阅读说明：本技术 求取车辆停留在行车道的车道上的停留概率的方法和设备 (Method and device for determining the probability of a vehicle staying in a lane of a roadway ) 是由 P·韦特 于 2021-04-13 设计创作，主要内容包括：本发明涉及用于求取车辆停留在行车道的多个车道中的一个车道上的停留概率的方法。方法具有读入步骤,其中读入：车道数据,该车道数据代表车辆行驶的行车道的车道的数量和/或在车道之间的变道可行性；和变道数据,该变道数据代表车辆的由变道传感器识别的车道变换；和行车道界限类型数据,该行车道界限类型数据代表车辆当前使用的车道的由行车道界限类型传感器识别的行车道界限的类型。方法还包括步骤：将车道数据、变道数据以及行车道界限类型数据关联,以确定车辆针对行车道的多个车道中的每个车道的停留概率。最后,方法包括步骤：选择具有所确定的最高的停留概率的车道作为车辆当前所位于的车道。(The invention relates to a method for determining a stopping probability of a vehicle stopping in one of a plurality of lanes of a traffic lane. The method has a reading step, in which: lane data representing the number of lanes of a traffic lane on which the vehicle is traveling and/or lane change feasibility between lanes; and lane-change data representing a lane change of the vehicle recognized by the lane-change sensor; and lane boundary type data representing the type of lane boundary of the lane currently used by the vehicle as identified by the lane boundary type sensor. The method further comprises the steps of: the lane data, lane change data, and lane boundary type data are correlated to determine a probability of a vehicle stopping for each of a plurality of lanes of the lane of travel. Finally, the method comprises the steps of: the lane with the highest determined probability of stopping is selected as the lane in which the vehicle is currently located.)

1. A method (200) for determining a stopping probability for a vehicle (100) to stop in one of a plurality of lanes (110) of a traffic lane (115), wherein the method (200) comprises the following steps:

read-in (210):

lane data (127) representing the number of lanes (110) of a traffic lane (115) traveled by the vehicle (100) and/or lane change feasibility between lanes (110); and

lane change data (137) representing a lane change of the vehicle (100) recognized by a lane change sensor (135); and

lane limit type data (141) representing the type of lane limit (143, 117) of a lane (110) currently used by the vehicle (100) as identified by a lane limit type sensor (137);

associating (220) the lane data (127), the lane change data (137), and the lane-boundary type data (141) to determine a probability of stay (145) of the vehicle (100) for each of a plurality of lanes (110) of the lane of travel (115); and is

Selecting the lane (110) having the highest determined stopping probability (145) as the lane (110) in which the vehicle (100) is currently located.

2. Method (200) according to claim 1, characterized in that in the reading-in step (210) the lane data (127) are read in from a digital map (129) and/or the lane data are read in by an optical sensor (120) and/or the lane change data (137) and/or the lane boundary type data (141) are read in by using an optical sensor (120).

3. The method (200) according to any one of the preceding claims,

in a reading-in step (210), previous position data (152'), previous lane data (127'), previous lane change data (137') and/or previous lane boundary type data (141') are also read in, and wherein

In a correlation step (220), the previous position data (152'), the previous lane data (127'), the previous lane change data (137') and/or the previous lane limit type data (141') are correlated with the lane data (127), the lane change data (137) and/or the lane limit type data (141) to determine the stopping probability (145),

wherein the previous position data (152') represent the position of the vehicle (100) on the traffic lane (115) at a past point in time (i-1),

the previous lane data (127') representing the number of previous lanes (100') of traffic lanes (115) traveled by the vehicle (100) at past points in time and lane change feasibility between the previous lanes (110'),

wherein the previous lane change data (137') represents a lane change of the vehicle (100) identified by the lane change sensor (135) at the past point in time, and

the previous lane limit type data (141') represents the type of lane limit (143, 117) of the lane (110) used by the vehicle (100) at the past point in time, which is recognized by the lane limit type sensor (137).

4. A method (200) according to claim 3, characterized in that in the reading-in step (210) previous position data corresponding to the lane (110) with the highest stopping probability (145) in the past selection method steps are read in.

5. The method (200) according to any one of the preceding claims,

in a reading step (210), at least one lane change data probability (161) and/or at least one lane boundary type data probability (167) are read in, and

in a correlation step (220), the at least one lane change data probability (161) and/or the at least one lane boundary type data probability (167) are used for determining the stopping probability (145) of the vehicle (100),

wherein the lane change data probability (161) represents a probability that the lane change identified by the lane change sensor (135) did actually occur, and

wherein the lane boundary type data probability (167) represents the probability that the lane boundary (143, 117) recognized by the lane boundary type sensor (137) corresponds to the type of lane boundary (143, 117) that is actually present.

6. The method (200) according to claim 5, characterized in that in the reading step (210) the at least one lane change data probability (161) is read in from a predetermined lane change data probability table (159) and/or the at least one lane limit type data probability (167) is read in from a predetermined lane limit type data probability table (165).

7. The method (200) according to claim 6, wherein in the reading step (210) the predetermined lane-change data probability table (159) is selected from a plurality of different lane-change data probability tables (159) depending on at least one environmental parameter, and/or the predetermined lane-limit type data probability table (165) is selected from a plurality of different lane-limit type data probability tables (165) depending on at least one environmental parameter, in particular wherein the environmental parameter represents a lighting condition in an environment surrounding the vehicle (100) and/or a lane road surface condition in an environment surrounding the vehicle (100).

8. The method (200) according to any one of claims 5 to 7, characterized in that, in the associating step (220), the at least one lane change data probability (161) is associated with the at least one lane boundary type data probability (167) in a multiplicative manner.

9. The method (200) according to any one of the preceding claims, wherein the steps (210, 220, 230) of the method (200) are performed repeatedly periodically, in particular at a frequency of not more than 1 hertz.

10. A device (105) configured to perform and/or to control the steps (210, 220, 230) of the method (200) according to any one of the preceding claims in a corresponding unit (131, 133, 150).

11. A computer program configured to perform and/or control the steps (210, 220, 230) of the method (200) according to any one of the preceding claims.

12. A machine readable storage medium on which a computer program according to claim 11 is stored.

Technical Field

The invention relates to a method and a device for determining a probability of a vehicle staying in one of a plurality of lanes of a traffic lane. The invention also relates to a computer program.

Background

Until now, accurate lane guidance has been performed, mainly by computationally complex evaluation of the signals by very simple sensors (e.g. video-based sensors), in order to enable accurate positioning of the vehicle on a map. Alternatively, expensive techniques (high precision GNSS sensors, radar sensors, lidar sensors) may be used to position the vehicle to the centimeter level. If a highly accurate map is also provided on the vehicle, the lane can be determined. However, this also requires a high level of computing power available in the vehicle, which is not typically available on this scale, or which is to be used in real time for other driver assistance tasks, especially in the field of highly automated driving. Therefore, there is a pressing need to create a simpler solution to accurately determine the position of the vehicle on the traffic lane.

Disclosure of Invention

On this background, the method according to the invention, and also the device configured for carrying out or controlling the method, and finally the corresponding computer program are described with the solution presented here.

The solution presented here provides a method for determining a stopping probability of a vehicle stopping in one of a plurality of lanes of a traffic lane, wherein the method comprises the following steps:

reading:

lane data representing the number of lanes of a traffic lane on which the vehicle is traveling and/or lane change feasibility between lanes; and

lane change data representing a lane change of the vehicle recognized by a lane change sensor; and

lane boundary type data representing the type of lane boundary of the lane currently used by the vehicle as identified by the lane boundary type sensor;

correlating the lane data, lane change data, and lane boundary type data to determine a stopping probability of the vehicle for each of a plurality of lanes of the lane; and is

The lane with the highest determined probability of stopping is selected as the lane in which the vehicle is currently located.

The lane data can be understood here as the following information: how many lanes there are and/or which of the available lanes lane change possibilities are provided on the traffic lane in which the vehicle is currently travelling. Lane change data may be understood here as the following information from the lane change sensor: whether the lane change sensor recognizes a lane change. This information can, for example, also indicate the direction to which the vehicle is changing or be present or supplemented in the form of a probability which gives the following indication: how high the probability is for the lane change sensor itself to recognize the actual lane change. The lane boundary type data can be understood here as information about the type of lane boundary. For example, such information about the kind or type of lane boundaries in the form of solid lines, dashed lines, lawn edges, sandbars, etc. may be formed by lane boundary type data. Correlation can be understood as a mathematical correlation, for example logically, algebraically or algorithmically. In the associating step, the stopping probability of the vehicle for a plurality of lanes available on the traffic lane (currently driven by the vehicle) may thus be determined.

The approach presented herein is based on the following recognition: by correlating the information that can be provided by sensors that are already present and/or low-cost, a very accurate recognition of the position of the vehicle on one of the possible lanes of the traffic lane can be achieved technically very simply. Thus, for the determination of the position, a certain uncertainty can be allowed, which can be kept very low by processing different kinds of information, which are preferably sought by different sensors or by evaluation of the information of the sensors. In this way, the numerical or circuit complexity for the highly accurate determination of the position of the vehicle on the traffic lane can be kept very low, so that advantages are achieved compared to previously known solutions.

In a particularly advantageous embodiment of the solution proposed here, in the read-in step, the lane data are read in from a digital map and/or the lane data are read in by an optical sensor and/or the lane change data and/or the lane boundary type data are read in by using an optical sensor. An advantage of this embodiment of the solution proposed here is that by reading the lane data from the digital map, information which is already largely already quite accurate can be obtained from sources which are already generally available. Such a digital map may be carried on a vehicle, for example, for use in a vehicle navigation system. Alternatively or additionally, the digital map stored in the online database may also be accessed, for example, via a communication connection, for example via an internet connection, or may be loaded into the vehicle, in particular in an excerpted manner. However, it is also conceivable to use signals from optical sensors (for example cameras) for the determination of lane data, wherein it should be noted in this case that modern camera systems can also already classify traffic lanes into different lanes with relatively simple technical algorithms. Previous position data can be understood as information, for example, about the position of the vehicle at a previous point in time (i.e. past in time relative to the observation point in time), here the past point in time or the previous time on a particular lane of the traffic lane. Similarly, the previous lane data may correspond to the previously mentioned information for the lane data at the previous point in time (i.e., when the vehicle is traveling on the traffic lane at the previous point in time). Likewise, the previous lane limit type data may correspond to the previously mentioned information for the road limit type data at the time of the previous point in time, i.e. to the type of the lane limit at which the vehicle was travelling on the lane at the previous point in time. An advantage of this embodiment of the method proposed here is that, by taking into account temporally preceding information, i.e. iteratively calculating the stopping probability, a further improvement in the accuracy is achieved when determining the stopping probability of the vehicle on one of the lanes of the traffic lane.

According to a further embodiment of the solution proposed here, previous position data, previous lane change data and/or previous lane boundary type data can be read in the read-in step, and wherein in the correlation step the previous position data, previous lane change data and/or previous lane boundary type data are correlated with the lane data, lane change data and/or lane boundary type data to determine the stopping probability, wherein the previous position data represent the position of the vehicle on the lane at a past point in time, the previous lane data represent the number of previous lanes of the lane traveled by the vehicle at a past point in time and the lane change feasibility between the previous lanes, wherein the previous lane change data represent the lane change of the vehicle recognized by the lane change sensor at the past point in time, and the previous lane boundary type data represents the type of lane boundary recognized by the lane boundary type sensor for the lane used by the vehicle at the past point in time.

In a further advantageous embodiment of the solution proposed here, in the reading-in step, previous position data corresponding to the lane with the highest stopping probability in the past selection method steps are read in. An advantage of this embodiment of the method proposed here is that by using the previous position data corresponding to the lane with the highest stopping probability in the past selection method steps, a very accurate, contiguous iterative calculation of the vehicle position is achieved. In this way, a further improvement of the prediction quality of the position determination of the vehicle on the available lane can be achieved with technically simple means.

In a particularly advantageous embodiment of the solution proposed here, in the reading step, at least one lane change data probability and/or at least one lane boundary type data probability is read in, and in the correlating step, the at least one lane change data probability and/or the at least one lane boundary type data probability is used to determine the stopping probability of the vehicle. Here, the lane change data probability may represent a probability that the lane change recognized by the lane change sensor actually occurs. The lane boundary type data probability may represent a probability that the lane boundary identified by the lane boundary type sensor corresponds to the type of lane boundary that actually exists. The lane change data probability can be understood as the following value, for example: a probability that an actually completed lane change is recognized as actually completed and/or a probability that an actually completed lane change is not recognized or a lane change that is not completed is recognized as a lane change. Similarly, the lane boundary type data probability can be understood as the following value: the probability that a type of actually present lane boundary (for example, on the left and/or right of the lane currently being traveled by the vehicle) is correctly recognized, or the probability that a type of actually present lane boundary is classified as another type of lane boundary or the probability that a type of lane boundary that is different from the type of actually present lane boundary is classified as the lane currently being traveled by the vehicle. An advantage of this embodiment of the solution proposed here is that, by taking into account the corresponding probabilities, the complexity of the solution proposed here in terms of value and/or circuitry can be kept low, and thus sufficiently accurate results can be obtained for the probability of a vehicle stopping in one of the possible lanes of the traffic lane.

An embodiment of the solution proposed here is also conceivable in which the at least one lane change data probability is read in from a predefined lane change data probability table and/or the at least one lane boundary type data probability is read in from a predefined lane boundary type data probability table. Such a lane change data probability table may be understood as a tabular generalization or data arrangement in which a probability that an actually existing lane is correctly associated with a recognized lane or a probability that another lane is an actually existing lane is entered. Similarly, a lane boundary type data probability table can also be understood as a tabular generalization or data arrangement in which the probability is entered that the type of the actually present lane boundary is correctly associated with the type of the recognized lane boundary or that a further probability is entered that the type of the recognized lane boundary is associated with the type of the actually present lane boundary. An advantage of this embodiment of the solution proposed here is that corresponding sensors, for example for ascertaining lane change data or lane boundary type data, are classified or assigned in advance with regard to their recognition quality, so that, by using corresponding, for example predetermined tables, a significant simplification is achieved in the determination of the stopping probability of the vehicle on one of the lanes of the traffic lane.

In accordance with a particular embodiment of the solution proposed here, in the reading-in step, a predetermined probability table of the zapping data can be selected from a plurality of different probability tables of the zapping data as a function of at least one environmental parameter. Alternatively or additionally, the predetermined lane boundary type data probability table may be selected from a plurality of different lane boundary type data probability tables depending on the at least one environmental parameter. In particular, the environmental parameter can represent a lighting condition in the environment surrounding the vehicle and/or a roadway surface condition in the environment surrounding the vehicle. An advantage of this embodiment of the solution proposed here is that the detection quality of the respective sensor can be taken into account in respect of different environmental parameters, for example whether the traffic lane is wet or icy due to rain or snow or whether a worse differentiation possibility for individual lane or lane limits occurs at night. In this way, depending on the application scenario, different tables can be used for determining lane change data or lane boundary type data, so that the accuracy of the stopping probability of the vehicle on one of the lanes of the traffic lane can be further increased by technically simple means.

In a particularly easy-to-implement embodiment of the solution proposed here, in the association step, the at least one lane change data probability is associated with the at least one lane boundary type data probability in a multiplicative manner. An advantage of this embodiment of the solution proposed here is that a sufficiently precise stopping probability of the vehicle on one of the lanes of the traffic lane can be determined by very simple circuit and/or numerical means. For example, in this context, the stopping probabilities can be normalized over all currently available lanes, so that errors can be avoided in which, for example, the sum of the stopping probabilities over all available lanes must not be 1.

In a particularly reliable working embodiment of the solution proposed here, the steps of the method are repeated periodically, in particular at a frequency of not more than 1 hz. An advantage of such an embodiment of the solution proposed here is that by repeatedly determining the stopping probability of the vehicle on one of a plurality of possible lanes of the traffic lane, the position of the vehicle can be monitored at intervals of as little time as possible, so that traffic safety hazards (for example due to the vehicle accidentally driving off the current lane) are identified as early as possible and appropriate countermeasures can be taken, for example issuing warnings or actively intervening in the vehicle control system. It is also conceivable that the method proposed here can be used to check the plausibility of another method for determining the position of a vehicle on a roadway.

The method can be implemented, for example, in the control device using software or hardware or a mixture of software and hardware.

The solution presented here also provides a device configured to perform, control or implement the steps of the variants of the method presented here in a corresponding apparatus. This object is likewise achieved quickly and efficiently by these embodiment variants of the invention in the form of a device.

To this end, the device may have: at least one processing unit for processing signals or data; at least one storage unit for storing signals or data; at least one interface for a sensor or an actuator for reading sensor signals from the sensor or for outputting data or control signals to the actuator; and/or at least one communication interface embedded in a communication protocol to read or output data. The calculation unit may be, for example, a signal processor, a microcontroller, etc., wherein the memory unit may be a flash memory, an EEPROM or a magnetic memory unit. The communication interface can be designed to read in or output data wirelessly and/or wired, and the communication interface can read in or output wired data, for example, it can read this data electrically or optically from or output it to the respective data transmission line.

A device is understood here to be an electrical device which processes sensor signals and outputs control and/or data signals as a function thereof. The device may have an interface that may be designed according to hardware and/or software. In the case of a hardware-based design, the interface may for example be part of a so-called system ASIC which contains the various functions of the device. However, the interface may also be a separate integrated circuit or be at least partly composed of discrete components. In the case of a software-based design, the interface may be a software module that is present on the microcontroller, for example, together with other software modules.

Advantageously, a computer program product or a computer program with a program code, which can be stored on a machine-readable carrier or storage medium (e.g. semiconductor memory, hard disk or optical memory), is used for performing, implementing and/or controlling the steps of the method according to one of the above-described embodiments, in particular when the program product or program is executed on a computer or device.

Drawings

Embodiments of the solution presented herein are shown in the drawings and are explained in detail in the following description. Wherein:

fig. 1 shows a block diagram of a vehicle with an exemplary embodiment of a device for determining a probability of a vehicle stopping in one of a plurality of lanes of a traffic lane;

FIG. 2 shows a flow chart of an embodiment of a method for determining a probability of a vehicle stopping in one of a plurality of lanes of a roadway;

FIG. 3 illustrates an exemplary transition matrix for reflecting the correct detection of a lane change by a lane change sensor;

FIG. 4 exemplarily shows a street condition and an example connectivity of lanes between two streets;

FIG. 5 illustrates Q for the above-described calculation shown in FIG. 4ⁱA graphical representation of an example of (a);

FIG. 6 illustrates an exemplary transition matrix for a lane marking type recognition sensor to reflect proper detection of the recognition quality of the type of lane marking by the lane marking type sensor; and

fig. 7 shows a diagram representing the flow of information for an embodiment of the scheme introduced herein.

Detailed Description

In the following description of advantageous embodiments of the invention, the same or similar reference numerals are used for the elements shown in the respective drawings and have similar functions, wherein repeated descriptions of these elements are omitted.

Fig. 1 shows a block diagram of a vehicle 100 with an exemplary embodiment of a device 105 for determining a stopping probability of the vehicle 100 on one of a plurality of lanes 110 of a traffic lane 115. As shown in fig. 1, the traffic lane 115 includes two lanes 110a and 110b, with the vehicle 100 located in the right lane 110 a. The right lane 110a is delimited on the right by a solid lane boundary 117a for marking the lane edge and on the left by means of a dashed line 117b in order to separate the right lane 110a from the left lane 110 b.

The vehicle further comprises an optical sensor 120, which is designed, for example, in the form of a single camera or a stereo camera and which provides an optical imaging of the traffic lane 115 or of the traffic lanes 110a and 110b arranged thereon as an image signal 122. According to this exemplary embodiment, the image signal 122 is fed to a lane data unit 125, in which lane data 127 are ascertained, which represent the number of lanes 110 of the traffic lane 115 traveled by the vehicle 100 and/or the lane change availability between the lanes 110. Alternatively or additionally, the lane data unit 125 may read data from the digital map 129 that represents lane data 127, which also represents the number of lanes 110 of the traffic lane 115 traveled by the vehicle 100 and/or lane change feasibility between the lanes 110. The lane data 127 are read in via the interface 131 of the device 105 and fed to the associating unit 133.

Furthermore, the vehicle 100 comprises a lane change sensor 135, which reads the image signal 122 from the optical sensor 120 and thereby generates lane change data 137, and which is provided to the combining unit 133 via the interface 131, wherein the lane change data 137 represents a change of the lane 110 of the vehicle 100 (i.e. a lane change) identified by the lane change sensor 135. For example, the lane-change sensor 135 may identify: whether one of the lane markers 117a or 117b was transmitted by the vehicle 100, and lane change data 137 may be generated based thereon.

The vehicle 100 further comprises a lane limit type sensor 139, which is designed to read in the image data 123 and to generate therefrom lane limit type data 141, which represent the type of lane limit 143 of the lane 110 currently traveled by the vehicle 100, which is recognized by the lane limit type sensor 139, for example the solid lane marking 117a here. The lane limit type data 141 are here likewise fed to the association unit 133 of the device 105 via the interface 131.

In the associating unit 133, the lane data 127, the lane change data 137 and the lane-boundary type data 141 are associated with each other to determine a stopping probability 145 of the vehicle 100 on one of the plurality of lanes 115 of the lane 100. For example, a first stopping probability 145a is determined, which represents the probability that the vehicle 100 is located in the right lane 110 a. Similarly, a second stopping probability 145b is determined, which represents the probability that the vehicle 100 is located in the left lane 110 b. These stopping probabilities 145 are fed to a selection unit 150, in which the lane 110 with the highest determined stopping probability is determined as the lane in which the vehicle 100 is currently located. This lane is encoded in the corresponding lane signal 152 as the current position of the vehicle 100 and forwarded, for example, to a driver assistance system 155 of the vehicle 100 for further processing. For example, the driver assistance system 155 may be a lane guidance system or a navigation system, which determines a steering intervention or a driving route suggestion depending on the lane 110 in which the vehicle is currently located and outputs it to an automatic steering assistance system of the vehicle 100 or to the driver of the vehicle 100.

Furthermore, the device 105 comprises a first storage unit 157 in which a probability table 159 of the zapping data is stored and from which at least one probability 161 of the zapping data is loaded into the associating unit 133 via the interface 131. The lane change data probability table 159 contains a plurality of lane change data probabilities 161 that reflect the quality of lane changes identified by the lane change sensor 135, for example, by: this probability 161 reflects whether lane change data recognized by the lane change sensor 135 actually occurred or with what probability a lane change that did not occur is evaluated as a lane change that actually occurred.

Similarly, the device 105 further comprises a second storage unit 163 in which a traffic lane limit type data probability table 165 is stored and from which at least one traffic lane limit type data probability 167 is loaded into the association unit 133 via the interface 131. The lane boundary type data probability table 165 likewise contains a plurality of lane boundary type data probabilities 167 which reflect the quality of the identification of the type of lane boundary by the lane boundary type sensor 139, for example, by: these probabilities reflect whether the type of traffic lane limit 143 detected by the traffic lane limit type sensor 139 corresponds to the type of actually present traffic lane limit 143 or is interpreted as a (wrong-) different type of traffic lane limit 143.

Then, in the association unit 133, the stay probability 145 is also found by using the lane change data probability 161 and the lane limit type data probability 167.

It is also conceivable to provide a third memory 169, in which the signals fed by the correlation unit 133 or determined by the correlation unit 133, i.e. the lane data 127, lane change data 137, lane boundary type data 141, lane change data probability tables 159, lane change data probabilities 161, lane boundary type data probability tables 165 and/or lane boundary type data probabilities 167 or the stopping probabilities 145 for the respective lane 110 are read in as input signals 171 and stored for a certain period of time. Alternatively or additionally, the current position of the vehicle 100 encoded in the respective lane signal 152 may also be stored as a previous position signal of the position of the vehicle 100 in the third memory 169 as a previous lane signal 152' and used for a subsequent association step in the association unit 133. After this time period, for example when a new integration period is taken for the determination of the stopping probability 145, the signals stored therein can be loaded from the memory 169 as previous values into the correlation unit for the determination of the current stopping probability 145, i.e. as lane data 127', previous lane change data 137', previous lane boundary type data 141', previous lane change data probability tables 159', previous lane change data probabilities 161', previous lane boundary type data probability tables 165', previous stopping probability 145 'and/or as previous lane boundary type data probability 167' for the respective lane 110 and taken into account for the determination of the stopping probability 145. In this way, by determining the stopping probabilities 145 of the vehicle on one of the lanes 110 of the traffic lane 115 in time sequence or continuously, these stopping probabilities 145 can be determined very precisely.

It is also conceivable to store in the first memory 157 a plurality of lane change data probability tables 159 which are associated with the probability of a lane change being correctly recognized by the lane change sensor 135 for different environmental conditions in which the vehicle 100 is currently located, for example on a dry or wet traffic lane or during the day or at night. For example, a matching table 159 or a matching probability 161 is then selected in the interface 131 and/or the association unit 133 for the respective current environmental conditions around the vehicle 100 and used for determining the current stay probability 145, respectively. Similarly, a plurality of traffic lane limit type data probability tables 165 may also be present in the second memory 165, which tables are associated with probabilities of correct recognition of the type of traffic lane limit by the traffic lane limit type sensor 139 for different environmental conditions in which the vehicle 100 is currently located, for example on dry or wet traffic lanes or during the day or at night. For example, a matching table 165 or a matching probability 167 is then selected in the interface 131 and/or the association unit 133 for the respective current environmental conditions around the vehicle 100 and used for determining the current stay probability 145, respectively.

Fig. 2 shows a flow chart of an embodiment of a method 200 for determining a stopping probability for a vehicle stopping in one of a plurality of lanes of a traffic lane. The method 200 includes step 210: reading in lane data representing the number of lanes of a lane of travel traveled by the vehicle and/or the lane change feasibility between the lanes, as well as lane change data representing a lane change of the vehicle identified by the lane change sensor, and lane boundary type data representing the type of lane boundary of the lane currently used by the vehicle identified by the lane boundary type sensor. The method 200 further comprises the following step 220: the lane data, lane change data, and lane boundary type data are correlated to determine a probability of a vehicle stopping for each of a plurality of lanes of the lane of travel. Finally, the method 200 includes the following step 230: the lane with the highest determined probability of stopping is selected as the lane in which the vehicle is currently located.

More details of determining the probability of the vehicle 100 stopping in one of the plurality of lanes 110 of the traffic lane 115 will be explained in more detail below. It should be noted here that the method proposed here contains a calculation model for sensor data fusion, the purpose of which is to locate the vehicle 100 on the street or traffic lane 115 lane precisely to the lane. In this context, systems that are able to locate vehicles with precision to the street are particularly helpful. The method described herein then determines, for example, which lane of the street the vehicle 100 is on (e.g., may also be referred to synonymously as lane 115). In this case, the data from the two sensors (here lane change sensor 135 and lane limit sensor 139) are merged with one another.

Lane change sensor 135, for example, identifies: whether the vehicle 100 has changed lanes, i.e. a change of lane 100, and may also indicate its direction (left change/right change), for example. The lane limit type detection sensor (also referred to as lane limit type sensor 139) detects the left and right lane limits 143 or lane limit markings 117 and can determine their type (solid/dashed line, curb, lawn edge, etc.) as the corresponding type of lane limit.

Neither of the sensors 135 or 139 may always work reliably, i.e., the individual measurements of these sensors 135 or 139 may be incorrect. The solution presented here combines the data from the two sensors 135 or 139 and thus achieves a higher positioning quality.

For example, to be stored in memory as a digital map 129 with a lane-accurate map, e.g., observations recorded over successive time intervals, may be used to determine the lane-accurate position of the vehicle 100 in the map 129 at any point in time.

For this purpose a pre-processing step may be performed. This determines the condition under which sensors 135 and 139 are operating, respectively. Respective transition matrices are determined for the two sensors, for example the transition matrices are stored in a first memory 157 as lane change data probability tables 159 in which corresponding lane change data probabilities 161 are stored, and in a second memory 163 as lane limit type data probability tables 165 in which corresponding lane limit type data probabilities 167 are stored. These transformation matrices indicate: one of the sensors 135 or 139 converts the frequency with the other in a given situation. For example, the vehicle 100 changes lanes to the left, but recognizes a lane change to the right or no lane change at all. The pre-processing step is performed only once before the method is performed. If necessary, the sensor 135 or 139 can also be subjected to a preprocessing step several times under different environmental conditions, for example for a wet or dry traffic lane 115 or for identifying the type of lane change or traffic lane boundary during the day or night, since under these different environmental conditions different probabilities of a correct or incorrect identification of the type of lane change or traffic lane boundary can also occur.

Furthermore, while the vehicle 100 is moving, sensor data, for example lane change data 137 and lane boundary type data 141 are queried periodically (for example once per second), for example, and the position of the vehicle 100 at the lane level or the current lane 110 is determined using these data as lane data 127 and preferably using map information. To this end, a probability 145 that the vehicle 100 is located in each lane 110 is calculated for that lane 110. Here, the method runs iteratively: the probabilities 145' or other previous data determined in the previous or previous iteration (i.e. the previous data stored in the third memory 169) are used as a starting point for the next (i.e. temporally subsequent) iteration. An important aspect of the approach presented herein is the calculation of the (stay) probability 145.

With the solution presented here, the lane 110 of the vehicle 100 can be determined very cost-effectively, wherein it can be constructed on the basis of the mostly already existing sensor signals.

In this case, it can be seen as an important aspect of the solution presented here that the stopping probability is calculated on all lanes of the street currently driven by the vehicle on the basis of the sensor information of the lane change recognition and the lane marking type recognition. For example, the following pattern is followed in an iterative manner:

1. acquiring lane data from a map for a current position;

2. receiving sensor data;

3. calculating a stopping probability of the lane at the current position from the sensor data of the lane change recognition;

4. calculating a stopping probability of the lane at the current position from the sensor data of the lane mark type recognition;

5. a current lane of the vehicle is determined.

Exemplary embodiments of each step are described in more detail below:

obtaining tracking data from a map for a current location

The manner in which the street-accurate location of the vehicle 100 on the digital map 129 is determined may be performed in different ways. The solution presented here is constructed, for example, on the basis of the following facts: there is information in the digital map 129 for each street about the number of lanes available on the street and its lane boundary marker type (e.g., solid line, dashed line, curbstone, etc.). In addition, the map 129 shows how lanes from different streets are connected to each other, i.e. connectivity or lane change possibilities between the individual lanes 110. If the street-accurate location of the vehicle 100 is sought, the type of lane boundary and connectivity can be read from the map 129.

Receiving sensor data

The vehicle may find the lane boundary marker type corresponding to the lane boundary marker 117 closest to the vehicle 100. These are identified below with a left lane marker 117b and a right lane marker 117a, respectively, in the direction of travel. Further, the vehicle 100 may seek when to change lanes. For each lane change, the direction, i.e. in particular the left/right in the direction of travel, can also be recognized.

Calculating a stopping probability of a lane for a current position from sensor data of lane change recognition

Fig. 3 shows an exemplary transformation matrix W for reflecting the correct recognition of a lane change by lane-change sensor 135, wherein this transformation matrix W corresponds, for example, to lane-change data probability tables 159 with individual lane-change data probabilities 161 stored in first memory 157. From W the uncertainty of the sensor can be obtained. Here, the rightward shift is described by the symbol R, the leftward shift is described by the symbol L, and the unrecognized lane shift is described by the symbol K. Thus, the matrix (which describes in the columns how the actual transformation is) illustrates: the probability of an event occurring is provided on the premise that the sensor has made a particular observation (identified transformations are described in these rows). Hereinafter, we use W [ a; b ] represents the probability that event b occurred, but that event a was detected by the sensor. W [ R (right); k (nothing) ] is here 5%, thus stating: although the sensor has identified a lane change to the right, the probability that a lane change will not occur is 5%.

In this step of the method, the calculated input is the result P of the calculation of the stay probability 145 from the last iteration (i-1)^i-1And connectivity K of the lane 110 between the last iteration (i-1) and the stopping location (street 115) of this iteration (i)ⁱ. Order Sⁱ＝{sⁱ ₁；...；sⁱ _nIs the number of lanes in the digital map in iteration i. Kⁱ[s^i-1 _u，sⁱ _v]Here, it is explained that: in the lane s^i-1 _u∈S^i-1(from the last iteration) and lane sⁱ _v∈SⁱWhether or not there is connectivity between (from the current iteration). If at most one lane change is necessary (and legal) where there is connectivity between the two lanes 110 in order to switch between lanes. If in lane s^i-1 _uAnd sⁱ _vThere is connectivity between, then Kⁱ[s^i-1 _u，sⁱ _v]1, otherwise Kⁱ[s^i-1 _u，sⁱ _v]0. When only the lane change recognition sensor is used, the calculation target Q at this stepⁱThe probability of stay in iteration i is illustrated. At iteration i, after the iteration is repeated until X ∈ { left; right; none, the probability of stay Q when a lane change is identifiedⁱ＝{qⁱ _1,…,qⁱ _nNow, for each lane u, can be calculated as follows:

wherein

Fig. 4 exemplarily shows a street condition. Here, an exemplary connectivity of a lane between two streets. In this case, the connectivity from the previous lane 110' to the lane 110-drivable in the next iteration step is as follows: 1 → {1, 2, 3}, 2 → {2, 3, 4}, 3 → {3, 4, 5 }. The lane of the stopping location of the last (previous) iteration (i-1) is in the lower region of the displayed lane 110' and the lane of the current (i) stopping location is in the upper region. The arrows illustrate: lane change sensor 135 has detected a lane change to the right.

FIG. 5 shows the above calculation QⁱA graphical representation of the example presented in fig. 4.

Calculating a stopping probability of a lane for a current position from sensor data of a lane marking type identification

From the digital map 129, the current stopping position for the vehicle 100 is known: which lane 110, which lane marker types 117 are used for the traffic lane 115. In addition, a conversion matrix M for the lane marker type recognition sensor is known.

Fig. 6 shows an exemplary transformation matrix M for the lane marking type detection sensor 139 for reflecting a correct detection of the quality of the type of lane boundary detected by the lane boundary type sensor 139, wherein the transformation matrix M corresponds, for example, to a lane boundary type data probability table 165 stored in the second memory 163 with lane boundary type data probabilities 167. Here, it is indicated by the sign N that the traffic lane mark 117 is not recognized, the dashed traffic lane mark 117 is indicated by the sign D, and the solid traffic lane mark is indicated by the sign S. Thus, the matrix (the actual type of travel to marker is shown in the columns) illustrates: the probability of an event occurring is provided on the premise that the sensor has made specific observations, in which rows the categories of recognition of the lane markings are described. Hereinafter, we use M ═ a; b ] represents the probability that event b occurred, but that event a was detected by the sensor. M [ D; n ═ 10% here, so we state: although the lane marker 117 is not present, the probability that the sensor has detected a dashed lane marker change is 10%.

Let T be { T }₁；…；t_mNow is the number of different lane marker types, e.g. "solid line" e.t. Let the identified left and right lane marker types be d for the current iteration_lE.g. T or d_rE.g. T. Let S be { S ═ S₁；...；s_nIs the number of lanes in the digital map, and leftMark: s → T is a function indicating the left lane marker type for the lane, rightMark: s → T is a function that indicates the right lane marker type for the lane. When using only lane marker type recognition sensors, the vehicle is in the lane s_xThe probability of stay on S can be calculated as: after the vehicle has identified the type d of the marking_lIs of the type leftMark(s)_x) Probability of being located on the left side of the vehicle, at which the mark type d has been identified_rIs of the type rightMark(s) under the premise of_x) Probability located on the right side of the vehicle. Alternatively, it is formulated as:

Pr(s_x)＝Pr(leftMark(s_x)|d_l)·Pr(rightMark(s_x)|d_r)

in which Pr (leftMark(s) can be read out directly from the conversion matrix M_x)|d_l) And Pr (rightMark(s)_x)|d_r)。

If it is assumed that the digital map is correct in terms of the number of lanes and the types of lane markers, it may be used when only the lane marker type recognition sensor 139 is usedTo calculate the vehicle is in the lane s_xE.g., the probability of staying on S.

Determining a current lane of a vehicle

In this step, Q is checked by means of lane marker type recognitionⁱWherein the probability of stay of the lane marker type recognition is weighted with the probability of the lane marker type recognition. Thereby defining Pⁱ＝{pⁱ ₁；...；pⁱ _nIs as

Where n represents the number of lanes on the street currently being driven. In iteration i, the current lane s_jIs correspondingly estimated as pⁱ _jLargest lane s_j。

As an alternative to the digital map 129 from which lane information is read (as the marker type and connectivity of the lane data 127), it may also be detected using the front camera 120. To this end, the image data 122 should be provided from the front-end camera 120 to the respective sensors 135 and 139, which can thereby determine the number of lanes 110, the lane marker type and the connectivity of the lanes to each other.

In a variant of an embodiment of the invention, the lane marker type recognition sensor is a (video) camera or processes data from such a camera 120. May be a front camera, a back camera, a side camera. The lane marker type may be derived from the camera image 122 by a neural network. In one embodiment variant of the invention, the lane change recognition sensor is a (video) camera or processes data from such a camera 120. May be a front camera, a back camera, a side camera. A lane change may always be recognized when the vehicle is traveling past the lane marking boundary. In the variant described here, the transformation matrix is fixed. It is contemplated that these may vary depending on the circumstances: in the dark, for example, recognition may become worse. This should be reflected in the conversion matrix M or W depending on the situation. There is a need for sensors that determine the external environment (e.g., bright, dark, rain, ice, snow, etc.), and a transformation matrix M, W that matches each of these situations. Accordingly, a corresponding matrix M, W may then be selected.

Fig. 7 shows a diagram of information flow for representing an embodiment of the scheme introduced herein. In this case, for example, lane data 127 are read in from a digital map 129, lane-change data 137 are read in from a lane-change sensor 135, lane-limit type data 141 are read in from a lane-limit type sensor 139, and the association unit 133 associates these with the aid of a mathematical model in order to obtain a stopping probability 145 for each of a plurality of lanes.

The solution presented here is particularly advantageous for use as a method for checking the reliability of a feature-based locator. It can also be used as a stand-alone positioning method, but this can lead to very high expenditure in comparison with a reliability check, which is particularly effective on street maps that are accurate to the lane.

If an embodiment includes the conjunction "and/or" between a first feature and a second feature, then this is to be interpreted as follows: the embodiment has both the first feature and the second feature according to one embodiment, and has only the first feature or only the second feature according to another embodiment.