The present invention relates to a method for ascertaining data of a traffic scenario. In addition, the present invention relates to a device for ascertaining data of a traffic scenario. The present invention also relates to a computer program product.
Vehicles driving in an automated or automatic driving manner require sensors and methods for detecting the environment. This detection of the environment is accomplished by suitable methods in such a way that the driving task is able to be carried out.
Existing methods for a scene interpretation directly utilize the sensors installed in the vehicle at the respective current instants.
Two conventional approaches for interpreting a scene are:
An object of the present invention is to provide an improved detection of a traffic scenario.
According to a first aspect of the present invention, the object may be achieved by an example method for ascertaining data of a traffic scenario, the example method having the steps:
This means that vehicles are able to profit from the wealth of experience of road users. In an advantageous manner, it is thereby possible to increase the safety while a vehicle is driven. A type of best practice aggregation is ultimately provided in this way, which takes into account behaviors of road users that are correct (“best practice”) and therefore enhances the safe driving operation of vehicles. This advantageously makes it possible to reduce the sensor expense for the vehicle.
According to a second aspect, the objective is achieved by a device for detecting a traffic scenario, the device including:
Advantageous further developments of the present method are described herein.
According to one advantageous further development of the present method, the combining and evaluating of the detected data of the environment and the behaviors of the road users is carried out inside or outside the vehicle. This provides different options for combining and evaluating the detected data.
One additional advantageous further development of the present method is characterized in that the combined and evaluated data are stored in an internal or an external digital map of the vehicle. This makes it easier to use both external and internal digital maps for the present method.
According to another advantageous further development of the present method, the combining and evaluating of the acquired data includes an averaging operation. A specific type of evaluation of the acquired data is thereby carried out.
According to another advantageous further development of the present method, the combining and evaluating of the acquired data includes an application of exclusion criteria. This provides another specific way of evaluating the acquired data.
According to another advantageous further development of the present method, at least one of the following is considered when combining and evaluating the acquired data: a local aspect, a temporal aspect, aspects pertaining to behavior patterns, and the use of external information. In this way, different aspects are taken into account when combining and evaluating acquired data.
According to another advantageous further development of the present method, the external information includes at least one of the following information: data pertaining to the weather, accident statistics, and police data. This advantageously utilizes different external information for the present method.
According to another advantageous further development of the present method, the combined and evaluated data are used for an information system and/or for a driver-assistance system of the vehicle. Advantageous application cases of the present method are thereby made available. For example, the combined and evaluated data may support a high availability of a longitudinal and/or transverse control of the vehicle.
Below, the present invention is described in detail together with further features and advantages on the basis of a plurality of figures. The figures are primarily intended to illustrate main features of the present invention and are not necessarily drawn true to scale.
Disclosed method features similarly result from correspondingly disclosed device features, and vice versa. This particularly means that features, technical advantages and embodiments pertaining to the present method result in a similar manner from corresponding embodiments, features and advantages relating to the present device, and vice versa.
Below, the term “automated vehicle” is synonymously used in the sense of a fully automated vehicle, a partly automated vehicle, a fully autonomous vehicle, and a partly autonomous vehicle.
One aspect of the present invention may particularly be understood as the creation of a database which considers a behavior of other road users and thereby contributes to a better quality of a digital map. Scene elements are utilized in the process and behavior patterns at the current and/or other point(s) in time are used by the ego vehicle and/or other vehicles. In accordance with the present invention, it is proposed to provide for the storage and aggregation of behavior patterns of vehicles and/or the interpretation of their behavior in the interaction with the infrastructure. These aspects are described in greater detail herein.
Due to the high complexity of a complete scene interpretation, conventional methods provide only a limited understanding of the scene, and thus only limited driving functions. Therefore, a comprehensive scene interpretation of automotive traffic situations, which will be necessary in the future, especially for autonomous driving, is provided.
The provided method uses a reciprocal context between the traffic infrastructure and the behavior of road users (all vehicles, pedestrians). On the one hand, the traffic infrastructure (e.g., the extension of a road) induces a specific behavior of the road users. On the other hand, when observing the behavior of road users, a specific development of the infrastructure is able to be inferred with the aid of the context (e.g., “the cars are driving on the road”). The detection range or the forecast of the extension of the current road is able to be greatly expanded when monitoring vehicles on the road.
The current behavior of a road user may be denoted as “best practice”, which describes a behavior of the road user that proves to be “correct” or “unproblematic” in the respective situation and contributes to a smooth traffic situation.
For example, one strategy for driving during the current situation may be to follow a vehicle that is driving ahead. As long as this vehicle obeys the applicable traffic laws, does not cause an accident, or in other words, implements a best practice, there is no reason (e.g., a traffic light turned red) not to trail said vehicle. As long as the vehicle driving ahead travels along the ego route, this may constitute a successful driving strategy.
If one observes the best practices of different road users in the current situation, then this may improve the interpretation of the current situation quite considerably. If the system according to the present invention remembers the best practices in a certain driving situation for a longer period of time, an expanded picture emerges of what is possible and advantageous in this particular situation in terms of behaviors and measures.
If this aspect of the present invention is expanded to apply to multiple locations and different points in time along a route a vehicle is traveling, then this may advantageously be used for driving the route. An additional expansion is achieved by linking other vehicles, which jointly cooperate in a crowd (what is known as “crowd sourcing”). A collective view of traffic situations is thereby generated or aggregated in the process.
Hereinafter, “aggregating” and “aggregation” denote compiling, combining and evaluating various items of information and contents and their storage in one or a plurality of suitable location(s). Suitable locations may be developed as digital maps, for example, which are located inside and/or outside the vehicle on a server device. In the case of an external server device, a communications device will be required in the vehicle with the aid of which the vehicle is able to communicate with the external server device and to transmit data to/from the external server device.
The information may relate to the following, for example:
The local information, for example, may relate to the following:
Temporal functions may pertain to the following, for example:
Behavior patterns or best practices, for instance, may relate to the following:
External marginal conditions may relate to the following, for example:
The other information, for instance, may describe the following:
In the mentioned collection, all enumerated information of a vehicle or a plurality of vehicles is detected by vehicle sensors (such as cameras and/or vehicle dynamics sensors) and/or radar sensors and/or navigation devices and/or further sensors, and transmitted to a combination device.
In the mentioned combination with the aid of the combination device, all items of information are compared to one another in order to arrive at the most uniform and correct image of the situation possible. The combined information is stored in a digital map based on its location information. An evaluation is carried out for this purpose in order to arrive at the correct information.
The example steps are able to be used in very many situations, a few of which are described in the following text, and they may be employed in many driver-assistance and automatic driving-function systems.
In an advantageous manner, this may be used especially for vehicles that are driving in an automated or automatic manner or for autonomously driving vehicles, which, in addition to their sensor-based environment detection, are able to utilize additional information in the form of aggregated data pertaining to best practices of other road users. Shortcomings in the area of reliability and availability of the situation awareness of traffic scenarios may be remedied in this manner.
The temporal and/or local aggregation is achieved with the aid of a second module 5.
The result of this aggregation is able to be stored in new, aggregated information 7. The information is synchronized with the aid of a synchronization process 9, based on which another aggregating situation detection 4 is able to be carried out. Aggregating situation detection 4, aggregated information 7, and synchronization process 9 may be processed or executed inside the vehicle and/or outside a vehicle, in what is referred to as the backend, for instance.
The results of second module 5 and, optionally, aggregated information 7 are combined into a situation interpretation 6 in the vehicle. It is used to derive a suitable, situation-appropriate behavior 8 for the vehicle.
Ultimately, an examination of the behavior of road users in the context of the infrastructure, external influences in the presence of temporal and/or local dependencies is carried out.
The method for a situation interpretation of the driving situation or the traffic scenario uses at least one sensor device for detecting an environment, e.g., a video camera and/or radar sensors and/or digital maps and/or locating information (e.g., GPS data) and/or further environment sensors and aggregated information from the mentioned sensor devices, for a description of the situation.
The objective is an improvement in the location- and/or time-specific driving behavior for automated and/or automatic and/or manual driving. The following aspects are being taken into account:
The provided method may make it easier to find answers to the above questions, thereby assisting in improving the interpretation of the situation of a traffic scenario, which may advantageously contribute to greater driving safety in that the situation interpretation of the traffic scenario is utilized in a specific manner (e.g., for a driver-information system, a driver-assistance system, a control system, etc. of the vehicle).
Below, examples of location-dependent traffic scenarios that are able to be detected and processed by the method according to the present invention are enumerated by way of example:
Driving situations differ considerably with regard to the respective road forms; on interstates, for instance, an evenly flowing traffic in the higher speed range is realized. Exceptions are the following events, which are able to be managed by the provided aggregating method, for example. The following lists are not to be considered complete but simply mention a few application cases by way of example:
In addition to the long-term topics that are based more on the infrastructure, there is also the following current information that might be relevant:
In addition to the interstate situations, the following additional situations and events that are able to be detected and processed by the aggregating method are encountered on highways:
On inner-city streets, the following further situations arise in addition to interstate and highway situations:
Regardless of the locality, traffic events that may have to be expected at the respective locality frequently occur, such as:
The local situations are described by the respective infrastructure and the road users that are involved. For example, elements of the infrastructure may include the following:
The road users move within the infrastructure listed above by way of example. A description of the road users may include the following features, although expansions are also possible:
The road users as a whole have an interrelationship with the infrastructure:
The current traffic flow may be allocated to individual infrastructures, such as:
The road users have the following characteristics:
Using the aforementioned observations, the current behavior (also known as action recognition) of the road users and—through a change in behavior—an intention of the user (also known as intention recognition) are able to be identified. There are observable indicators that announce said intentions, such as:
Monitoring the presence, the behavior and the intentions of the road users allows for indirect inferences in connection with the infrastructure, in the following manner:
The following time-related information may be examined when detecting and processing the respective traffic scenario:
The following external influences may be examined for the detected and processed traffic scenario:
The detection of the respective information in connection with the situation, the infrastructure and the behavior of road users and the own behavior is carried out using suitable environment sensors, it being possible to use the following sensor devices:
The mentioned aggregation uses external information (for instance accident statistics and police data) and carries out an aggregation on the basis of observations by other road users (crowd sourcing), police and highway traffic authorities.
All of the following or a selection of the following is/are aggregated:
The mentioned aggregation, i.e. the detecting of the behaviors of the road users with the aid of the sensor device, and the combining and evaluating of the acquired data of the environment, may be carried out in the ego vehicle and/or in on an external system and be correspondingly stored internally and/or externally in a memory or a plurality of memories. All of this may be employed to enable the ego vehicle to know a great number of imponderables of a route and specifically utilize them, as a result of a situation-specific aggregation of behavior patterns. In an advantageous manner, the safety during a driving operation may be considerably increased in this manner.
It is provided to sense and detect illustrated traffic scenario 100 using the provided method, the detected data being combined and evaluated so that the data ascertained in this manner are able to be used for specific purposes. For example, a driver-assistance system of a vehicle may thereby become aware of the danger potential when approaching the intersection situation of
In this case, as well, a detection with the aid of sensors, a combination and evaluation of the traffic scenario including the behavior of the road users is able to be carried out. The corresponding data are able to be shared with other road users so that future vehicles approaching traffic scenario 100 in Figure may advantageously profit from the ‘wealth of experience’ of vehicles that have already passed through the area.
In this case, as well, a detection of the traffic situation by sensors with the aid of the provided method is carried out, including a detection of the behavior patterns of road users, e.g., bus 70, pedestrian 60, 61, and this information is combined and evaluated in order to form aggregated data; the data may be used to ensure that future road users proceed with a greater level of alertness when approaching traffic scenario 100 of
A cooperative driving behavior of vehicles 30, 40, and 41: Vehicle 40 enters the traffic circle in the right/outer lane and leaves the traffic circle at the first exit, or in other words, carries out a right-turn maneuver. A further vehicle 30 enters the traffic circle in the center lane and leaves the traffic circle at the second exit, thus realizing straight-ahead driving. A further vehicle 41 enters the traffic circle in the left/inner lane and leaves the traffic circle at the third exit and thereby realizes a left-turn maneuver.
However, there is also an uncooperative driving mode of a further vehicle 42, which enters the traffic circle in the right/outer lane and permanently remains in the right/outer lane, leaving the traffic circle at the third exit. Vehicle 42 thereby realizes uncooperative turning because it crosses multiple intersections and crosses traffic lanes.
This example is meant to illustrate how many possible driving modes there may exist in certain driving situations and that all of them are part of a common practice in traffic situations. The best practices in the case of traffic scenario 100 of
The combining and evaluating of the acquired data may be accomplished in the form of averaging or in the form of defining exclusion criteria, but many other types of combining and evaluating of the acquired data are possible as well.
The provided method may advantageously be used for high-performance automatic and/or (partly) automated driving functions. The (partly) automated driving in the urban environment, on highways and on interstates is relevant in this context. However, the present method may advantageously also be used for manual driving, in which case optical and/or acoustic warning signals, for example, are output to the driver of the vehicle.
The present method advantageously makes it possible for vehicles to profit from data of other vehicles that were acquired with the aid of sensors. Ultimately, a reduced sensor expense is thereby necessary for vehicles because they profit from a sensor infrastructure of other vehicles.
In an advantageous manner, the method of the present invention may be used to provide high availability of a longitudinal and transverse control of vehicles, for example.
In a step 200, an environment of a vehicle 30, 40, 41, 42 is detected with the aid of a sensor device.
In a step 210, behaviors of road users are detected with the aid of the sensor device.
In a step 220, the detected data of the environment and the behaviors of the road users are combined and evaluated. In a step 230, the combined and evaluated data are stored.
It is of course understood that the sequence of steps 200 and 210 may be chosen as desired.
In an advantageous manner, the provided method is able to be realized with the aid of a software program using suitable program-code means, which runs on a device for ascertaining data of a traffic scenario. This allows for a simple adaptation of the present method.
One skilled in the art will modify the features of the present invention in a suitable manner and/or combine them with one another without departing from the core of the present invention.
Number | Date | Country | Kind |
---|---|---|---|
10 2017 206 343.2 | Apr 2017 | DE | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2018/057743 | 3/27/2018 | WO | 00 |