DE102021101130A1 - Method for determining a risk of collision and for mitigating or avoiding a collision, computer program, computer-readable medium, collision risk monitoring system and moving object - Google Patents

Abstract

The present invention relates in particular to a method (300) for determining a risk of collision between a moving object (20, 20'), in particular a vehicle (20, 20'), and one located in the vicinity of the object (20, 20'). Foreign object (10) and comprises the steps: a) providing (S3-1) detected state variables (Z1..n),b) determining (S3-3) collision risk parameter values (KP1..3) depending on one or more of the provided state variables (Z1..3),c) determination (S3-4) of individual collision time values (TTC1..3) depending on the associated collision risk parameter values (KP1..3),d) determination (S3-5) of an overall collision time value (TTCtotal), which characterizes an overall risk of collision, from the individual, individual collision time values (TTC1..3) using weighting factors (W1..3),e) evaluating (S3-6) the total collision time value (TTCtotal), andf) generating (S3-6) at least one collision state of danger signal (KZS).

Description

The present invention relates to a method, in particular a computer-implemented method, for determining a risk of collision between a moving object, in particular a vehicle, and a foreign object located in the vicinity of the object.

The present invention also relates to a method, in particular a computer-implemented method, for mitigating or avoiding a collision between a moving object, in particular a vehicle, and a foreign object in the vicinity of the object, the object having at least one component that can be controlled by a control device , which is designed and set up to mitigate or avoid a collision between the object and the foreign object and/or to output an indication of an imminent collision.

Furthermore, the present invention relates to a computer program and a computer-readable medium, in particular a computer-readable storage medium, each comprising instructions which, when executed by a computer, cause the computer to carry out one of the aforementioned methods.

In addition, the present invention relates to a collision risk monitoring system, at least for determining a risk of collision between a moving object, in particular a vehicle, and a foreign object located in the vicinity of the object.

In addition, the present invention relates to a moving object, in particular a vehicle, with an aforementioned collision risk monitoring system.

Methods for determining a risk of collision and for mitigating or avoiding a collision between a moving object, in particular a moving object such as a vehicle, and a foreign object, including moving foreign objects such as vehicles driving ahead, are known in principle from the prior art, for example from US 2014/0343749 A1 . Likewise, collision monitoring systems and moving objects, such as vehicles, are known with such systems, which are set up to support, for example, a brake application adapted to the situation or to initiate or automatically carry out emergency braking when a risk of collision has been detected.

Against this background, it is an object of the present invention to provide an alternative, in particular improved, method for determining a risk of collision between a moving object, in particular a vehicle, and a foreign object in the vicinity of the object, as well as an alternative, in particular improved, method to mitigate or avoid a collision, to provide an alternative, in particular improved, system for collision monitoring and an alternative, in particular improved, moving object, in particular a vehicle, with such a system, as well as a corresponding computer program and a computer-readable medium.

This object is achieved according to the invention by a method according to the invention for determining a risk of collision between a moving object, in particular a vehicle, and a foreign object in the vicinity of the object, by a method for mitigating or avoiding a collision, by a collision risk monitoring system, by a computer program, solved by a computer-readable medium and by a movable object having the features according to the respective independent patent claims. Advantageous embodiments of the invention are the subject matter of the dependent patent claims, the description and the figures and are explained in more detail below. The patent claims are made part of the content of the description by express reference.

1. a) Bereitstellen von erfassten Zustandsgrößen, die jeweils einen Zustand, eine Eigenschaft oder eine Bewegung des Objekts charakterisieren, sowie von Zustandsgrößen, die jeweils einen Zustand, eine Eigenschaft oder eine Bewegung des Fremdobjekts charakterisieren,
2. b) Ermitteln von wenigstens zwei, eine Kollisionsgefahr charakterisierenden Kollisionsgefahrparameterwerten in Abhängigkeit von einer oder mehreren der bereitgestellten Zustandsgrößen,
3. c) Ermitteln eines individuellen Kollisionszeitpunktwertes zu wenigstens einem Kollisionsgefahrparameter, von dem ein Kollisionsgefahrparameterwert ermittelt worden ist, in Abhängigkeit von dem zugehörigen Kollisionsgefahrparameterwert,
4. d) Ermitteln eines Gesamt-Kollisionszeitpunktwertes, der eine Gesamt-Kollisionsgefahr charakterisiert, aus den einzelnen, individuellen Kollisionszeitpunktwerten mithilfe von Gewichtungsfaktoren,
5. e) Auswerten des Gesamt-Kollisionszeitwertes, und
6. f) Erzeugen wenigstens eines Kollisionsgefahrzustandssignals.

A method according to the invention, in particular a computer-implemented method, for determining a risk of collision between a moving object, in particular a vehicle, and a foreign object located in the vicinity of the object, comprises at least the following steps:

1. a) providing detected state variables, each of which characterizes a state, a property or a movement of the object, and state variables, each of which characterizes a state, a property or a movement of the foreign object,
2. b) determination of at least two collision risk parameter values characterizing a risk of collision as a function of one or more of the provided state variables,
3. c) determining an individual collision time value for at least one collision risk parameter, from which a collision risk parameter value has been determined, as a function of the associated collision risk parameter value,
4. d) determining an overall collision time value, which characterizes an overall risk of collision, from the individual collision time values using weighting factors,
5. e) evaluating the total collision time value, and
6. f) generating at least one collision risk condition signal.

By determining at least two collision risk parameter values that characterize a risk of collision as a function of one or more of the state variables provided, subsequent determination of an individual collision time value as a function of the associated collision risk parameter value and determination of an overall collision time value that characterizes an overall risk of collision from the individual, individual collision time values using weighting factors, the actual risk of collision can be determined more precisely.

In particular, a more accurate mapping/simulation of the collision situation can be achieved by determining a plurality of collision risk parameters and the subsequent weighting of the individual collision time values determined as a function of the individual collision risk parameters. In this way, in particular the accuracy of the determination of the risk of collision can be improved, in particular the rate of collisions identified as falsely positive can be reduced and/or the rate of collisions identified as correctly positive can be increased.

As a result, an improved assistance function and/or greater safety, in particular improved collision mitigation or better collision avoidance, can be achieved with a method according to the invention.

A "movable object" within the meaning of the present invention is in particular an object that is movable, i.e. the location of which can be changed. A “movable object” within the meaning of the present invention is in particular a vehicle, preferably a motor vehicle. However, a moving object within the meaning of the present invention can also be a rail vehicle, a watercraft, a flying object, for example a drone, an air taxi or the like. An object preferably has at least one drive device, by means of which it can be driven.

The “movable object” can be three-dimensionally movable in space or two-dimensionally only on the ground, whereby the object can be freely movable or can be guided along a guided path, for example with the help of rails.

A “foreign object” within the meaning of the present invention is in particular another object, such as another vehicle, in particular another motor vehicle, such as a truck, car or the like. A foreign object within the meaning of the present invention can also be a movable object, but does not have to be movable. A foreign object can therefore be any type of "obstacle", such as a crash barrier, a wall, a pedestrian, a building or the like.

In the context of the present invention, the term “risk of collision” is understood to mean the risk of a collision between a moving object, in particular a moving object, and a foreign object, with the object in particular being at a distance from the foreign object and the object being in particular with a Object speed moves and the foreign object in particular with a foreign object speed. The object can move in particular at a current object speed relative to the foreign object, with the fundamental risk of collision between the object and the foreign object if there is a collision between the object without taking measures, for example decelerating the object or changing the trajectory of the object and foreign object is to be expected.

A “state variable” within the meaning of the present invention is in particular a measured variable or a defined, stored variable that identifies or characterizes any type of state, property or movement of the object, the foreign object or an environment of the object.

On the one hand, state variables that characterize the object are required to determine the risk of collision. These can be, for example, state variables such as an object speed, an object acceleration, an object position, an object dimension or object expansion, an object movement trajectory, in particular a predicted (pre-calculated, assumed for the near future) movement trajectory, Collision area variables, which enable collision areas to be classified and/or a potential collision to be assigned to a specific area and/or state variables which, for example, characterize a type of movement of the object (forward, backward, straight ahead, along a curve, etc.) and/or state variables that characterize the trustworthiness of one or more object state variables.

Dynamic state variables, i.e. state variables that change over time and characterize the object, such as object speed, object acceleration and object position, are particularly preferably measured variables, while static state variables, i.e. state variables that do not change over time, such as state variables, which characterize the object dimension or its expansion/measurements, can also be permanently stored variables.

State variables that characterize the foreign object are also required. These can be, for example, state variables such as a foreign object speed, a foreign object acceleration, a foreign object position, a foreign object dimension or foreign object extent, a foreign object movement trajectory, in particular a predicted (pre-calculated, assumed for the near future) foreign object Movement trajectory, and/or state variables, which contain a time stamp, for example, to characterize, for example, a point in time at which a specific foreign object was first detected or the like, and/or state variables, which, for example, characterize a type of movement of the foreign object (forward, backward , straight ahead, along a curve, etc.) and/or state variables that characterize the trustworthiness of one or more foreign object state variables.

Furthermore, state variables can be provided which each characterize a state or a property of the environment, for example as a function of GPS data.

A “collision risk parameter” within the meaning of the present invention is a parameter that characterizes a collision risk, in particular as a function of certain state variables, as a function of which this collision risk parameter has been determined.

A first “collision risk parameter” can be, for example, a speed risk of collision parameter, which has been determined, for example, from an object speed and a foreign object speed and which characterizes the risk of collision based on the speeds of the object and foreign object, for example based on one of the individual speeds determined relative speed between object and foreign object.

Another "collision risk parameter" can be, for example, a degree of overlap collision risk parameter, which has been determined, for example, depending on the positions of the object and foreign object and depending on their dimensions, and which determines the risk of collision due to a depending on the positions and dimensions of the object and foreign object determined degree of overlap characterized.

A "collision risk parameter value" within the meaning of the present invention is in each case a specific measure for a collision risk parameter, i.e. in particular a specific numerical value.

When a method according to the invention is carried out, preferably not only two but more than two collision risk parameter values characterizing a risk of collision are determined as a function of one or more of the state variables provided. As a result, the accuracy of the determination of the risk of collision can be improved, in particular the rate of collisions identified as falsely positive can be reduced and/or the rate of collisions identified as positively correct can be increased.

An “individual collision time value” in the sense of the present invention characterizes the intensity of the collision risk for the respective collision risk parameter or is a measure for this, in each case based on the determination or calculation on which the respective collision risk parameter is based. In particular, an “individual collision time value” is a value, in particular a TTC time (Time To Collision time), which indicates the remaining time up to the collision, which has been determined based on the determination or calculation on which the respective risk of collision parameter is based.

When carrying out a method according to the invention, an individual collision time value is preferably determined for each collision risk parameter that has been determined. As a result, the accuracy of the determination of the risk of collision can be improved, in particular the rate of collisions identified as falsely positive can be reduced and/or the rate of collisions identified as positively correct can be increased.

An “overall collision time value” in the sense of the present invention correspondingly characterizes the intensity of the overall risk of collision or the “overall” risk of collision across all parameters of the risk of collision or is a measure for this. In particular, a “total collision time value” is a value, in particular a TTC time (Time To Collision time) showing a remaining time up to collision which has been determined based on all collision risk parameters.

The determination of the overall collision time value in step d) from the individual, individual collision time values is preferably carried out using a number of weighting factors, with each individual collision time value being assigned a (own) weighting factor in particular. As a result, the accuracy of the determination of the risk of collision can be improved, in particular the rate of collisions identified as falsely positive can be reduced and/or the rate of collisions identified as positively correct can be increased.

A method according to the invention makes it possible, for example, to avoid the false positive detection of a risk of collision due to a detected overlap if it has been detected at the same time that the foreign object is moving along a movement trajectory at a foreign object speed such that a collision cannot occur expect is. As a result, the rate of false positive detections can be reduced.

Furthermore, by taking into account more than one collision risk parameter, in particular if several collision risk parameters are taken into account, the rate of correctly positively recognized collisions or falsely negatively recognized collisions can be increased, since more parameters, in particular more state variables, and in particular different evaluations of these are taken into account be able.

In a method according to the invention, the overall collision time value is evaluated in particular by analyzing the overall collision time value, for example by comparing the overall collision time value with one or more threshold values, with the collision risk state signal being generated in step f) of a method according to the invention , is generated in particular as a function of the result of the analysis of the total collision time value.

A "collision risk status signal" within the meaning of the present invention is in particular a signal which indicates or characterizes a collision risk status, i.e. which contains corresponding status information which in particular indicates whether there is a risk of collision or not.

The collision risk status signal can not only contain status information that indicates whether there is a risk of collision or not, but can (but does not have to) also contain information on how critical the collision situation is assessed, i.e. whether there is a high or low risk of collision . I.e. the collision risk status signal can also contain collision risk level information in particular.

The collision risk status signal can also not only be a pure status information signal that contains information about the collision risk status, but can already be a control signal, in particular a control signal that does not require any further conversion or modification or conversion or the like in order to control at least one component of the object to be used or to be able to be used.

The collision risk status signal is preferably not only generated, but also output, i.e. made available in particular for further processing or further use. In particular, the collision risk status signal can be output, for example, to a control device for controlling at least one component of the object to mitigate or avoid a collision, the respective component of the object being set up in particular to receive the collision risk status signal and to be controlled by it. Such a component can be, for example, a deceleration device such as a braking device, or an optical, acoustic or haptic display device or a belt tensioner, a steering device or the like.

In an advantageous embodiment of a method for determining a risk of collision according to the present invention, the state variables to be provided for determining the respective collision parameter values are recorded in particular, preferably in a further step, in particular before step a), in particular by means of at least one sensor device or at least one sensor module which is designed in each case to detect one or more state variables that characterize a state, a property and/or a movement of the object or the foreign object and/or a state or a property of the environment.

One or more state variables are particularly preferably recorded in particular by the object itself, i.e. in particular by sensor devices assigned to it, which in particular has one or more corresponding internal sensor devices or sensor modules for this purpose.

An “internal sensor device” within the meaning of the present invention is a sensor device that is arranged or installed in the object.

In particular, the object can have at least one camera device and/or a laser scanner and/or a radar device for this purpose. The object preferably also has further sensor devices or modules, such as wheel speed sensors and/or acceleration sensors and/or GPS sensors or the like, in particular for detecting object state variables.

Alternatively or additionally, however, external sensor devices can also be used to record one or more state variables, i.e. sensor devices arranged or installed outside of the object, which, however, are preferably each designed accordingly to provide the recorded state variables for further processing or in particular to a corresponding, in particular to transmit built-in object collision risk parameter determination module for determining the collision risk parameters.

An “external sensor device” within the meaning of the present invention is accordingly a sensor device that is located outside the object, in particular arranged or installed outside of it.

An external sensor device can, for example, be a camera or a sensor device of a foreign object, for example a GPS sensor device which, for example, records GPS data of the foreign object and is coupled or connected to a corresponding collision risk parameter determination module of an object, for example in a data or signal-communicating manner .

The state variables required to characterize the object are particularly preferably detected using the sensor devices installed in the object for operating the object, i.e. in particular using corresponding internal sensor devices. In a vehicle, for example, by means of wheel speed sensors, acceleration sensors, steering angle sensors, torque sensors, brake pressure sensors or the like.

In a particularly advantageous embodiment of a method for determining a risk of collision according to the present invention, the state variables required to characterize the foreign object are recorded in particular by means of at least one camera device and/or a laser scanner and/or a radar device. This enables status variables of all possible foreign objects to be detected without being dependent on the respective foreign object in each case.

In a particularly advantageous embodiment of a method for determining a risk of collision according to the present invention, the state variables of the foreign object are detected using a front sensor device and/or a corner sensor device and/or a side sensor device. As a result, a risk of collision in an area surrounding an object can be achieved particularly well, in particular in an area in front of and/or to the side of an object.

A "front sensor device" within the meaning of the present invention is a sensor device for detecting an environment in front of the object, in particular in the area of a front, with such a sensor device being arranged in particular in the front of the object or integrated into the front of the object a vehicle, for example in the radiator grille, particularly preferably in each case in a central area.

A "corner sensor device" within the meaning of the present invention is accordingly a sensor device for detecting an environment in the area of a corner of the object, in particular in the area of a front corner of the object, and is arranged in particular in the area of a corner of the object, preferably in the area of a front corner.

Particularly preferably, two corner sensor devices are provided on an object, in particular if the object is a vehicle, in particular on the two front corners to cover the two front corner regions. A foreign object can thus be detected and recorded not only in the front area immediately in front of the object, but also in front and to the side, as a result of which a risk of collision can also be determined in an angular area in front of the object and not just in a straight corridor in front of the object.

A "side sensor device" within the meaning of the present invention is accordingly a sensor device for detecting a lateral environment in the area of a side of the object, in particular in the front side area, and is arranged in particular on or in the side area of the object, in particular in the front area of a side of the Object, preferably in each case on both sides of the object (left and right) to cover both sides a side sensor device is provided with an object. As a result, a risk of collision can also be determined in an area to the side of the object.

A front sensor device, a corner sensor device and/or a side sensor device can in particular each be a camera, a laser scanner or a radar sensor device.

In a further advantageous embodiment of a method for determining a collision risk according to the present invention, the collision risk parameter values are determined in step b), in particular by means of a collision risk parameter determination module, which is particularly preferably assigned to the object. Alternatively, the collision risk parameter determination module can also be integrated into a cloud and only be connected to the object for data and signal transmission, in particular by radio.

In a further advantageous embodiment of a method for determining a risk of collision according to the present invention, the determination of the individual collision time values is particularly preferably carried out using a collision time determination module, which is also particularly preferably assigned to the object, but alternatively can also be integrated into a cloud and with can only be connected to the object for data and signal transmission, in particular by radio.

• - einen Positions-Kollisionsgefahrparameter,
• - einen Geschwindigkeits-Kollisionsgefahrparameter,
• - einen Beschleunigungs-Kollisionsgefahrparameter,
• - einen Dimensions-Kollisionsgefahrparameter,
• - einen Klassifizierungs-Kollisionsgefahrparameter,
• - einen Alter-Kollisionsgefahrparameter,
• - einen Überlappungsgrad-Kollisionsgefahrparameter,
• - einen Eindringungsregion-Kollisionsgefahrparameter,
• - einen Bewegungstrajektorien-Kollisionsgefahrparameter,
• - einen Bewegungsart-Kollisionsgefahrparameter und
• - einen Vertrauenswürdigkeits-Kollisionsgefahrparameter.

In an advantageous embodiment of a method for determining a collision risk according to the present invention, at least two collision risk parameter values that are determined in step b) are values of collision risk parameters that are selected from a group containing at least

• - a position collision risk parameter,
• - a speed collision risk parameter,
• - an acceleration collision risk parameter,
• - a dimensional collision risk parameter,
• - a classification collision risk parameter,
• - an age collision risk parameter,
• - an overlap degree collision risk parameter,
• - an intrusion region collision risk parameter,
• - a motion trajectory collision risk parameter,
• - a movement type collision risk parameter and
• - a confidence level collision risk parameter.

A value is preferably determined in step b) for each of the collision risk parameters mentioned in this group. Furthermore, an associated variance (standard deviation, spread) and/or an associated covariance is particularly preferably determined in each case, in particular a covariance as a function of the state variables on which the determination of the respective parameter value is based.

A "position risk of collision parameter" within the meaning of the present invention is a risk of collision parameter which has been determined in particular as a function of an object position and/or a foreign object position and which characterizes the risk of collision in particular based on the positions of the object and/or foreign object , for example based on a distance determined from the individual positions.

A "speed collision risk parameter" in the sense of the present invention is a collision risk parameter that has been determined in particular as a function of an object speed and/or a foreign object speed and which characterizes the risk of collision based in particular on the speeds of the object and/or foreign object , For example, based on a determined from the individual speeds relative speed between object and foreign object.

An "acceleration risk of collision parameter" within the meaning of the present invention is a risk of collision parameter which has been determined in particular as a function of an object acceleration and/or a foreign object acceleration and which characterizes the risk of collision based in particular on the accelerations of the object and/or foreign object , For example, based on a determined from the individual accelerations relative acceleration between object and foreign object.

A “dimension collision risk parameter” within the meaning of the present invention is a collision risk parameter that has been determined in particular as a function of an object dimension, ie as a function of an object dimension, and/or a foreign object dimension, ie a foreign object dimension is and characterizes the risk of collision, in particular based on the dimensions of the object and/or foreign object, for example based on a distance determined from the individual dimensions, esp special in connection with the positions of object and foreign object.

A "classification risk of collision parameter" in the sense of the present invention is a risk of collision parameter, which depends in particular on an object classification, i.e. depending on an assignment of an object to a defined class based on its properties, and/or a foreign object classification, i.e. a foreign object classification, has been determined and which characterizes the risk of collision, in particular based on a determined classification of the object and/or foreign object.

An "age collision risk parameter" within the meaning of the present invention is a collision risk parameter that has been determined in particular as a function of an "age" of the foreign object, for example as a function of a time stamp that indicates when a specific foreign object was in an environment for the first time of an object has been detected, and which characterizes the risk of collision, in particular based on an “age” of the foreign object.

A "degree of overlap - risk of collision parameter" within the meaning of the present invention is a risk of collision parameter that has been determined in particular as a function of the positions of the object and / or foreign object and as a function of their dimensions and the risk of collision due to a depending on the positions and of characterized by the dimensions of the object and/or foreign object determined degree of overlap.

A "penetration region collision risk parameter" within the meaning of the present invention is a collision risk parameter that has been determined in particular as a function of the positions, dimensions and (predicted) movement trajectories of the object and / or foreign object and the risk of collision or in particular its severity based on a determined Collision area on the object or a determined area in which the foreign object penetrates in a collision characterized.

A “movement trajectory collision risk parameter” in the sense of the present invention is a collision risk parameter that has been determined in particular as a function of the movement trajectories of the object and foreign object, in particular as a function of the respectively predicted movement trajectories, and which calculates the risk of collision based on ( predicted) movement trajectories characterized.

A "movement type collision risk parameter" within the meaning of the present invention is a collision risk parameter that depends in particular on a classification of an object movement, i.e. depending on an assignment of an object movement to a defined class based on certain characteristics of the object movement, and /or a classification of a foreign object movement has been determined and which characterizes the risk of collision in particular based on a determined classification of object movement and/or foreign object movement, for example depending on a type of object movement (forward, backward, straight ahead, along a curve) and/or depending on a type of foreign object movement (forward, backward, straight ahead, along a curve).

A “trustworthiness risk of collision parameter” in the sense of the present invention is a risk of collision parameter that has been determined in particular as a function of trustworthiness, for example as a function of the trustworthiness of at least one state variable, and that characterizes the risk of collision in particular based on a determined trustworthiness.

In a particularly advantageous embodiment of a method for determining a risk of collision according to the present invention, in particular in a development of a method for determining a risk of collision, at least one collision risk parameter value is determined as a function of an environment model and/or an object model.

Particularly preferably, at least one collision risk parameter value is determined as a function of an environment model that has been created using the provided state variables, in particular using the state variables that characterize the foreign object, with the environment model particularly preferably containing the foreign object and/or as a function of a Object model that has been created using the provided state variables, in particular using the state variables that characterize the object, with static state variables such as object dimensions or the like being stored.

The environment model is particularly preferably created using image processing methods, for example using camera images provided by a front camera and other state variables that have been recorded using a corresponding environment sensor system, such as one or more laser scanners or radar sensor devices, which can be used as front, corner or Side sensor devices can be installed in the object.

Particularly good accuracy can be achieved when determining the collision risk parameter values by using the environment and/or object model. For this purpose, for example, a parameterized environment model and/or a corresponding object model can be stored, the parameters of which can be assigned the provided state values in order to determine the respective risk of collision parameter values.

In some cases it can be advantageous if a weighting is also carried out when the environment model and/or the object model is created, in particular a weighting of the individual state variables and, if necessary, a preliminary evaluation, and in particular only a determination of the collision risk parameter values takes place in the further course if a Collision cannot be ruled out based on the preliminary evaluation of the state variables. This allows resources to be conserved or kept free. Furthermore, the quality of the determination of the risk of collision can be further improved by this additional checking step.

In a further advantageous embodiment of a method for determining a risk of collision according to the present invention, at least one state parameter value is determined, in particular also in step b), depending on one or more of the provided state variables.

In the context of the present invention, a “state parameter” is understood to mean a parameter that characterizes any type of state, in particular a parameter that characterizes an object state, an object property or an environmental state. A state parameter can be a state variable or a parameter derived or calculated from one or more state variables.

Accordingly, a "state parameter value" is a concrete value of a state parameter.

In a particularly advantageous embodiment of a method for determining a risk of collision according to the present invention, at least one state parameter value is determined, which is a value for a state parameter that characterizes the ratio of a current object speed to a current foreign object speed or to a current relative speed between object and foreign object . In this way, a particularly advantageous evaluation of the risk of collision can be achieved, which will be explained in more detail below.

Alternatively or additionally, one or more other state parameter values can also be determined, e.g. a current distance value between object and foreign object or a current relative acceleration value.

In a particularly advantageous embodiment of a method for determining a collision risk according to the present invention, in particular in a development, at least one individual collision time value is determined in step c) not only as a function of the respective associated collision risk parameter value, but also as a function of at least one determined State parameter value, in particular as a function of a determined state parameter value that characterizes the ratio of a current object speed to a current foreign object speed. This allows the risk of collision to be determined particularly well, since the relative speed between the object and foreign object or the ratio of a current object speed to a current foreign object speed or a current relative speed has a significant influence on the risk of collision.

In a particularly advantageous embodiment of a method for determining a collision risk according to the present invention, at least one individual collision time value is determined in step c) using a stored characteristic map, preferably using a two-dimensional characteristic map or using a two-dimensional matrix, in particular as a function of the respective associated collision risk parameter value and a state parameter value, the state parameter value being in particular the value of a state parameter which characterizes the ratio of a current object speed to a current foreign object speed. Each individual collision time value is particularly preferably determined in this way in step c). As a result, the individual collision time values can be determined in step c) in a particularly simple, resource-saving, but extremely advantageous manner.

A two-dimensional map can in particular be a two-dimensional table or matrix, according to which an individual collision time value is assigned to the various value ranges of collision risk parameter values and state parameter values, which for the respective station wagon nation of collision risk parameter value and state parameter value can be taken from the table. As a result, the individual collision time values can be determined in step c) in a particularly simple, resource-saving, but extremely advantageous manner.

It is also conceivable, however, to determine the individual collision time value using an interpolation or extrapolation of the values stored in the characteristics map. In some cases, this allows the accuracy of the determination of the risk of collision to be improved even further. However, this procedure requires more resources (memory, CPU time, etc.) than the previously described, pure extraction of the individual collision time value from the characteristics map or the table.

The characteristics map can also be multi-dimensional, for example to represent or map a dependency on various state parameter values. In some cases, this allows the accuracy of the determination of the risk of collision to be improved even further.

In a further advantageous embodiment of a method for determining a collision risk according to the present invention, in particular in a further development, the overall collision time value is determined from the individual collision time values in step d), in particular with the aid of predetermined, permanently stored weighting factors, in particular with the aid of in weighting factors stored in a database for the individual, individual collision risk parameters, with at least one separate, associated weighting factor being stored in particular for each collision risk parameter. Alternatively, the individual weighting factors can also be determined as a function of one or more state variables and/or state parameters or can be corrected as a function of these. In some cases, this allows the accuracy of the determination of the risk of collision to be improved even further.

In a further advantageous embodiment of a method for determining a risk of collision according to the present invention, the overall collision time value is determined in step d), in particular using a classification and/or regression method, using an artificial neural network, using fuzzy logic or using a Combination of these, for example using a so-called random forest method, using a “convolutional neural network” (CNN) or a “deep learning neural network” (DNN) and/or using fuzzy logic. In this way, particularly good results can be achieved, in particular for many different foreign objects.

1. i) Erfassen von Zustandsgrößen, die jeweils einen Zustand, eine Eigenschaft oder eine Bewegung des Objekts charakterisieren, sowie von Zustandsgrößen, die jeweils einen Zustand, eine Eigenschaft oder eine Bewegung des Fremdobjekts charakterisieren, mittels wenigstens einer Sensoreinrichtung,
2. ii) Durchführen eines vorbeschriebenen erfindungsgemäßen Verfahrens zum Bestimmen einer Kollisionsgefahr, und
3. iii) Ansteuern der wenigstens einen Komponente des Objekts mittels einer Steuerungseinrichtung in Abhängigkeit von dem erzeugten und ausgegebenen Kollisionsgefahrzustandssignals derart, dass, wenn eine Kollisionsgefahr vorliegt, eine Milderung oder Vermeidung einer Kollision des Objekts mit dem Fremdobjekt bewirkt wird und/oder ein Hinweis auf eine bevorstehende Kollision ausgegeben wird.

A method according to the invention, in particular a computer-implemented method, for mitigating or avoiding a collision between a moving object, in particular a vehicle, and a foreign object in the vicinity of the object, the object having at least one component which can be controlled by a control device and is designed and set up for this purpose To bring about a mitigation or avoidance of a collision of the object with the foreign object and/or to issue an indication of an imminent collision comprises at least the following steps:

1. i) detection of state variables, each of which characterizes a state, a property or a movement of the object, and of state variables, each of which characterizes a state, a property or a movement of the foreign object, by means of at least one sensor device,
2. ii) carrying out a previously described method according to the invention for determining a risk of collision, and
3. iii) Controlling the at least one component of the object by means of a control device depending on the generated and output collision risk status signal in such a way that, if there is a risk of collision, a mitigation or avoidance of a collision of the object with the foreign object is effected and/or an indication of an imminent one collision is issued.

With such a method according to the invention, which includes the method steps of a method according to the invention for determining a risk of collision, an improved mitigation or avoidance of collision can be achieved, in particular by the determination of the risk of collision according to the invention. In particular, collision mitigation or avoidance better adapted to a current situation and/or state.

The component that is controlled in step iii) is particularly preferably a delay device and/or an acoustic, optical or haptic display device of the object.

If the component is a deceleration device, this is controlled in particular in such a way that the object is decelerated to such an extent that a collision between the object and the foreign object is avoided or is mitigated in the event of an unavoidable collision.

If the component is a display device, it is controlled in particular in such a way that an indication of an imminent collision is output in such a way that, for example, a driver of the object quickly recognizes from the indication that a collision is imminent and/or is prompted to mitigate or avoid a collision Initiate measures, such as changing lanes, taking evasive action, etc.

In the context of the present invention, a “deceleration device” is understood in particular to mean a device which is set up to bring about a deceleration of the object, the deceleration device being in particular controllable and a deceleration of the object being effected when the deceleration is actuated accordingly.

A deceleration device preferably has a braking device or is a braking device. As an alternative or in addition, however, the deceleration device can also have other components that bring about a deceleration of the object. For example, a drive device that can be controlled in such a way that the object can be decelerated, for example by removing the gas.

A computer program according to the invention comprises instructions which, when the computer program is executed by a computer, cause the computer to carry out at least one method according to the invention for determining a risk of collision, in a further development in particular also a method according to the invention for mitigating or avoiding a collision.

A computer-readable medium according to the invention, in particular a computer-readable storage medium, comprises instructions which, when executed by a computer, cause the computer to carry out at least one method according to the invention for determining a risk of collision, in a further development in particular also a method according to the invention for mitigating or avoiding a collision.

A collision risk monitoring system according to the invention is designed at least to determine a risk of collision between a moving object, in particular a vehicle, and a foreign object in the vicinity of the object, in particular in a further development also to mitigate and avoid a collision, and comprises at least means for executing an inventive Method for determining a risk of collision, in a further development in particular also means for carrying out a method according to the invention for mitigating or avoiding a collision.

A "means" or "module" within the meaning of the present invention can be formed by pure hardware components, i. H. in particular be a pure hardware module. However, a means or module can also include software. A means or module can in particular have one or more memories, in particular fixed memories, i. H. ROMs (Read Only Memories) or buffers and/or include one or more transmitting and receiving devices for signal and data transmission and/or processing, in particular for signal/data reception and for signal and data transmission, and for example with one or more Be connected to signal and data buses and/or have one or more digital and/or analog processing devices, in particular microprocessor units (CPU) or the like, and/or one or more programs or program modules. If a microprocessor unit is provided, it can be set up in particular to process commands that are implemented as a program stored in a memory system. Likewise, the microprocessor unit can be designed to receive input signals, in particular from a data bus or the like, to evaluate and process data and/or to emit output signals, preferably also to a data bus or the like. An existing storage means can in particular include one or more, in particular different, storage media. Both optical and magnetic storage media can be provided, or solid-state storage media and/or other non-volatile media.

A collision risk monitoring system according to the invention preferably has in particular a receiving device for receiving and providing at least some of the detected state variables, and also in particular a collision risk parameter determination module and a collision time determination module.

In a particularly preferred embodiment of a collision risk monitoring system according to the invention, this also has in particular one or more sensor devices or sensor modules for detecting the state variables to be provided and/or a transmission device for outputting and sending the collision risk state signal to a control device and/or a control device for controlling at least one component and in particular also prefers a corresponding Appropriate, controllable component, in particular a component that can be controlled by the control device, which is designed and set up to mitigate or avoid a collision of the object with the foreign object when there is a risk of collision and/or to output an indication of an impending collision.

Such a component can in particular be a controllable deceleration device for decelerating the object, for example a braking device, wherein the deceleration device can be controlled by the control device in such a way that the object can be decelerated. However, such a component can also be an optical, acoustic or haptic display device or a belt tensioner, a steering device or the like.

With the present invention, a particularly flexible collision risk monitoring system can be provided with little effort. The modular structure makes it particularly easy to provide a collision risk monitoring system with a flexibly adaptable system architecture. In particular, a collision risk monitoring system can be provided with which an improved (reduced) rate of falsely positively recognized collisions and/or an improved (increased) rate of correctly positively recognized collisions can be achieved.

A moving object according to the invention is in particular a vehicle and is designed at least to carry out a method according to the invention for determining a risk of collision, in a further development in particular also for mitigating and avoiding a collision, and/or has a collision risk monitoring system according to the invention.

With such an object, a risk of collision can be determined in a particularly advantageous manner and, for example, displayed to a driver or mitigated or in some cases even completely avoided, for example by automatically effecting emergency braking when a risk of collision has been detected. In this way, in particular, the operational safety of the object can be improved and improved operational comfort of the object can be achieved.

The preferred embodiments presented with reference to a method according to the invention for determining a risk of collision and the advantages thereof also apply accordingly to a method according to the invention for mitigating or avoiding a collision, to a computer program according to the invention, to a computer-readable storage medium according to the invention, to a collision risk monitoring system according to the invention and to a system according to the invention , moving object.

Further features of the invention result from the claims, the figures and the description of the figures. All the features and combinations of features mentioned above in the description and the features and combinations of features mentioned below in the description of the figures and/or shown alone in the figures can be used not only in the combination specified in each case, but also in other combinations or on their own, provided these is technically possible in principle, i.e. feasible.

The invention will now be explained in more detail on the basis of two non-limiting, possible and advantageous exemplary embodiments with reference to the attached drawings.

• 1 ein erstes Ausführungsbeispiel eines erfindungsgemäßen, beweglichen Objekts mit einem ersten Ausführungsbeispiel eines erfindungsgemäßen Kollisionsgefahrüberwachungssystem mit einem Abstand zu einem Fremdobjekt,
• 2 das erfindungsgemäße, bewegliche Objekt aus 1 in vergrößerter Darstellung,
• 3 ein Ablaufschema eines erfindungsgemäßen Verfahrens zur Kollisionsmilderung und -vermeidung,
• 4 ein Ablaufschema eines erfindungsgemäßen Verfahrens zur Bestimmung einer Kollisionsgefahr,
• 5 ein erstes Beispiel einer Situation, in der Kollisionsgefahr besteht und ein zweites Beispiel einer Situation, in der Kollisionsgefahr besteht,
• 6 ein Beispiel eines zweidimensionalen Kennfeldes zur Ermittlung eines ersten individuellen Kollisionszeitwertes in Abhängigkeit eines ersten Kollisionsgefahrparameterwertes und eines ersten Zustandsparameterwertes,
• 7 ein Beispiel eines zweidimensionalen Kennfeldes zur Ermittlung eines zweiten individuellen Kollisionszeitwertes in Abhängigkeit eines zweiten Kollisionsgefahrparameterwertes und des ersten Zustandsparameterwertes,
• 8 ein Blockschaltbild zur Visualisierung einer Ermittlung des Gesamt-Kollisionszeitpunktwertes, und
• 9 ein weiteres Ausführungsbeispiel eines erfindungsgemäßen Kollisionsgefahrüberwachungssystems mit einem alternativen Ausführungsbeispiel eines erfindungsgemäßen, beweglichen Objekts.

It shows:

• 1 a first embodiment of a moving object according to the invention with a first embodiment of a collision risk monitoring system according to the invention with a distance to a foreign object,
• 2 the moving object according to the invention 1 in enlarged view,
• 3 a flow chart of a method according to the invention for collision mitigation and avoidance,
• 4 a flow chart of a method according to the invention for determining a risk of collision,
• 5 a first example of a situation in which there is a risk of collision and a second example of a situation in which there is a risk of collision,
• 6 an example of a two-dimensional map for determining a first individual collision time value depending on a first collision risk parameter value and a first state parameter value,
• 7 an example of a two-dimensional map for determining a second individual collision time value depending on a second collision risk parameter value and the first state parameter value,
• 8th a block diagram for visualizing a determination of the overall collision time value, and
• 9 a further embodiment of a collision risk monitoring system according to the invention with an alternative embodiment of a moving object according to the invention.

1 shows a first exemplary embodiment of a moving object 20 according to the invention with a first exemplary embodiment of a collision risk monitoring system 100 according to the invention, which is shown in 2 is shown enlarged, the object 20 being a vehicle, in particular a two-lane passenger car 20, which is currently at a distance d from a foreign object 10, which is also a corresponding vehicle. In this case, the collision monitoring system 100 is integrated into the vehicle 20 and the vehicle 20 is designed to carry out a method according to the invention for mitigating or avoiding a collision.

The vehicle 20 drives at a current object speed vO with the current distance d behind the foreign object 10, which is moving at a current foreign object speed vF, so that the moving object 20, i. H. the vehicle 20 moves relative to the foreign object 10 at a current relative speed vR=vF-vO.

The object 20 according to the invention, ie the vehicle 20 according to the invention, has a sensor module 30 for acquiring status data Z1..n (see FIG 3 and 4 ), each characterizing a state of vehicle 20 itself, such as object speed vO, a position of vehicle 20 and its dimensions, and a state of foreign object 10, such as foreign object speed vF, its position and dimensions.

In this exemplary embodiment, vehicle 20 also has a collision risk parameter determination module 40, a collision time determination module 50, a control device 60, and a controllable component 70 in the form of a delay device 70, in particular a braking device 70.

In this example, the sensor module 30 comprises in particular a front sensor device 30A in the form of a front camera 30A (cf. 2 ), which is arranged in the central area of vehicle 20 (relative to a vehicle transverse direction) in the radiator grille and is designed to capture images of the area in front of vehicle 20 and transmit the captured data as state variables Z1..n to collision risk parameter determination module 40 transmit and thus provide.

The sensor module 30 also includes a left-hand corner sensor device 30B in the form of a laser scanner 30B and a right-hand corner sensor device 30C, which is also designed as a laser scanner 30C (cf. 2 ). These are arranged at the front corners of the vehicle 20, respectively. As a result, the surroundings of the vehicle 20 can also be detected in the front corner areas of the object 20 and/or can be monitored for foreign objects 10 which could pose a risk of collision. The two laser scanners 30B and 30C are also designed to generate sensor data from foreign objects 10 located in the detection area and to transmit the recorded data as state variables Z1..n to the collision risk parameter determination module 40 and thus make them available.

The collision risk parameter determination module 40 is set up to, among other things, from the detected status data Z1..n (cf. 3 and 4 ) to determine several collision risk parameter values KP1..3 characterizing a collision risk and to forward or transfer these to the collision time determination module 50 for further processing (cf. 4 ).

Furthermore, in this exemplary embodiment, the collision time determination module 50 is also set up to determine a first state parameter value ZP1 from the detected and provided state variables Z1..n (cf. 4 ), which in this case characterizes the ratio of the object speed vO to the relative speed vR = vF-vO, and also to transfer this to the collision time determination module 50 for further processing (cf. 4 ).

The collision time determination module 50 is also designed and set up for each determined collision risk parameter KP1..3 initially as a function of the associated collision risk parameter value KP1..3 and the determined state parameter value ZP1, which in this case is the ratio of the object speed vO to the relative speed vR = vF-vO characterizes an individual collision time value TTC1 ..3 using a table (cf. 6 and 7 ) to be determined, which will be explained in more detail later in this application.

The collision time determination module 50 can then determine an overall collision time value TTCtotal from the respectively determined, individual collision time values TTC1..3 (cf. 4 and 8th ), which is then evaluated by the collision time determination module 50.

According to the invention, a collision risk status signal KZS is generated as a function of this evaluation result (cf. 3 and 4 ), which in this exemplary embodiment is output directly as a control signal to the control device 60, which can control the deceleration device 70 in the event that a risk of collision is detected in such a way that a corresponding deceleration, particularly suitable for mitigating or avoiding a collision, is effected.

3 shows an embodiment of a flowchart for the previously based on 1 and 2 described, inventive method 200 for collision mitigation and - avoidance, wherein the in 3 illustrated exemplary embodiment of a method 200 according to the invention comprises steps S0 to S5. Step S0 represents a start of the method and S5 an end of the method, whereby steps S0 to S5 can optionally be repeated, which in 3 is indicated by the dashed arrow.

The method according to the invention includes a step S1 of detecting state variables Z1 . . . n that characterize foreign object 10 . Step S2 symbolizes the detection of state variables Z1 . . . n that characterize vehicle 20 .

The third step S3 includes the implementation of a method 300 according to the invention, described in more detail below, for determining a collision risk, the result of step S3 according to the invention being the collision risk state signal KZS.

In step 4, deceleration device 70 of object 20 is actuated by control device 60 as a function of the status information generated and output in step S3 when determining the risk of collision in the form of the collision risk status signal KZS, with object 20 being decelerated if there is a risk of collision is controlled, in particular by the deceleration device 70 being controlled in such a way that the object 20, in this case the vehicle 20, is decelerated in such a way that a collision with the foreign object 10, in this case the vehicle 10 driving ahead, is mitigated or even completely avoided.

The method steps S1 and S2 are in 3 depicted as parallel, ie essentially simultaneous, method steps. As an alternative, however, this can also be carried out one after the other, in parallel or in any order or combination, ie partially in parallel or completely sequentially, provided it is technically sensible.

4 shows an embodiment of an associated flowchart for the individual method steps of step S3 3 for the method according to the invention for determining the risk of collision 300, the method for determining the risk of collision 300 starting with step S3-0 and ending with step S3-7.

In step S3-1, the detected status data Z1..3 or Z1..n are first provided, which are then further processed in step S3-2, with the aid of the status data Z1..3 in this exemplary embodiment, in particular both an environment model UM and an object model OM is created, with the help of which different collision risk parameter values KP1..3 for different collision risk parameters are determined from the detected and provided state variables Z1..n in step S3-3.

For example, a degree of overlap collision risk parameter value KP1 is determined, which characterizes the risk of collision based on a degree of overlap Ü determined as a function of the positions and dimensions of object 20 and/or foreign object 10 .

To illustrate the degree of overlap Ü shows 5 in the upper half a first example of a situation in which there is a risk of collision and a second example of a situation in which there is a risk of collision, the degree of overlap Ü being 100% in the upper case and only about 20% in the lower case.

Furthermore, a trustworthiness collision risk parameter value KP2 is determined, for example, which characterizes the collision risk, in particular based on a determined trustworthiness.

• - einen Positions-Kollisionsgefahrparameter,
• - einen Geschwindigkeits-Kollisionsgefahrparameter,
• - einen Beschleunigungs-Kollisionsgefahrparameter,
• - einen Dimensions-Kollisionsgefahrparameter,
• - einen Klassifizierungs-Kollisionsgefahrparameter,
• - einen Alter-Kollisionsgefahrparameter,
• - einen Eindringungsregion-Kollisionsgefahrparameter,
• - einen Bewegungstrajektorien-Kollisionsgefahrparameter, und
• - einen Bewegungsart-Kollisionsgefahrparameter.

In addition, another value KP3 is determined for a risk of collision parameter, which is selected from a group containing at least

• - a position collision risk parameter,
• - a speed collision risk parameter,
• - an acceleration collision risk parameter,
• - a dimensional collision risk parameter,
• - a classification collision risk parameter,
• - an age collision risk parameter,
• - an intrusion region collision risk parameter,
• - a movement trajectory collision risk parameter, and
• - a movement type collision risk parameter.

In addition, as already explained above, in this exemplary embodiment the state parameter value ZP1 is also determined, which in this case characterizes the ratio of the object speed vO to the relative speed vR=vF-vO.

In a further step S3-4, an individual collision time value TTC1 ..TTC3 is determined for each collision risk parameter (value) depending on the respective collision risk parameter value KP1..3 and the state parameter value ZP1. In particular, in each case with the help of a two-dimensional map in the form of a table, as shown in the example in 6 for the collision risk parameter KP1 (overlap degree collision risk parameter) and in 7 for the collision risk parameter KP2 (confidence collision risk parameter).

Depending on the value of the respective collision risk parameter KP1 or KP2 or KP3 and in this case also depending on the value of the associated state parameter value ZP1, there are different individual collision time values TTC1..3, in particular different time buffers remaining until the collision, and thus different urgencies for the Taking appropriate measures to mitigate or avoid an impending collision for the respective collision risk parameters.

With a state parameter value ZP1 in the CD range, this results in a degree of overlap Ü of 100% (see upper example in 5 ) or an overlap collision risk parameter value KP1 of 100%, an individual collision time value TTC1 according to the table in 6 from TTC1-C5 and for a degree of overlap Ü of 20% (see example below in 5 ) or an overlap collision risk parameter value KP1 of 20%, an individual collision time value TTC1 from TTC1-C1.

In this way, the fact that in the case of a degree of overlap Ü of 100% can be taken into account in a simple manner, measures to mitigate or avoid collisions have to be taken earlier than with a degree of overlap of only 20%. In particular, in the latter case, emergency braking can be triggered considerably later because the driver has more time and space for an evasive steering movement.

If the respective associated individual collision time values TTC1 ..3 have been determined in this way for all collision risk parameters - for the collision risk parameter KP2 (trustworthiness collision risk parameter), for example according to the table in 7 - In the further course of the method, in particular in a further step S3-5, according to the invention from the individual individual collision time values TTC1..3 or TTC1..n using weighting factors W1..W3 or W1..Wn (cf. 4 and 8th ) determines a total collision time value TTCtotal.

In this exemplary embodiment, an associated weighting factor W1..Wn is stored in a corresponding database for each individual collision time value TTC1..3 or TTC1..n.

It has been found that a particularly good determination of the risk of collision can be achieved if the total time-of-collision value TTCtotal is determined using a classification and/or regression method, using an artificial neural network, using fuzzy logic or using a combination thereof.

In the example described here, the total collision time value TTCtotal is determined using a so-called "deep learning neural network" (DNN), which is 8th illustrated with a simple block diagram.

9 shows a further exemplary embodiment of a collision risk monitoring system 400 according to the invention with an alternative exemplary embodiment of a mobile object 20 ′ according to the invention, with the collision risk parameter determination module 40 and the collision time determination module 50 not being arranged within the vehicle 20 ′ in this exemplary embodiment, but being part of a cloud. In other words, the collision risk monitoring system 400 is designed as a decentralized system 400 in this example.

However, the same method steps S0 to S5 are carried out and the collision risk monitoring system 400 according to the invention basically also has the same components as the collision risk monitoring system 100 according to the invention from FIGS 1 and 2 .

In this exemplary embodiment, the detected state variables Z1 . . . n are evaluated in the cloud, with the detected state variables Z1 . . Radio communication network, for example using a WLAN network or the like, from the vehicle 20 'can be transmitted to the cloud.

As in the collision risk monitoring system 100 described above, the collision time determination module 50 can also transmit a collision risk status signal KZS to the object or vehicle 20', in particular to the control device 60 of the object 20', and depending on this, the delay device 70 of the Vehicle 20 'are controlled accordingly.

This exemplary embodiment of a collision avoidance system 400 according to the invention is also designed to record at least one additional state variable Z1..n, in particular the current speed vF and a GPS position of the foreign object 10, thereby achieving an improved and also simpler determination of the current relative speed value vR in the cloud can be.

This appears particularly advantageous with regard to autonomous driving, for which collision-free driving is an essential prerequisite. A collision risk monitoring system 400 according to the invention is therefore particularly preferably designed to be integrated in a system for autonomous driving, with the collision risk monitoring system 400 being set up in particular to collect data from a number of objects 20', 10, in particular an ego object 20' and several to receive foreign objects 10 and, depending on the received data, to determine a risk of collision and to generate and output status information in order to initiate appropriate measures, for example to delay the ego object 20' and/or to delay and/or accelerate foreign objects 10 or to steer or the like.

Alternatively, other divisions of the components of the collision risk monitoring system 400 than according to the 9 division shown as an example possible.

Reference List

100, 400

collision risk monitoring system according to the invention

200

inventive method for collision mitigation or avoidance

300

inventive method for determining a risk of collision

10

foreign object

20, 20'

moving object according to the invention

30

sensor module

30A

front camera

30B

left corner sensor device / laser scanner

30C

right corner sensor device / laser scanner

40

Collision Hazard Parameter Determination Module

50

Collision Time Determination Module

60

control device

70

delay device

i.e

Distance between object and foreign object

DNN

"Deep Learning Neural Network"

KP1..n

collision risk parameter value

KZS

Collision Danger Condition Signal

S0..S5

process steps

TTC1..n

individual collision time value

TTCtotal

Total time of collision value

AROUND

environment model

Ü

overlap

OM

object model

vF

Foreign Object Velocity

vO

object speed

vR

Relative speed between foreign object and object

W1..n

weight factor

Z1..n

state variable

ZP1

state parameter value

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• US 2014/0343749 A1 [0006]

Claims (15)

Method (300), in particular computer-implemented method, for determining a risk of collision between a moving object (20, 20'), in particular a vehicle (20, 20'), and a foreign object located in the vicinity of the object (20, 20'). (10), characterized by the steps: a) providing (S3-1) detected state variables (Z1..n), each of which characterize a state, a property or a movement of the object (20, 20'), and state variables (Z1..n), each of which characterizes a state, a property or a movement of the foreign object (10), b) determining (S3-3) at least two collision risk parameter values (KP1..3) characterizing a risk of collision as a function of a or several of the provided state variables (Z1..n), c) determining (S3-4) an individual collision time value (TTC1..3) for at least one collision risk parameter, from which a collision risk parameter value (KP1..3) is determined w is ordered, depending on the associated collision risk parameter value (KP1..3), d) determining (S3-5) a total collision time value (TTCtotal), which characterizes a total collision risk, from the individual, individual collision time values (TTC1..3 ) using weighting factors (W1..3), e) evaluating (S3-6) the total collision time value (TTCtotal), and f) generating (S3-6) at least one collision risk status signal (KZS). Method (300) according to claim 1 , wherein in a further step (S1, S2), in particular before step a), the state variables (Z1..n) to be provided for determining the respective collision parameter values (KP1..3) are recorded. Method (300) according to claim 1 or 2 , wherein the at least two collision risk parameter values (KP1..3) determined in step b) (S3-3) are values of collision risk parameters that are selected from a group containing at least - a position collision risk parameter, - a speed collision risk parameter , - an acceleration collision hazard parameter, - a dimension collision hazard parameter, - a classification collision hazard parameter, - an age collision hazard parameter, - an overlap degree collision hazard parameter, - an intrusion region collision hazard parameter, - a movement trajectories collision hazard parameter, - a movement type collision hazard parameter, and - a confidence collision risk parameter. Method (300) according to one of the preceding claims, wherein at least one collision risk parameter value (KP1..3) is determined as a function of an environment model (UM) and/or an object model (OM). Method (300) according to one of the preceding claims, wherein at least one state parameter value (ZP1) is also determined in step b) (S3-3) as a function of one or more of the provided state variables (Z1..n). Method (300) according to claim 5 , wherein at least one state parameter value (ZP1) that is determined is a value for a state parameter that is the ratio of a current object speed (vO) to a current foreign object speed (vF) or a current relative speed (vR = vF-vO) between object ( 20, 20') and foreign object (10). Method (300) according to claim 6 , whereby at least one individual collision time value (TTC1..3) is determined not only as a function of the respective associated collision risk parameter value (KP1..3), but also as a function of the determined state parameter value (ZP1), which represents the ratio of a current object speed (vO ) to a current foreign object speed (vF) or to a current relative speed (vR=vF-vO) between the object (20, 20') and the foreign object (10). Method (300) according to one of the preceding claims, wherein at least one individual collision time value (TTC1..3) is determined using a stored characteristic map. Method (300) according to one of the preceding claims, wherein the total collision time value (TTCtotal) is determined from the individual collision time values (TTC1..3) in step d) using predetermined, permanently stored weighting factors (W1..n). Method (300) according to one of the preceding claims, wherein the determination of the total collision time value (TTCtotal) in step d) using a classification and / or regression method, using an artificial neural network, using fuzzy logic or using a combination thereof. Method (200), in particular a computer-implemented method, for mitigating or avoiding a collision between a moving object (20, 20'), in particular a vehicle (20, 20'), and a vehicle in the vicinity of the object (20, 20') located foreign object (10), wherein the object (20, 20 ') has at least one by means of a control device (60) controllable component (70) which is designed and set up to mitigate or avoid a collision of the object (20, 20' ) to effect with the foreign object (10) and / or to issue an indication of an imminent collision, characterized by the steps: i) detecting (S1, S2) of state variables (Z1..n), each a state, a property or characterize a movement of the object (20, 20'), and state variables (Z1..n), each of which characterize a state, a property or a movement of the foreign object (10), by means of at least one sensor device (30), ii) by perform a method (300) (S3-0 to S3-7) according to any one of Claims 1 until 10 , and iii) controlling (S4) the at least one component (70) of the object by means of a control device (60) as a function of the generated collision risk status signal (KZS) in such a way that, if there is a risk of collision, a mitigation or avoidance of a collision of the object ( 20, 20') with the foreign object (10) and/or an indication of an imminent collision is output. Computer program, comprising instructions which, when the computer program is executed by a computer, cause the computer to use a method (300, 200) according to one of Claims 1 until 11 to execute. Computer-readable medium, in particular computer-readable storage medium, comprising instructions which, when executed by a computer, cause it to carry out a method (300, 200) according to one of Claims 1 until 11 to execute. Collision risk monitoring system (100, 400), at least for determining a collision risk between a moving object (20, 20'), in particular a vehicle (20, 20'), and a foreign object located in the vicinity of the object (20, 20') ( 10), comprising means (30, 40, 50, 60, 70) for carrying out a method (300, 200) according to one of Claims 1 until 11 . Movable object (20, 20 '), in particular vehicle, wherein the object (20, 20') for carrying out a method (300, 200) according to one of Claims 1 until 11 is formed and / or a collision risk monitoring system (100, 400). Claim 14 having.

Images