Methods for traffic space forecasting

Abstract

Method for predicting the traffic space available to an ego vehicle (1), comprising the following steps: - At least two foreign vehicles (2,3) are recorded, - The condition of foreign vehicles (2,3) is categorized, - Movements of foreign vehicles (2,3) are predicted, - In the case of a predicted collision of the foreign vehicles (2,3), a restricted area (SP) to be avoided by the ego vehicle (1) is determined, which depends on the categorization of the foreign vehicles (2,3), characterized in that it is checked whether, in the case of a predicted collision of the foreign vehicles (2,3), a free route for the ego vehicle (1) will result in the area of ​​the current routes of the foreign vehicles (2,3) as a result of the collision.

Description

[0001] The invention relates to a method for predicting the amount of traffic space available to a vehicle. Furthermore, the invention relates to a system for carrying out such a prediction method.

[0002] German patent DE 10 2012 006 986 A1 discloses a method for traffic space prediction, which is intended to be applicable particularly to autonomous vehicles, but also to vehicles driven by a human driver. The vehicle applying the traffic space prediction method is in any case referred to as the "ego vehicle." To distinguish it from this vehicle, other vehicles using the same traffic space are referred to as "other vehicles." Within the framework of the method, the movements of the other vehicles are predicted, and predictions are made regarding potential interference between the predicted movements, which indicate that an evasive maneuver by a other vehicle is imminent. An evasive maneuver is considered a deviation from the predicted movement and leads to the definition of restricted zones within the traffic space, which are not available to the ego vehicle.As part of the traffic space prediction process, the driver of the ego-vehicle may receive a notification, for example in the form of an acoustic warning signal, regarding the available road space. If the ego-vehicle is an autonomous vehicle, intervention in its steering and / or speed control may also be possible.

[0003] German patent DE 102 44 205 A1 discloses a method for preventing vehicle collisions, in which a hazard level is determined from a degree of right-of-way violation and a collision risk level. Depending on the determined hazard level, various steps can be initiated to reduce the risk of a collision. These steps can include, in particular, interventions in the vehicle's braking control system.

[0004] DE 102006029995 A1 describes a method for detecting a critical situation in an area in front of a motor vehicle using an object detection sensor.

[0005] DE 102006002746 A1 discloses a device for controlling personal protective equipment depending on at least one environmental signal and at least one impact signal.

[0006] DE 102010009667 A1 discloses a method for determining the severity of a collision between a motor vehicle and an object.

[0007] DE 102005037961 A1 discloses a method for detecting a lateral impact point, in which physical and / or vehicle dynamic parameters are recorded and / or evaluated via appropriate sensor units.

[0008] DE 102008034229 A1 discloses a method for reducing the consequences of accidents after collisions, especially for motor vehicles in road traffic.

[0009] DE 10 2012 009 297 A1 discloses a method for assisting a driver in driving a vehicle, whereby driver instructions are issued within the vehicle depending on a predicted future potential collision risk and / or risk of subsequent collisions between the vehicle and other road users.

[0010] DE 102012107186 A1 discloses a method for detecting a hazardous situation of a vehicle using at least one sensor.

[0011] The invention is based on the objective of further developing a method for traffic space monitoring compared to the aforementioned prior art with regard to specific reaction possibilities to different traffic situations.

[0012] This problem is solved according to the invention by a method for predicting available traffic space with the features of claim 1 and by a system designed for carrying out such a prediction method according to claim 7. The embodiments and advantages of the invention explained below in connection with the method also apply mutatis mutandis to the device, i.e., the system for predicting traffic space, and vice versa.

[0013] The method assumes a traffic situation with at least three vehicles. These are the vehicle referred to as the "ego vehicle," which performs the traffic space prediction, and at least two other vehicles, also referred to as "external vehicles." The technical equipment used to perform the method is located at least partially, and preferably entirely, in or on the ego vehicle. The traffic space prediction system is therefore a vehicle-based system. The ego vehicle is a motorized vehicle, either single-track or multi-track. The external vehicles can be any type of motorized or non-motorized vehicle.

[0014] The first step in the traffic space prediction process involves the detection of at least two other vehicles by the vehicle-based sensors of the ego vehicle. The ego vehicle's sensors can be based on any physical principle. Examples include optical systems, radar, and lidar systems. Ultrasound-based systems are also suitable for short-range detection. Crucial to the further execution of the process is the categorization of the other vehicles' states by the vehicle-based prediction system. A vehicle's state encompasses both its state of motion and its design. In simpler implementations of the process, the categorization refers either exclusively to the other vehicles' state of motion or exclusively to their design.In refined versions of the prediction method, both the state of motion and the type of a foreign vehicle are included in its respective categorization.

[0015] In all cases, the movements of other vehicles are predicted, and if a collision between them is predicted, a no-go zone is determined that the ego vehicle must avoid. The geometry of this no-go zone depends on the categorization of the other vehicles.

[0016] Different types of crash scenarios can be advantageously assigned to specific categories of the other vehicles' states. Different crash scenarios can be distinguished, for example, by the angle at which the predicted trajectories of the two other vehicles intersect. Different categories of vehicle types can also be defined to establish different crash scenarios, differing, for example, in vehicle mass and / or vehicle dimensions.

[0017] Crash scenario I, for example, is defined as a collision between two unrelated vehicles that make side-to-side contact, meaning they are essentially traveling in the same direction and have only a small difference in speed. In this type of collision, it can be assumed that the two vehicles involved will continue moving essentially in their original direction of travel without skidding. However, due to the collision itself and / or reflexive reactions of the drivers of the unrelated vehicles, a significant lateral component in the vehicles' movements after the collision is also to be expected.Depending on the collision parameters, particularly the absolute and differential speeds, it may be most advantageous for the driver of the ego vehicle to use a road area significantly separate from the road area used by the other vehicles, or to specifically target the area of ​​the road where the collision occurs, as the other vehicles are expected to move away from this area very quickly. The road areas that the driver of the ego vehicle should avoid are designated as restricted areas.

[0018] A second crash scenario, which differs significantly from crash scenario I, involves a collision between a first, relatively light vehicle (a passenger car) and a second, considerably heavier vehicle (a truck), which is initially traveling much slower than the first vehicle. Furthermore, in crash scenario II, there is a significant lateral offset between the two vehicles, so that the first vehicle, the passenger car, only makes contact with the truck traveling in the same direction with part of its front. Unlike crash scenario I, in this case, it can be assumed that the passenger car will immediately go into a skid upon impact, rapidly lose speed, and move sharply to the side relative to its original direction of travel. The result is a restricted area that is significantly larger than the restricted area in the first crash scenario.

[0019] According to the invention, the system further checks whether, in the event of a predicted collision between the other vehicles, a clear path for the ego vehicle will result within the area of ​​the other vehicles' current paths. This is based on the consideration that, in some selected crash scenarios, the collision itself imparts a strong impulse to opposite sides of the vehicles, thereby creating a space between them that may offer a clear path for the ego vehicle. In this case, the crash parameters are checked and evaluated appropriately.

[0020] Furthermore, in a preferred configuration, a risk assessment is carried out for different options, in particular for different routes for the ego vehicle, and in particular the one with the lowest risk potential for the ego vehicle is selected and suggested to the driver or automatically taken.

[0021] In principle, different measures can be triggered in the ego vehicle in the various crash scenarios by determining the restricted zones and clear lanes. If the risk of the ego vehicle being involved in the accident is relatively low and the available reaction time is sufficiently long for the driver of the ego vehicle, especially due to a large distance to the point of collision between the other vehicles, then a visual and / or audible warning directed at the driver of the ego vehicle may be sufficient. However, if there is a high risk of the ego vehicle colliding with one of the other vehicles, then automatic intervention in the steering of the ego vehicle may be implemented.In the event of extreme danger to the Ego vehicle, further measures that are known in themselves can be taken automatically, such as the automatic application of the brakes, the raising of the seat backs of the Ego vehicle, the closing of any sunroof and the activation of seat belt tensioners.

[0022] Several embodiments of the invention are explained in more detail below with reference to a drawing. This drawing shows: Fig. 1. A first collision event between two other vehicles, observed from an ego vehicle. Fig. 2 in a representation analogous Fig. 1 a second collision event, Fig. 3. A table categorizes different crash scenarios.

[0023] The Fig. Figure 1 illustrates the collision event described above as Crash Scenario I. One ego vehicle is designated 1, and two other vehicles are designated 2 and 3. Vehicles 1, 2, and 3 are traveling on a highway 4, which has three lanes 5, 6, and 7 in the same direction. Ego vehicle 1 is in the left lane 5, the first other vehicle 2 is in the middle lane 6, and the second other vehicle 3 is in the right lane 7. Ego vehicle 1 is at a distance, in Fig. Vehicle 1 is not shown to scale, is behind the other vehicles 2 and 3, and is traveling slightly faster than the other vehicles 2 and 3. There is only a small speed difference between the other vehicles 2 and 3.

[0024] In a collision between the two vehicles 2 and 3, they make lateral contact over a large portion of their length. Sensors (not shown) within the ego vehicle 1, which include radar and / or optical sensors, provide data to an evaluation unit also located in the ego vehicle 1. This evaluation unit is a further component of the vehicle-based prediction system. Based on the information in Fig. The traffic situation outlined in section 1 suggests an imminent collision between the two vehicles 2 and 3. The directions of travel of the two vehicles 2 and 3 after the predicted collision are shown in [reference to diagram]. Fig. 1 is indicated by dashed arrows. Between these dashed arrows lies a restricted area SP, which the Ego vehicle 1 must avoid. As shown in Fig. As shown in Figure 1, the restricted area SP initially lies entirely within lanes 6 and 7, viewed from the point of collision. If the ego vehicle 1 has a sufficient speed advantage over the other vehicles 2 and 3, continuing to drive in the left lane 5 is the best course of action to avoid being involved in an accident as the driver of the ego vehicle 1. A corresponding signal is displayed to the driver inside the ego vehicle 1, for example, in the form of a straight-ahead arrow on a display. No intervention in the steering, braking, or engine control of the ego vehicle 1 is required.

[0025] The in Fig. The collision event outlined above corresponds to the scenario already designated as Crash Scenario II. It is similar to Crash Scenario I in that the ego vehicle 1 is traveling on the three-lane highway 4 behind the other vehicles 2 and 3 in the left lane 5. However, unlike Crash Scenario I, the second other vehicle 3 is not a passenger car, but a truck. The categorization of the second other vehicle 3 as a truck is carried out by the prediction system installed in the ego vehicle 1. In addition to the sensors already mentioned, acoustic sensors can also be used in individual cases, depending on the vehicle type and traffic situation, to categorize other vehicles 2 and 3.

[0026] In the Fig. In the traffic situation outlined in Figure 2, the foreign vehicle 2, which is traveling primarily in the middle lane 6 and partly in the right lane 7, is a passenger car and is traveling much faster than the second foreign vehicle 3, i.e., the truck. The inevitable collision between the foreign vehicles 2 and 3, which was predicted in advance by the forecasting system in the ego vehicle 1, causes the first foreign vehicle 2 to receive a strong impulse to the left, towards the left lane 5. This is in Fig. 2 is indicated by a dashed arrow, which originates from the first foreign vehicle 2 and limits the restricted area SP to one side.

[0027] In Crash Scenario II, it is not only assumed that the first foreign vehicle 2 will be thrown onto the left lane 5 after the collision with the second foreign vehicle 3, but also that the first foreign vehicle 2 will rapidly lose speed in the process. Accordingly, measures must be taken as quickly as possible in the Ego vehicle 1 to avoid entering the restricted zone SP. These measures may include, in particular, an immediate, automatic braking intervention after the detection of the Fig. 2. The traffic situation depicted, even before the collision between the foreign vehicles 2,3.

[0028] If, based on the available crash parameters, it is expected that one or both other vehicles 2, 3 will experience a strong lateral impulse in crash scenario I or crash scenario II, the system, in a preferred embodiment, checks and determines whether, in this case, the point of collision and / or the vehicle 2's previous lane 6 might even be a possible alternative route, i.e., not a restricted area SP, for the ego vehicle 1. In this case, it is expected that one or both of the colliding vehicles 2, 3 will move to the side. If such an option is detected, it is signaled to the driver of the ego vehicle 1, or the vehicle automatically assumes the route, possibly after weighing the risk to other possible routes.

[0029] In Fig. Section 3 provides a table summarizing some possible ways of classifying crash scenarios. Different types of vehicle collisions (FK) are listed in separate columns of the table. The first type of vehicle collision, designated FK1 in the table, is a collision between two cars, as shown in... Fig. Figure 1 shows another type of vehicle collision, designated FK2 in the table, which is a collision between a car and a truck, as shown in Figure 1. Fig. Figure 2 illustrates this. In the individual rows of the table, crash scenarios are differentiated according to the angle at which the other vehicles 2, 3 involved in the collision collide. A first angle range, W1, encompasses scenarios in which the two other vehicles 2, 3 are initially traveling in the same or nearly the same direction, as in the example shown in Figure 2. Fig. 1 as well as in the example after Fig. 2 is the case.

[0030] In contrast, the angle range W2 encompasses collision scenarios in which the two vehicles 2, 3 collide at an angle significantly greater than zero, for example, an angle of 45° to 90°. Other collision scenarios, also not shown in the figures, may include, for example, head-on collisions between similar or dissimilar vehicles 2, 3. The table below Fig. The three exemplary crash scenarios are designated C1 / 1, C1 / 2, C2 / 1 and C2 / 2, with the crash scenarios being based on the Fig. 1 and Fig. 2. This concerns crash scenarios C1 / 1 and C1 / 2. In each crash scenario C1 / 1, C1 / 2, C2 / 1 and C2 / 2, an adapted reaction is automatically triggered in the Ego vehicle 1, which depends in particular on the type of other vehicles 2,3 involved. Reference symbol list 1 Ego vehicle 2 foreign vehicles 3 foreign vehicles 4 motorway 5 left lane 6 middle lane 7 right lane C1 / 1 Crash Scenario C1 / 2 Crash Scenario C2 / 1 Crash Scenario C2 / 2 Crash Scenario FK1 Vehicle collision FK2 Vehicle Collision SP Restricted Area W1 angular range W2 angular range

Claims

[1] Method for predicting a traffic space available for an ego vehicle (1), comprising the following steps: - At least two other vehicles (2,3) are recorded, - Conditions of other vehicles (2,3) are categorised, - Movements of other vehicles (2,3) are forecast, - in the case of a collision prediction between the other vehicles (2,3), a restricted area (SP) to be avoided by the ego vehicle (1) is determined, which depends on the categorisation of the other vehicles (2,3), characterized by that it is checked whether, in the case of a collision between the other vehicles (2,3) being predicted, a free route for the ego vehicle (1) will result as a result of the collision in the area of ​​the current routes of the other vehicles (2,3). [2] Method according to claim 1, characterized by that the categorization of the states of the other vehicles (2,3) corresponds to different types of crash scenarios (C1 / 1,C1 / 2,C2 / 1,C2 / 2). [3] Method according to claim 2, characterized by that the different types of crash scenarios (C1 / 1,C1 / 2,C2 / 1,C2 / 2) are differentiated from each other with regard to the angle at which the trajectories of the other vehicles (2,3) meet. [4] Method according to one of claims 1 to 3, characterized by that the third-party vehicles (2,3) are categorised according to vehicle types. [5] Method according to claim 4, characterized by that the third-party vehicles (2,3) are categorised according to weight classes. [6] Method according to one of claims 1 to 5, characterized by that in the case of a collision prediction between the other vehicles (2,3), an automatic intervention in the control of the ego vehicle (1) takes place, depending on their categorization. [7] System for detecting and evaluating traffic situations, designed to carry out the method according to claim 1. [8] System according to claim 7, characterized bythat this is designed as a vehicle-based system

Images