to determine whether the model object following the trajectory overlaps with the collision zone can be evaluated with lower algorithmic complexity than is possible, for example, for geometric shapes containing nonlinear components such as circular segments. The outer corner points of automobiles, two-wheelers, or other vehicles, represented in the model by the outer corner points of polygons, might have a higher probability of coming into contact with a point on the vehicle in the event of a collision within the collision zone, compared to other points, even if these points do not necessarily coincide with the point of the traffic object that collides with the vehicle first. By restricting the reference points to the outer corner points of the polygon, an effective reduction of the number of points to be checked for overlap with the warning zone or the collision zone could be achieved. Furthermore, representing the vehicle as a polygon, especially a rectangle, could allow for a comparatively realistic representation of its geometry in the model, thus enabling a particularly reliable overlap check for this type of object. Many options are available when choosing the geometric shape of the polygon. If the vehicle is a car, it could, for example, be represented in the model by a rectangle whose vertices approximately correspond to the vertices of the car. If the vehicle is a two-wheeler (e.g., a bicycle, motorcycle, scooter, e-scooter, etc.), it could also be represented by a rectangle or by another quadrilateral (e.g., rhombus- or diamond-shaped), where, for example, one vertex corresponds to a front point of a front wheel and two other vertices each correspond to an endpoint of a handlebar of the two-wheeler. However, numerous other polygons are conceivable, such as a polygon with more than four vertices that represents the vehicle's outer contour with greater detail. For a car, the polygon could, for example, have additional protrusions to approximate the appearance of side mirrors.A two-wheeler could even be represented, for example, by a cross-shaped polygon having two pairs of opposing protrusions that are aligned relative to each other at an angle other than 0°, in particular 90°. - In the event that the specific object type is a pedestrian, the set of reference points contains exactly one reference point, and the specified geometric shape encloses this exact one reference point, where the exact one reference point is, in particular, a geometric center or centroid of the specified geometric shape. A pedestrian could, for example, be distinguished from vehicles by its comparatively small dimensions (in cross-section parallel to a ground plane of the traffic environment) and / or its own speed. Reducing the set to exactly one reference point could enable a particularly efficient overlap check for this object type, since for each position checked, only the exact one reference point would need to be checked for overlap with the warning or collision zone. The restriction to exactly one reference point could thus be...2024PF01922 15 The reference point, as well as the fact that the geometric shape encloses exactly one reference point, takes into account the special geometry of a pedestrian, for whom lateral and longitudinal directions, as well as extrema such as corners or endpoints, would be difficult to define unambiguously. On the other hand, multiple reference points of a pedestrian would inevitably be close to each other, so the increased data processing effort of an independent overlap check of these multiple reference points would only with a comparatively low probability result in a significant difference in the overlap check result. Thus, an efficient and reliable overlap check could also be achieved for this type of object in this way.

[0039] In one example, the warning routine additionally includes the selection of a reduced subset of reference points. This subset is limited to reference points for which entry into the collision zone is most likely compared to the other reference points in the set. Specifically, reference points with the highest probability of entering the collision zone are those located on the side of the model object facing the vehicle and / or the collision zone. Restricting the selection to a subset of reference points with a comparatively high collision risk could further increase the efficiency of the overlap checks by, for example, excluding reference points facing away from or further away from the vehicle.The selection of reference points located on one of the sides of the model object facing the collision zone and / or the vehicle could, for example, be based on a determined orientation and / or direction of movement of the traffic object and, for instance, the position of known outer edges of the model object relative to the vehicle. This could reduce the data processing effort required for selecting the subset by, for example, eliminating the need to compare the distances of the reference points to a nearest point on the vehicle or the collision zone. For example, the reference points for the subset of a car could be the two corners of the car on the side facing the vehicle or the two front corners. In another example, the reference points for the subset of a two-wheeler could be a front endpoint of the longitudinal axis and an endpoint of the transverse axis facing the vehicle (e.g., the front of the front wheel).handlebar) of the two-wheeler.

[0040] In one example, checking whether one of the reference points enters the collision zone when the model object follows the expected trajectory involves attaching the trajectory to one of the reference points and checking whether the attached trajectory intersects an edge of the collision zone. This could simplify overlap checking by reducing the problem of determining an overlap. The process of identifying the traffic object with the collision zone is reduced to calculating an intersection point or checking for the existence of an intersection point between the attached trajectory and the edge of the collision zone. This could reduce the data processing capacity required for the overlap check.

[0041] In one example, the warning response is selected from a group that includes: emitting a warning signal inside the vehicle; securing movable side wall elements against movement into the warning zone; and emitting a warning signal into the surrounding traffic. One or more of these warning responses could contribute to preventing the detected potential collision. For example, emitting a warning signal inside the vehicle could prompt occupants to focus their attention on the traffic object in the warning zone and / or prevent side wall elements from moving into the collision zone for the duration of the warning signal. The warning signal could be emitted, for example, via a loudspeaker, a screen, and / or an electromechanically adjustable visual or haptic signaling device inside the vehicle.Securing movable side wall elements against movement into the collision zone could minimize the number of structures in the collision zone with which the traffic object could collide. These movable side wall elements could include, for example, doors, fuel filler flaps, and / or exterior mirrors of the vehicle, which could be secured by locking, retracting, interrupting their control, etc. Preferably, the securing mechanism is automatically released once the warning response is complete. A warning signal could be emitted into the surrounding traffic environment, for example, via signaling devices installed on the exterior walls / body of the vehicle, preferably at the rear. These devices could include visual light signals such as turn signals, brake lights, rear fog lights, license plate lights, reversing lights, etc., and / or audible signals such as an external loudspeaker.This could alert entities in the traffic environment, especially a person or a computer unit controlling the movement of the traffic object, to the risk of collision in the collision zone and cause them, for example, to perform an evasive maneuver that bypasses the collision zone.

[0042] A warning response could, for example, also involve transmitting a warning signal to a computer unit other than the main computer unit, which controls the vehicle's movement and could be configured to adjust the vehicle's movement control in response to receiving the warning signal in such a way as to reduce the probability of a collision with the traffic object in the warning zone, preferably if the current traffic situation allows such an adjustment without further increasing the potential for danger. 2024PF01922 17

[0043] In one example, at least one of the following is true: The warning zone adjoins the reference side wall and extends forward from a rear warning boundary located behind the vehicle to a front warning boundary located laterally to the reference side wall, with the rear warning boundary being at least 5 meters, preferably at least 14 meters, away from the front warning boundary. The front and rear warning boundaries could be defined as the outermost points of the warning zone in the forward direction, or analogously as maximum and minimum coordinates on a longitudinal axis of the vehicle, or as half-lines perpendicular to the longitudinal axis that touch the warning zone at most at one edge of the warning zone. Accordingly, the specified distances between the front and rear warning boundaries could be measured parallel to the forward direction or longitudinal axis of the vehicle.A distance of at least 5 meters could ensure that, from the initial detection of the traffic object in the warning zone, there is still enough time to generate the warning response early enough before a potential collision, allowing an entity receiving or perceiving the warning response to react in a collision-avoidance manner. A distance of at least 14 meters could ensure that the aforementioned duration between initial detection and the collision-avoidance response is at least 1 second when the traffic object approaches the vehicle at 50 kilometers per hour (km / h). - The warning zone and / or collision zone extends perpendicular to the forward direction over a distance that is at least equal to the distance between the reference sidewall and an outermost point the vehicle would have when all movable elements of the reference sidewall are moved as far away from the reference sidewall as possible. This could ensure that the check for overlap of the model object with the collision zone covers all possible collision scenarios involving sidewall elements spaced away from the reference sidewall. - The reference sidewall is the only sidewall of the vehicle adjacent to the warning zone. This could increase the selectivity of the procedure for potential collisions on the same side of the vehicle facing the reference sidewall, since a traffic object approaching the collision zone specific to the reference sidewall from the rear of the vehicle or from a side of the vehicle facing away from the reference sidewall would necessarily have to drive around the rear of the vehicle and from there significantly change its direction into the collision zone. This might be unlikely for an unintentional collision at higher speeds due to the inertia of the traffic object. At lower speeds, e.g.In the case of a pedestrian, there might still be enough time, in the case of a warning zone specifically defined for the aforementioned reference side wall, to initiate the warning reaction in time if the traffic object spontaneously changes direction towards the warning zone.2024PF01922 18. to spend. In this way, the proportion of unnecessary warning reactions could be reduced. - The vehicle is an automobile, with the rear collision boundary adjacent to a rear bumper and the front collision boundary adjacent to a front axle. This could minimize the forward-direction extension of the collision zone and thus reduce the proportion of unnecessary warning reactions due to an overly generously defined warning zone. - The warning zone lies entirely to the side of the reference sidewall. In particular, in this case, the warning zone, measured perpendicular to a central longitudinal axis of the vehicle, could be located further from the longitudinal axis than any point, some point, or at least one point on the reference sidewall. Similar to the case discussed above, where the reference sidewall is the only sidewall of the vehicle to which the warning zone borders, this could increase the selectivity of the method for potential collisions on the same side of the vehicle facing the reference sidewall, since a traffic object approaching the collision zone specific to the reference sidewall from the rear of the vehicle would necessarily have to drive around the rear of the vehicle and from there significantly change its direction into the collision zone.This might be unlikely for an unintentional collision at higher speeds due to the inertia of the vehicle. At lower speeds, for example with a pedestrian, there might still be enough time for the warning signal to be triggered in the event of a sudden change of direction by the vehicle towards the warning zone, even if the warning zone is located entirely to the side of the reference wall. In this way, the number of unnecessary warning signals could be reduced.

[0044] In one example, in response to the detection that the model object following the expected trajectory no longer overlaps the collision zone, the warning response is only terminated after a predefined holding time has elapsed and / or when the model object has left a predefined hysteresis zone surrounding the warning zone. This could suppress high-frequency switching between activating and deactivating the warning response in borderline cases where the model object following the expected trajectory just barely overlaps the collision zone. It could also make the warning response more robust against unforeseen course changes of the traffic object, where the traffic object, which initially moved towards the collision zone and later changed course away from the collision zone, subsequently changes course again towards the collision zone at a later time.If the warning response is activated at a later time, the remaining time for a collision-avoidance response to the warning might be too short. 2024PF01922 19.

[0045] In one example, generating the model object involves determining at least one of the following object properties based on sensor data: the current position of the traffic object; the current orientation of the traffic object; the current direction of movement of the traffic object; the length and / or width of the traffic object; the current velocity component of the traffic object; the number of measurement cycles in which the traffic object was detected and / or tracked; the probability of the traffic object existing; and the object type of the traffic object. Checking whether the warning condition is met is additionally performed based on at least one of these specific object properties. This could reduce the false-positive and / or true-negative detection rates of the fulfilled warning condition, thus increasing the reliability of the warning condition check.

[0046] In one example, obtaining the expected trajectory involves selecting a predefined trajectory geometry based on sensor data and then determining the expected trajectory based on that selected geometry. This could enable a more realistic representation of the current traffic situation in the environment model. For example, a set of predefined trajectory geometries could include a straight line, a curve, an S-shape, etc. From this predefined set, a trajectory could then be determined based on, for example, a speed component and / or acceleration component of the traffic object determined from the sensor data, a detected lane path, a detected position, speed, acceleration, orientation, etc., of an obstacle and / or other traffic objects in the traffic environment, a detected position, function, and / or current status of a traffic signal such as a traffic sign, a traffic light, etc.A trajectory geometry best suited to the identified traffic situation is selected and adapted to the measured speeds, positions, etc.

[0047] In one example, the collision zone is composed of collision sectors, with each collision sector encompassing an area where a rigid element of the reference sidewall extends or where a movable element of the reference sidewall can be moved away from the reference sidewall. This could allow for a more realistic assessment of the collision risk of the traffic object with the respective sidewall element projecting from the sidewall. For example, an exterior mirror and a fuel tank or charging port cover could each be represented by a small collision sector adjacent to the reference side wall, and a door by a collision sector larger according to a lateral dimension of the door. For example, all collision sectors could be rectangles to facilitate, for instance, determining the overlap with the traffic object following the trajectory in Cartesian coordinates, or, for instance, movable around a hinge. Side wall elements such as flaps and doors can be represented by circular sectors whose center point lies in or at the respective hinge and whose radius corresponds to a lateral dimension of the respective side wall element that can be rotated around the hinge. In this way, the false positive rate and / or the true negative rate of the overlap test could be reduced.

[0048] In one example, the detection of an overlap between the model object following the expected trajectory and the collision zone indicates: Defining a set of control points discretely distributed along the expected trajectory, wherein the set of control points contains fewer control points the greater the current speed of the traffic object determined from the sensor data; Mapping the model object within the environment model such that a first reference point of the model object lying on the expected trajectory is mapped to a second reference point selected from the set of vertices; checking whether the mapped model object has a point that lies on an edge of the collision zone or within the collision zone.

[0049] In this way, the reduced number of overlap checks at higher vehicle speeds could allow for earlier detection of a potentially imminent collision compared to, for example, a speed-independent number of checkpoints. This could lead to an earlier warning response at higher vehicle speeds, giving people in the surrounding traffic and / or inside the vehicle more time to react and potentially prevent the collision. Thus, the reduced time to collision (TTC) resulting from the higher vehicle speed could be offset by the possibility of an earlier warning response.

[0050] The following refers to the drawings, with similar elements being marked with the same reference symbols.

[0051] Fig. 1 shows a block diagram of a vehicle 100 equipped with a computer unit 110. The computer unit could have a processor 112 and a memory 114, which the processor 112 can access and which stores programs 116 that can be executed by the processor 112. Among the programs 116 could be program instructions 118 that could implement the method 200 described herein when executed by the processor 112.

[0052] The vehicle 100 could furthermore have a set of sensors 120 configured to detect the traffic environment of the vehicle 100. Examples of sensors 120 include a camera 122, a radar sensor 124, and a 2024PF01922 21. Infrared sensor 126 and an ultrasonic sensor 128. The sensors 200 could further be configured to provide sensor signals to the computer unit 110 via an interface, encoding sensor data that describe the traffic environment.

[0053] The vehicle 100 could further comprise a set of vehicle components 130, which could be controllably connected to the computer unit 110 via an interface. Examples of vehicle components 130 include a lighting system 132, which could include visual light signaling devices such as turn signals, brake lights, a rear fog light, license plate lights, reversing lights, etc.; a horn 134; a display 136, e.g., of a passenger information system in the interior of the vehicle 100; and loudspeakers 138, e.g., for audio playback in the interior and / or to the outside traffic environment.

[0054] Fig. 2 shows a flowchart with steps of a method 200 that enables computer-based warning of a potentially imminent collision between a vehicle 100 and a traffic object approaching from behind in the area of ​​a side wall of the vehicle 100. Execution of the program instructions 118 by the processor 112 could cause the computer unit 110 to carry out the method 200.

[0055] Method 200 could include a step 202 in which the computer unit 110 receives sensor data from the sensors 120, which detect the traffic environment of the vehicle 100. Method 200 could also include a step 204 in which the computer unit 110 operates an environment model that is continuously updated based on the sensor data. This would allow the environment model to represent the traffic environment in a way that is processable by the computer system.

[0056] The environment model could define an associated collision zone 302 for at least one side wall of the vehicle 100, for which this side wall serves as a reference side wall. The collision zone 302 could extend in a predetermined forward direction 304 of the vehicle 100 from a rear collision boundary 306 located laterally to the reference side wall to a front collision boundary 308 located laterally to the reference side wall. The environment model could also define at least one warning zone 300 extending behind the vehicle 100. In the example shown in Figures 3-7, each side wall of the vehicle 100 is assigned its own warning zone 300, which extends behind the vehicle 100 and laterally to the respective reference side wall, and its own collision zone 302, which adjoins the respective reference side wall.

[0057] The computer unit 110 could be configured by one of the programs 116 or the program instructions 118 to detect a traffic object based on the received sensor data, which is located in the traffic environment of the 2024PF01922 22 Vehicle 100 is located. Possible traffic objects could be, for example, a piece of road equipment (e.g., a street lamp), an obstacle, or another road user (e.g., a vehicle or a pedestrian). If a new traffic object is detected based on the sensor data, the computer unit 110 could create a model object 400 in the environment model that corresponds to the newly detected traffic object 206, which, for example, describes the traffic object at a point in time currently recorded in the sensor data by a geometric shape. Depending on the design of the environment model, for example, all or several traffic objects could be represented by an identical geometric shape (e.g., a rectangle), or some or all model objects 400 could have individual geometric shapes. Regardless of the shape, the individual geometric shape could be scaled based on the dimensions of a traffic object determined from the sensor data.The generated model object 400 could additionally be enriched with further information describing the corresponding traffic object, e.g., determined from the sensor data, such as an object type of the traffic object (e.g., automobile, two-wheeler, pedestrian, obstacle), spatial orientation information, speed and acceleration components, etc.

[0058] The computer unit 110 could also be configured by the program instructions 118 to execute a warning routine. The warning routine could monitor the environment model 208 to see if any of the model objects 400 overlap with a warning zone 300. If such an overlap is detected, this could cause the computer unit 110 to obtain a predicted trajectory 406 of the model object 400 210. The generation of the predicted trajectory could be a function of the program instructions 118, one of the programs 116, or another computing unit specializing in trajectory calculations based on sensor data.

[0059] The program instructions 118 could then cause the computer unit 110 to perform an overlap check using the predicted trajectory 406. With this overlap check, the computer system 110 could determine whether the environmental model, in conjunction with the sensor data, allows a projection of a future state of the traffic environment in which the model object 400, detected in warning zone 300, could enter a spatial region of the traffic environment corresponding to collision zone 302 in the environmental model. For this purpose, for example, the model object 400, or a copy thereof, could be moved along the predicted trajectory 406 to determine whether the moved model object 400 has a point that lies within collision zone 302 or on an edge of collision zone 302.The edge of collision zone 302 could include, in particular, the rear collision limit 306 and the front collision limit 308. In an alternative example, the 2024PF01922 23 could be included. The expected trajectory 406 with a starting point is attached to the model object 400 to determine whether there is an intersection of the expected trajectory 406 with the collision zone 302 or its edge.

[0060] The overlap check could be part of a more comprehensive warning condition, which, in addition to the overlap check, could contain further criteria that, in combination with one another (e.g., in a combination predefined by logical AND, OR, AND / OR operators), must be met in order for the warning routine to continue. Procedure 200 could thus perform the overlap check as part of a check 212 of the warning condition. If the warning condition is met, the program instructions 118 could cause the computer system 110 to perform a predefined warning response. This warning response could be configured to stop an entity that controls a movement and / or state of the vehicle 100 or the traffic object (e.g.,to alert a person or computer unit on board vehicle 100, and / or a person or computer unit associated with the motion control of traffic object 400, to the detected, potentially imminent collision risk between the traffic object and vehicle 100 within warning zone 302. In this way, the entity could be prompted to verify the potential collision risk and, if necessary, take a collision-preventing action.

[0061] Figures 4-7 show diagrams of various exemplary traffic situations represented in an environment model. The features shown in the drawings are based on representation options, such as the geometries of the elements shown, which were chosen solely for illustrative purposes and have no limiting effect on the description herein. The environment models shown each depict a vehicle 100 with a left side wall and a right side wall. The left side wall is the reference side wall for a left collision zone 302 adjacent to the left side wall and a left warning zone 300, which partially adjoins the left side wall and extends laterally to the left of the vehicle 100 behind the vehicle 100.The right side wall is the reference side wall for a right collision zone 302 adjacent to the right side wall and a right warning zone 300, which partially adjoins the right side wall and extends laterally to the right of the vehicle 100 behind the vehicle 100. The warning zones 300 and the collision zones 302 are each rectangular in shape with two sides parallel to the specified forward direction 304 of the vehicle 100 and two sides perpendicular to the forward direction 304.

[0062] In the examples shown in Figures 4-7, the warning and collision zones are each surrounded by overlapping hysteresis zones 410, which form a narrow border around the zones, e.g., approximately 10-50 cm wide. The left warning, collision, and hysteresis zones are separated from the right warning, collision, and hysteresis zones by a gap extending behind the vehicle 100. The 2024PF01922 24 Hysteresis zones 410 could have the function of maintaining a 214 warning reaction once triggered, as long as the traffic object 400 still overlaps with one of the hysteresis zones 410.

[0063] In Figures 4-6, the traffic object is a vehicle identified from sensor data in multiple measurements. The corresponding model object 400 is a rectangle that encloses the measurement points 402 corresponding to the measurements. The corner points of the rectangle facing the collision zone 302 were defined as reference points 404. A copy of a predicted trajectory 406 is attached to each of the reference points 404. The trajectories 406 each have a linear geometry. In the traffic situations shown, one of the attached trajectories 406 has an intersection point 408 with the rear collision boundary 306. The other attached trajectory 406 is shown with its own length and terminates at an endpoint 408 outside the collision zone 302, corresponding to the intersection point 408.

[0064] In the traffic situation shown in Fig. 4, the model object 400 is located entirely within the left warning zone 300. Since one of the attached predicted trajectories 406 intersects the left collision zone 302, the overlap condition that the model object 400 following the predicted trajectory 406 overlaps with the collision zone 302 is fulfilled. Without additional sub-conditions of the warning condition, the warning condition would be fulfilled in this example, and the predefined warning reaction 214 would be triggered.

[0065] In the traffic situation shown in Fig. 5, the model object 400 is located with its left front corner, i.e., a reference point 404, within the left warning zone 300 and with its right front corner, i.e., another reference point 404, in the gap between the two warning zones 300. The trajectory 406 attached to the reference point 404 located outside the left warning zone 300 intersects the left collision zone 302. The trajectory 406 attached to the reference point 404 located inside the left warning zone 300 terminates at an endpoint 408 located outside the left collision zone 302. The example in Fig. 5 illustrates that in some traffic situations it might be useful to check all reference points 404 for an overlap with the collision zone 302, regardless of where they are located relative to the warning zone 300 at a given time.Without additional sub-conditions of the warning condition, the warning condition would also be fulfilled in this example, and the specified warning reaction would be triggered. 214 In the example of Fig. 5, the warning routine is configured to continue the trajectories 406 in the forward direction 304 at least to the coordinate of the rear collision boundary 306 in order to check whether the respective trajectory 406 overlaps with the collision zone 302. In the drawing, this case is represented by a dashed line perpendicular to the forward direction 304, on which the endpoint 408' of the trajectory 406' originating from the left front corner point of the model object 400 lies. It is fundamentally possible that 2024PF01922 25. Different trajectories 406, 406' for the same model object 400 are extrapolated to different distances in order to perform the overlap check.

[0066] In the traffic situation shown in Fig. 6, the model object 400, with both reference points 404, is located within the right warning zone 300. One of the two attached trajectories 406 has an intersection point 408 with the left collision zone 302. In this case, an overlap condition is also met. However, whether the warning reaction 214 is triggered in this case could be chosen depending on the implementation. For example, if one assumes that the shallow angle at which the attached trajectory 406 intersects the rear collision boundary 306, compared to Figs. 4 and 5, indicates that the real traffic object will likely behave in a collision-preventing manner, then the warning reaction in this scenario could, for example,This can be suppressed by the fact that the warning condition contains an additional sub-condition requiring that the intersection point 408 must be on the same side of the vehicle as the reference point 404, which is connected to the intersection point 408 by the expected trajectory 406, or on the same side of the vehicle as the warning zone 300 with which the model object 400 overlaps. However, if one assumes that a possible warning response should always be issued in case of doubt, in order to maximize the potential safety contribution of the procedure 200, then the aforementioned additional condition could be omitted. Thus, in this case, the warning response 214 would also be carried out, even though the model object 400 and the expected intersection point 408 are located on different sides of the vehicle 100.

[0067] Fig. 7 shows another diagram of an environment model that describes a traffic object by a model object 400 with a predicted trajectory 406 leading into one of the collision zones 302. In this example, the traffic object is a pedestrian, represented in the model by a model object 400 having the geometric shape of an ellipse. The program instructions 118 could, for example, be configured to specify the object type of a traffic object as a pedestrian if the traffic object has speed components and / or dimensions typical for pedestrians. In the case of an approaching pedestrian, the program instructions could provide special handling that might differ from the case of a vehicle as the traffic object.

[0068] In the example of Fig. 7, the pedestrian approaches the right-hand warning zone 300 from the far right. Upon entering the collision zone 302, a geometric center point of the model object 400 representing the pedestrian could, for example, be defined as the sole reference point 404. Furthermore, the expected trajectory 406 of the model object could be determined, for example, by fitting it (e.g., using a least-squares algorithm) to the pedestrian's previous positions. In the example of Fig. 7, an arc-shaped expected trajectory 406 would best fit the previous positions of the model object 400. By attaching the 2024PF01922 26 Based on the predicted trajectory 406 to the reference point 404, the computer unit 110 could recognize that the predicted trajectory 406 leads into the right collision zone 302 and, for example, trigger the warning reaction 214 without further conditions. Optionally, the computer unit 110 could be configured to trigger the warning reaction only after reaching a predetermined minimum number (e.g., 2, 3, 4, ...) of measurement and test cycles 214 in which an overlap of the predicted trajectory 406 with one of the collision zones 302 has been detected. This could be done, for example, without a significant increase in the collision risk due to deceleration, since pedestrians often travel at lower speeds than vehicles.

[0069] Many other traffic situations are conceivable in which the warning reaction 214 is triggered, and likewise, in which the warning reaction is not triggered. Examples of warning-free traffic situations could include: a model object 400 moving within a hysteresis zone 410 without entering the warning zone 300; a model object 400 moving in a gap between two warning zones 300; a model object 400 moving within a warning zone 300 but, if it follows the expected trajectory 406, not overlapping with any of the collision zones 302, which could include a missing intersection 408 with any of the collision zones 302 for all attached trajectories 406 and / or a missing overlap of the traffic object 400, displaced along the expected trajectory 406, with any of the collision zones 302;a model object 400 overlapping a warning zone 300, which, following the expected trajectory 406, overlaps a collision zone 302 that is not on the same side of the vehicle 100 as the warning zone 300 with which the model object 400 overlaps; a model object 400 approaching a collision zone 302 without overlapping any of the warning zones 300 (e.g., approaching from the side or from the front); a model object 400 that is already in one of the collision zones 302 and could therefore likely be noticed by persons on board the vehicle 100.

[0070] Although the invention is illustrated and described in detail in the drawings and the preceding description, this illustration and description is to be regarded as exemplary and not limiting; the invention is not limited to the disclosed embodiments.

[0071] Although the invention is illustrated and described in detail in the drawings and the preceding description, this illustration and description is to be regarded as exemplary and not limiting; the invention is not limited to the disclosed embodiments.

[0072] As will be clear to those skilled in the art, aspects of the present invention could be embodied in the form of a device, a method, or a computer program product. Accordingly, aspects of the present invention could take the form of a purely hardware variant, a purely software variant (including firmware, resident software, microcode, etc.), or a variant that Software and hardware aspects are combined, which may be generally referred to here as a "circuit", "module", or "system". Furthermore, aspects of the present invention could take the form of a computer program product embodied in one or more computer-readable media containing computer-executable code.

[0073] Any combination of one or more computer-readable media could be used. The computer-readable medium could be a computer-readable signaling medium or a computer-readable storage medium. A "computer-readable storage medium," as used here, includes any physical storage medium capable of storing instructions that can be executed by a processor or computing system of a computer device. The computer-readable storage medium could be referred to as a computer-readable non-transitory storage medium. The computer-readable storage medium could also be referred to as a physical computer-readable medium. In some embodiments, a computer-readable storage medium could also be capable of storing data that the computing system of the computer device can access. Examples of computer-readable storage media include, but are not limited to: a floppy disk, a magnetic hard disk drive, a solid-state drive, flash memory, a USB flash drive, random access memory (RAM), read-only memory (ROM), an optical disk, a magneto-optical disk, and the computer system's register file. Examples of optical disks include compact discs (CDs) and digital versatile discs (DVDs), such as CD-ROM, CD-RW, CD-R, DVD-ROM, DVD-RW, or DVD-R discs. The term "computer-readable storage medium" also refers to various types of recording media that the computer can access via a network or communication connection. For example, data could be retrieved via a modem, the internet, or a local area network. Computer-executable code embodied on a computer-readable medium could be transmitted via any suitable medium, including but not limited to wireless transmission, wired transmission, fiber optic cable, radio frequency transmission, etc., or via a suitable combination of the aforementioned media.

[0074] A computer-readable signaling medium could contain a propagating data signal with computer-executable code embodied therein, for example, in the baseband or as part of a carrier wave. Such a transmitted signal could take any form, including, but not limited to, electromagnetic or optical signals, or a suitable combination thereof. A computer-readable signaling medium could be any computer-readable medium that is not a computer-readable storage medium and that can transmit, propagate, or transport a program for use by or in conjunction with a command execution system, apparatus, or device. 2024PF01922 28

[0075] "Computer memory" or "memory" is an example of a computer-readable storage medium. Computer memory is any storage that a computing system can directly access. Another example of a computer-readable storage medium is "computer storage" or "storage." Computer storage is any non-volatile, computer-readable storage medium. In some embodiments, computer storage could also be computer memory, or vice versa.

[0076] A “computer system,” as used here, comprises an electronic component capable of executing a program, machine-executable instruction, or computer-executable code. References to the computer system that include the example “a computer system” should be understood to mean that it may contain more than one computer system or processing core. The computer system could, for example, be a multi-core processor. A computer system could also refer to a collection of computer systems within a single computer system or distributed across multiple computer systems. The term “computer system” should also be interpreted as potentially referring to a collection or network of computer devices, each comprising a processor or computer systems.The machine-executable code or instructions could be executed by multiple computing systems or processors, which may be located within the same computer device or even distributed across multiple computer devices.

[0077] Machine-executable instructions or computer-executable code could include instructions or a program that causes a processor or other computing system to execute an aspect of the present invention. Computer-executable code for performing operations for aspects of the present invention could be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, Smalltalk, C++, or similar languages, and conventional procedural programming languages ​​such as the programming language "C" or similar languages, and compiled into machine-executable instructions. In some cases, the computer-executable code could be in the form of a higher-level programming language or in pre-compiled form and used in conjunction with an interpreter that generates the machine-executable instructions on the fly. In other cases, the machine-executable instructions or the computer-executable code could be in the form of programming for programmable logic gate arrays.

[0078] The executable computer code could run entirely on the user's computer, partially on the user's computer as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. Im2024PF01922 29 In the latter case, the remote computer could be connected to the user's computer via any network, including a local area network (LAN) or a wide area network (WAN), or the connection could be made with an external computer (e.g., via the Internet with the help of an Internet service provider).

[0079] Aspects of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the invention. It is understood that each block or part of the blocks of the flowchart, illustrations, and / or block diagrams can be implemented by computer program instructions in the form of computer-executable code, where applicable. It is further understood that combinations of blocks in different flowcharts, illustrations, and / or block diagrams could be combined, provided they are not mutually exclusive.These computer program instructions could be provided to a computing system of a general-purpose computer, a special-purpose computer, or any other programmable data processing device to create a machine, such that the instructions executed through the computing system of the computer or any other programmable data processing device provide means for implementing the functions / actions specified in the flowchart and / or block diagram block or blocks.

[0080] These machine-executable instructions or computer program instructions could also be stored in a computer-readable medium capable of instructing a computer, other programmable data processing device, or other apparatus to operate in a particular manner, such that the instructions stored in the computer-readable medium produce a manufactured item containing instructions to perform the function / action specified in the flowchart and / or block diagram block or blocks.

[0081] The machine-executable instructions or computer program instructions could also be loaded onto a computer, other programmable data processing device, or other equipment to initiate a series of operational steps that are executed on the computer, other programmable device, or other equipment to produce a computer-implemented process, so that the instructions executed on the computer or other programmable equipment provide processes for implementing the functions / actions specified in the flowchart and / or block diagram block or blocks.

[0082] A “user interface”, as used here, is an interface that allows a user or operator to interact with a computer or 2024PF01922 30 A user interface (UI) is a device used to interact with a computer system. A "user interface" could also be called a "human-interface device." A user interface could provide information or data to the operator and / or receive information or data from the operator. A user interface could allow an operator's input to be received by the computer, and it could provide the computer's output to the user. In other words, the user interface could allow an operator to control or influence a computer, and the interface could allow the computer to display the effects of the operator's control or influence. Displaying data or information on a screen or graphical user interface is an example of providing information to an operator.Receiving data via keyboard, mouse, trackball, touchpad, trackpoint, graphics tablet, joystick, gamepad, webcam, headset, pedals, wired glove, remote control, and accelerometer are all examples of user interface components that enable the reception of information or data from an operator. 2024PF01922 31. LIST OF REFERENCE MARKS 100 vehicles 110 computer unit 112 processor 114 storage 116 programs 118 program instructions 120 sensors 122 Camera 124 radar sensor 126 Infrared sensor 128 Ultrasonic sensor 130 vehicle components 132 Lighting system 134 Horn 136 Display 138 speakers 200 procedures 202 sensor data received 204 Operating the environmental model 206 Create model object Check warning zone 208 210 Obtain expected trajectory 212 Check warning condition 214 Carry out warning response 300 warning zone 302 Collision Zone 304 Forward direction 306 Rear Collision Limit 308 Front Collision Limit 400 model objects 402 Measuring point 404 Reference point 406 Expected trajectory 408 Intersection 410 Hysteresis zone

Claims

2024PF01922 32 REQUIREMENTS 1. Method (200), comprising, by means of a computer unit (110) of a vehicle (100): Receiving (202) sensor data from sensors (120) that detect a traffic environment of the vehicle (100); Operating (204) an environment model that is continuously updated based on sensor data, wherein the environment model has a reference side wall of the vehicle (100): a predefined warning zone (300) located behind the vehicle (100) and a predefined collision zone (302) extending in a predefined forward direction (304) of the vehicle (100) from a rear collision limit (306) located laterally to the reference side wall to a front collision limit (308) located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating (206) a model object (400) that describes the traffic object in the environment model; Performing a warning routine that includes: in response to a detection that the model object overlaps with the warning zone, obtain (210) a predicted trajectory (406) of the model object (400) in the environment model based on the sensor data; Check (212) whether a predefined warning condition is met, wherein the met warning condition shows an overlap of the model object (400) following the expected trajectory (406) with the collision zone (302); Perform (214) a predetermined warning response when the warning condition is met.

2. Method (200) according to claim 1, wherein the fulfilled warning condition additionally has at least one of the following fulfilled additional conditions: a speed of the traffic object (400) reaches or exceeds a specified minimum object speed; an expected duration until a collision of the traffic object with the vehicle (100) within the collision zone (302) reaches or falls below a specified maximum duration; 2024PF01922 33 a probability of existence of the traffic object reaches or exceeds a predetermined minimum probability; a certain amount of the speed of the vehicle (100) in the forward direction (304) reaches or falls below a predetermined maximum speed, preferably with the vehicle (100) at rest. Optionally, the warning routine is only executed if one of the additional conditions is met.

3. Method (200) according to one of the preceding claims, wherein the recognition of the traffic object comprises the recognition of an object type of the traffic object, wherein the model object (400) is generated with a geometric shape specified for the particular object type.

4. Method (200) according to one of the preceding claims, wherein the test to determine whether the model object (400) following the expected trajectory (406) overlaps with the collision zone (302) comprises determining a reduced set of reference points (404) representing the model object (400) and is limited to checking whether one of the reference points (404) enters the collision zone (302) when the model object (400) follows the expected trajectory (406), where optionally checking whether one of the reference points (404) enters the collision zone (302) when the model object (400) follows the expected trajectory (406), attaching the trajectory (406) to one of the reference points (404), and checking whether the attached trajectory (406) intersects an edge of the collision zone (302).

5. Method (200) according to claim 4, wherein at least one of the following applies: in the case that the specified object type is a vehicle, the specified geometric shape is a polygon, in particular a rectangle, and the reference points (404) are selected from outer vertices of the polygon; In the case that the specified object type is a pedestrian, the set of reference points (404) contains exactly one reference point (404) and encloses the specified geometric shape the exactly one reference point (404), wherein the exactly one reference point (404) is in particular a geometric center or centroid of the specified geometric shape.

6. Method (200) according to claim 4 or 5, wherein the warning routine additionally comprises selecting a reduced subset from the set of reference points (404), wherein for the subset only reference points (404) are selected for which an entry into the collision zone (302) is most probable in comparison of the reference points (404) of the set of reference points (404) among themselves, wherein in particular the reference points (404) for which an entry into the 2024PF01922 34 In the comparison of the reference points (404) of the set of reference points (404) that are most likely to be the collision zone (302), reference points (404) are selected that are located on a side of the model object (400) facing the vehicle (100) and / or the collision zone (302).

7. Method (200) according to one of the preceding claims, wherein, in response to a detection that the model object (400) following the expected trajectory (406) no longer overlaps with the collision zone (302), the warning reaction is only terminated when a predetermined holding time has elapsed and / or when the model object (400) has left a predetermined hysteresis zone (410) surrounding the warning zone (302).

8. Method (200) according to one of the preceding claims, wherein obtaining (210) the expected trajectory (406) comprises selecting a predetermined trajectory geometry based on the sensor data and determining the expected trajectory (406) based on the selected trajectory geometry.

9. Method (200) according to one of the preceding claims, wherein the collision zone (302) is composed of collision sectors, each collision sector enclosing an area in which a rigid element of the reference side wall extends or in which a movable element of the reference side wall can be moved away from the reference side wall.

10. Method (200) according to one of the preceding claims, comprising the detection of an overlap of the model object (400) following the expected trajectory (406) with the collision zone (302): Defining a set of support points discretely distributed along the expected trajectory (406), wherein the set of support points contains fewer support points the greater the current velocity of the traffic object determined from the sensor data; Mapping the model object (400) within the environment model such that a first reference point of the model object (400) lying on the expected trajectory (406) is mapped to a second reference point selected from the set of vertices; Check if the depicted model object (400) has a point that lies on an edge of the collision zone (302) or within the collision zone (302).

11. Computer unit (110) for a vehicle (100), wherein the computer unit (110) comprises a processor (112) and a memory (114) functionally connected to the processor (112), wherein the memory (114) stores program instructions (118) which, when executed by the processor (112), cause the computer unit (110) to perform a procedure (200) comprising: 2024PF01922 35 Receiving (202) sensor data from sensors (120) that detect a traffic environment of the vehicle (100); Operating (204) an environment model that is continuously updated based on sensor data, wherein the environment model has a reference side wall of the vehicle (100): a predefined warning zone (300) located behind the vehicle (100) and a predefined collision zone (302) extending in a predefined forward direction (304) of the vehicle (100) from a rear collision limit (306) located laterally to the reference side wall to a front collision limit (308) located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating (206) a model object (400) that describes the traffic object in the environment model; Performing a warning routine that includes: In response to a detection that the model object (400) overlaps with the warning zone (300), obtain (210) a predicted trajectory (406) of the model object (400) in the environment model based on the sensor data; check (212) whether a predefined warning condition is met, wherein the met warning condition is an overlap of the model object (400) following the predicted trajectory (406) with the collision zone (302); execute (214) a predefined warning response if the warning condition is met.

12. Computer program product, in particular a computer-readable medium, wherein the computer program product carries computer-executable code (118) for execution by a processor (112) of a computer unit (110) of a vehicle (100), wherein the execution of the instructions (118) causes the processor (112) to perform a method (200) comprising: Receiving (202) sensor data from sensors (120) that detect a traffic environment of the vehicle (100); Operating (204) an environment model that is continuously updated based on sensor data, wherein the environment model has a reference side wall of the vehicle (100): a predefined warning zone (300) located behind the vehicle (100), and a predefined collision zone (302) extending in a predefined forward direction (304) of the vehicle (100) from a lateral object. extends from the rear collision limit (306) located at the reference side wall to a front collision limit (308) located laterally to the reference side wall; in response to the detection of a traffic object within the traffic environment based on sensor data, generating (206) a model object (400) that describes the traffic object in the environment model; Performing a warning routine that includes: In response to a detection that the model object (400) overlaps with the warning zone (300), obtain (210) a predicted trajectory (406) of the model object (400) in the environment model based on the sensor data; check (212) whether a predefined warning condition is met, wherein the met warning condition is an overlap of the model object (400) following the predicted trajectory (406) with the collision zone (302); execute (214) a predefined warning response if the warning condition is met.

Images