Method for predicting the size of an open area adjacent to a vehicle, data processing device

Abstract

Method for predicting the size of an open area (151) adjacent to a vehicle (100), wherein a distance between the vehicle (100) and an object is determined using an environmental sensor (130) and an actual size of the open area (150) is determined from this, characterized in that at a first time point, a first prediction (151) for the size of the open area is determined for a later second time point, that at the second time point, the actual size of the open area (150) is measured using the environmental sensor (130) and the first prediction of the size of the open area (151) is compared with the actual size of the open area (150), and that a result of the comparison is taken into account when determining a second prediction for the size of the open area (151) for a still later third time point.wherein differences between the actual size of the free space (150) and the prediction of the size of the free space (151) are determined for several time points and the determined differences are taken into account in the prediction of the free space (151) and / or in the calculation of the probabilities for different distances and / or speeds and / or accelerations of the object relative to the vehicle (100).

Description

[0001] The invention relates to a method for predicting the size of an open area adjacent to a vehicle. Furthermore, the invention relates to a data processing device configured to perform a method for predicting the size of an open area adjacent to a vehicle. State of the art

[0002] In order for vehicles to automatically execute driving functions, they must use sensor data to estimate how their environment will develop. This is achieved using predefined catalogs of road scenarios. Probabilities for these scenarios are calculated, and based on these probabilities, a road scenario is selected. A trajectory for the vehicle during the automated execution of at least one driving function is then provided based on these probabilities. Such a method is described in DE 10 2013 003 944 A1.

[0003] German patent DE 10 2009 058 488 A1 discloses a method for assisting the driver of a motor vehicle. Sensors identify points in the surrounding space that are not occupied by an object and combine them into a "free space object." Based on a prediction of the future development of this free space object, control commands for the vehicle are generated.

[0004] German patent DE 10 2013 214 233 A1 discloses an intelligent forward collision warning system that calculates the time to collision (TTC) with one or more vehicles ahead. Based on the calculated TTC values ​​and the detection of critical conditions, such as a potential lane change by another vehicle, a warning signal is generated for the driver. Depending on the situation, warning thresholds can be adjusted. Disclosure of the invention

[0005] One object of the invention is to improve a method for predicting the size of a free area adjacent to a vehicle. A further object of the invention is to provide a data processing device configured to execute such a method.

[0006] These tasks are solved using the method and data processing device of the independent patent claims.

[0007] In a method for predicting the size of an open area adjacent to a vehicle, a sensor measures the distance between the vehicle and an object, and from this, the actual size of the open area adjacent to the vehicle is determined. At a first time point, an initial prediction for the size of the open area is generated for a later time point. At the second time point, the actual size of the open area is measured using the sensor, and the measured size is compared with the initial prediction. The result of this comparison is then used to determine the prediction for the size of the open area for a third, later time point.

[0008] By comparing the actual size of an open area at a given time with its predicted size for that time, it is possible to assess whether the model used to estimate the size provides realistic values. If the comparison shows that realistic values ​​are provided, the method can be used as is to determine the size of the open area for a later time. However, if the comparison reveals a discrepancy between the actual and predicted sizes of the open area, this discrepancy can be taken into account when predicting the size of the open area for a later time.

[0009] In one embodiment, the comparison of the actual size of the open space with the predicted size of the open space is taken into account when predicting the size of the open space in such a way that, if there is a deviation of the actual size of the open space from the predicted size of the open space, the prediction of the size of the open space for the third time point is increased or decreased according to the deviation.

[0010] In one embodiment, the environment sensor determines, in addition to the distance, a velocity and / or an acceleration and / or a trajectory of the object relative to the vehicle, and the velocity and / or acceleration and / or the trajectory of the object relative to the vehicle are taken into account when predicting the size of the free space by predicting a position of the object relative to the vehicle for the third time point based on the velocity and / or acceleration and / or the trajectory of the object relative to the vehicle, and based on this predicted position, the prediction of the size of the free space is adjusted by excluding the predicted position of the object from the predicted free space.

[0011] If, in addition to the distance of an object to the vehicle, the speed and / or acceleration and / or trajectory of the object are recorded, a more accurate prediction of the size of the open area can be made.

[0012] In one embodiment, a probability for the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle is calculated to predict the size of the free area, taking into account the comparison between the predicted size of the free area and the actual size of the free area.By calculating the probability for the distance and / or speed and / or acceleration and / or trajectory of the object, different distances and / or speeds and / or accelerations and / or trajectories of the object can be taken into account by determining different movement possibilities of the object when the actual size of the free space deviates from the predicted size of the free space and assigning probabilities to the movement possibilities, using a database of movement possibilities of objects and probabilities of movement of objects.

[0013] This allows for the determination of different probabilities for the resulting size of the open space for the various conceivable movements of the object. In many driving situations, it is conceivable that an object on the road will perform different movements. By calculating the probability of these different movements, probabilities for the sizes of the open spaces dependent on these movements can be calculated. This improves the prediction of the open space size.

[0014] In one embodiment, at least two comparisons between the prediction of the size of the free space and the actual size of the free space are taken into account when predicting the size of the free space and / or when calculating the probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle.

[0015] By considering multiple comparisons between the actual size of the open space and the calculated size of the open space, the prediction of the size of the open space and / or the calculation of the probability for different movement scenarios of the object can be improved.

[0016] According to the invention, the differences between the actual free space and the prediction of the free space are determined for several times, and the determined differences are taken into account in the prediction of the free space and / or in the calculation of the probabilities for different distances and / or speeds and / or accelerations of the object relative to the vehicle.

[0017] By considering multiple comparisons of the actual with the predicted size of the open area for several times, the prediction of other sizes of the open area can be improved, as this allows for a more accurate assessment of whether the model used to estimate the size of the open area for that time provides realistic values.

[0018] In one embodiment, the differences between the actual size of the open space and the prediction of the size of the open space for different times are weighted differently.

[0019] This allows, for example, identified measurement errors to be given less weight when determining the actual size of the open area.

[0020] In one embodiment, the weighting of the difference between the actual size of the open area and the prediction of the size of the open area is lower the greater the time interval between the prediction of the size of the open area and the current time.

[0021] This allows comparisons between the predicted size of the open space and the actual size of the open space to be considered for different points in time, with earlier points in time being weighted less heavily than more recent points in time.

[0022] In one embodiment, a faulty environmental sensor is identified by evaluating a distance and / or a velocity and / or an acceleration and / or a trajectory of the object relative to the vehicle and / or by comparing the predicted size of the free area with the actual size of the free area, wherein the faulty environmental sensor provides measurement data that results in a different outcome for determining the actual size of the free area than the actual size of the free area determined based on measurement data from other environmental sensors and / or the result for the actual size of the free area based on the faulty environmental sensor does not agree with the prediction of the size of the free area.

[0023] This allows environmental sensors that provide faulty and / or unreliable measurement data to be identified. The values ​​for the actual size of the open area determined based on these sensors can then be corrected or discarded.

[0024] In one embodiment, the prediction of the size of the free space and / or the probability of the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle is passed on to a method for the automated execution of at least one driving function, wherein the method for the automated execution of at least one driving function controls a driving function based on the passed-on prediction of the size of the free space and / or probability of the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle.

[0025] This allows the procedure for executing a driving function to access the determined values ​​and take them into account. Particularly in the case of faulty sensors, a more conservative trajectory for the vehicle can then be generated.

[0026] In one embodiment, the vehicle is accelerated, decelerated and / or steered by the method for the automated execution of at least one driving function based on at least one prediction of the size of the free space and / or a probability for the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle.

[0027] In one embodiment, the method for the automated execution of at least one driving function takes into account at least a prediction of the size of the free space and / or a probability for the distance and / or speed and / or acceleration and / or the trajectory of the object relative to the vehicle when creating a trajectory for the automatic execution of the driving function by placing the trajectory in the predicted free space.

[0028] By comparing the actual size of the free space with the predicted size, a method for the automated execution of at least one driving function is able to plan a trajectory for the vehicle that takes into account the predicted free space and / or the probabilities for the various movement scenarios of objects.

[0029] In one embodiment, the variance of the difference between the actual size of the open space and the predicted size is determined during the evaluation. A quality factor for predicting the size of the open space is calculated from this variance. If the quality factor falls below a predetermined value, the method for automatically executing a driving function is adjusted so that the vehicle drives more slowly and / or maintains a greater safety distance and / or accelerates less and / or performs fewer lane changes or overtaking maneuvers than if the predetermined quality factor were exceeded.

[0030] Determining the variance of the difference is a common statistical method for evaluating deviations of a measured quantity from a predicted quantity. By calculating a quality factor and adjusting the execution of the driving function when a predefined quality factor is undershot, the vehicle is switched to a safer driving mode if the predictions of the free area size fall below a certain quality level.

[0031] Exemplary embodiments of the invention are explained with reference to the following drawings. The schematic drawing shows: Fig. 1 a vehicle with the necessary components to carry out the invention; Fig. 2 and Fig. 3 Flowcharts of different embodiments of the process; Fig. 4 a vehicle that determines a free space at a first point in time; Fig. 5 a vehicle that determines a free space at a second time; and Fig. 6 a vehicle that calculates probabilities for two movement scenarios of an object.

[0032] The Fig. Figure 1 shows a vehicle 100 with a data processing device 110, a device for the automated execution of a driving function 120, and an environmental sensor 130 in the front area 105 of the vehicle 100. In one embodiment, the vehicle 100 comprises several environmental sensors 130. The environmental sensor 130 is configured to determine the distance between an object and the vehicle 100. The environmental sensor 130 can, for example, consist of a radar or lidar module, an ultrasonic sensor, or a camera. However, other sensors with which the distance between the vehicle and the object can be determined are also conceivable. Objects can be, for example, other vehicles, buildings at the roadside, curbs, guardrails, lampposts, traffic signs, trees, construction sites, traffic jams, warning beacons, pedestrians, cyclists, animals, or other obstacles that the vehicle should ideally not touch.Each environmental sensor 130 has a defined range within which it can detect objects. If the environmental sensor 130 cannot detect an object and therefore cannot determine its distance, it can be assumed that the clear area extends across the entire range of the environmental sensor 130. The distance data from the environmental sensor 130 is processed in the data processing unit 110. By measuring the distance to one or more objects, the size of an area adjacent to the vehicle 100 and free of other objects is determined—a so-called clear area. The term "clear area" is not limited to one- or two-dimensional geometries; the three-dimensional volume that the vehicle 100 will occupy on its path of travel can also be relevant for determining the clear area.For example, a hanging obstacle above the roadway, a bridge, or a tunnel can restrict the clear space for vehicle 100. The clear space represents the area within the detection range of the environment sensor 130 that vehicle 100 can traverse without touching an obstacle. A safety zone may also be defined around the objects, which vehicle 100 should also avoid and which further restricts the clear space.

[0033] By evaluating multiple measurement points from the environmental sensor 130, the object's velocity, acceleration, and / or trajectory relative to the vehicle can be determined. Based on the determined free space, velocity, acceleration, and / or trajectory of the object relative to the vehicle, the data processing unit 110 generates a prediction for a free space at a future point in time. When predicting the size of the free space, a comparison is made between the size determined by the environmental sensor 130 and the predicted size. Alternatively or additionally, the probability of different object trajectories, and thus the probability of different free space sizes, can also be determined.The prediction for the size of the open area and / or the probability of different object trajectories, and thus for different open area sizes, is transmitted to the device for the automated execution of a driving function 120 and taken into account by the device when executing the driving function. This can be done, for example, by changing the speed and / or a steering movement. Alternatively or additionally, it is conceivable that the device for the automated execution of a driving function 120 modifies or plans a trajectory for the vehicle 100 based on the prediction for the size of the open area and / or the probability of different object trajectories.

[0034] If the predicted free space is smaller than the actual free space, then the vehicle has 100 more space available. The vehicle can react to this, for example, by accelerating.

[0035] If the predicted open space is larger than the actual open space, the vehicle will slow down because the smaller open space results in a shorter distance to objects.

[0036] It is also conceivable to provide an environmental sensor 130 in the rear area 106 of the vehicle 100 to determine a clear area behind the vehicle. Another possibility is to install the environmental sensor in a side area 107 of the vehicle 100 to determine a clear area to the side of the vehicle.

[0037] The data processing device 110 has a data input for measurement data from the environmental sensor 130 and a data output for transmitting data to a device for the automated execution of a driving function 120. The automated execution of a driving function can mean that the vehicle 100 is able to independently perform steering movements, changes in speed, lane changes, and / or overtaking maneuvers without driver intervention. Other automated driving functions are also conceivable. From the measurement data of the environmental sensor 130, which can include the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle 100, the data processing device 110 determines the size of the free space around the vehicle 100 and a prediction of the free space for a later time. The free space prediction is transmitted to a device for the automated execution of a driving function.

[0038] In one embodiment, the data processing device has a database in which possible movements of objects and probabilities for movements of objects are stored.

[0039] The Fig. Figure 2 shows a flowchart 200 of a procedure for predicting the size of an open area adjacent to the vehicle 100. In a first program step 201, a prediction of the size of the open area for a later second time is determined at a first time point. At the second time point, in a second program step 202, the actual size of the open area is determined, in particular measured, using the environmental sensor 130. In a third program step 203, the prediction of the size of the open area from the first program step 201 is compared with the actual size of the open area from the second program step 202.A result of the comparison is taken into account in a fourth program step 204 when predicting the open space for a later, third time point. This is done by adjusting the predicted size of the open space for the third time point accordingly if the actual size of the open space deviates from the predicted size. For example, if the third program step 203 shows that the predicted size of the open space from the first program step 201 is 20% larger than the actual size of the open space from the second program step 202 (i.e., the actual size of the open space is 20% smaller), then the predicted open space is also reduced by 20% in the fourth program step 204.Should the third program step 203 reveal that the predicted size of the open space from the first program step 201 is 15% smaller than the actual size of the open space from the second program step 202, the predicted open space will also be increased by 15% in the fourth program step 204. After the fourth program step 204, the program can return to the second program step 202, i.e., the determination of the actual size of the open space. This is in . Fig. 2 is represented by a dashed line.

[0040] In one embodiment, the deviation of the predicted size of the open area from the actual size of the open area is not taken into account with the same percentage in the next prediction, but with a reduced percentage. This means that, for example, if the predicted size of the open area from the first program step 201 is 30% smaller than the actual size of the open area from the second program step 202, the predicted open area in the fourth program step 204 is only increased by 20%. Similarly, it is conceivable to provide the full percentage value when the predicted size of the open area decreases, whereas only the reduced percentage is provided when the size of the open area increases. Other combinations can also be provided by those skilled in the art without departing from the scope of the invention.

[0041] The Fig. Figure 3 shows a flowchart 200 of a further embodiment of the method for predicting the size of a free area adjacent to a vehicle 100. Program steps 201 to 204 correspond to the program steps of the Fig. 2. In a fifth program step 205, the actual size of the open area is determined at the third time point using environmental sensor 130 and compared in a sixth program step 206 with the prediction from the fourth program step 204. In a seventh program step 207, a further prediction for the size of the open area is determined for a fourth time point. In the seventh program step 207, the result of the comparison of the actual size of the open area with the predicted size of the open area, as well as the comparison from the sixth program step 206 and the comparison from the third program step 203, are taken into account. If the actual size of the open area deviates from the predicted size of the open area, the prediction of the size of the open area for the fourth time point is changed accordingly.If, in the sixth program step 206, it turns out that the prediction of the open area size from the fourth program step 204 is 10% larger than the actual open area size from the fifth program step 205, i.e., the actual open area size is 10% smaller, then in the seventh program step 207 the predicted open area will also be reduced by 10%. Other adjustments to the percentage values ​​can also be provided by a specialist. Should it turn out in the sixth program step 206 that the prediction of the open area size from the fourth program step 204 is 20% smaller than the actual open area size from the fifth program step 205, then in the seventh program step 207 the predicted open area will also be increased by 20%. This is indicated by a dashed line. Fig. 3 again indicated that the second program step 202 can be continued, whereby in the second and all subsequent iterations of the fourth program step 204 the respective comparisons between the actual size of the open space and the predicted size of the open space from program steps 203 and 206 are taken into account in the same way as in the previous iteration of the fourth program step 204, and in all subsequent iterations of the seventh program step 207 the respective comparisons between the actual size of the open space and the predicted size of the open space from program steps 203 and 206 are taken into account in the same way as in the previous iteration of the seventh program step 207.

[0042] In one embodiment, comparisons from further back in time between the actual size of the open space and the measured size of the open space can also be taken into account. This would mean, in Fig. It is possible to insert three more program steps before returning to the second program step, 202. For example, the last four comparisons between the actual size of the open area and the predicted size of the open area could be taken into account. It is also conceivable to consider the comparisons determined within a specific time period, such as the last 5 to 8 seconds, in the prediction.

[0043] In one embodiment, the differences between the actual size of the open space and the predicted size of the open space are weighted more heavily the more recent the comparison. For example, with four comparisons considered, the weightings could be 40% for the most recent, 30% for the second-to-last, 20% for the third-to-last, and 10% for the fourth-to-last. Alternatively, it is conceivable to weight a number n of comparisons with a factor of 1 for the most recent (i.e., last) comparison, ½ for the second-to-last, 1 / 3 for the third-to-last, and so on, up to a factor of 1 / n for the most recent. Further weightings can be provided by those skilled in the art without exceeding the scope of the invention.

[0044] The Fig. Figure 4 shows a vehicle 100 on a roadway 101 at a first point in time, with a data processing device 110 in the vehicle 100 executing the procedure. The roadway 101 has two lanes, and the vehicle 100 is located in the left lane of the roadway 101. An environmental sensor 130 is mounted in the vehicle 100, which can detect objects in a predefined area 131 in the direction of travel in front of the vehicle 100. The predefined area 131 extends to the range of the environmental sensor 130. Within the predefined area 131, there is an object, here another vehicle 140, in the right lane of the roadway 101. The actual clear area 150, which is located in front of the vehicle 100 on the roadway 101, is shown hatched. The actual clear area is determined using the environmental sensor 130 by measuring the distance to an object, here the other vehicle 140.The actual free space 150 is limited by the other vehicle 140, since the area of ​​the right lane of roadway 101 occupied by the other vehicle 140 cannot be traversed by vehicle 100. A device for the automated execution of a driving function 120 in vehicle 100 controls vehicle 100 on a trajectory 102, which will lead vehicle 100 to a location 160. Location 160 lies on roadway 101, and vehicle 100 will reach location 160 at a second time due to the trajectory 102. A prediction of the free space for vehicle 100 is to be generated for this second time.

[0045] The data processing device 110 takes into account that the other vehicle 140 will also move between the first and second times, so that the other vehicle 140 is expected to be in area 141. The open space that the vehicle 100 predicts for the second time must therefore account for the movement of the other vehicle 140 between the first and second times. This is done by measuring the distance between vehicle 100 and the other vehicle 140 at the first time and assuming that the other vehicle 140 is moving at the same speed as vehicle 100, thus making the distance at the second time identical to the distance at the first time. A predicted open space 151 is located in the direction of travel before location 160, which vehicle 100 will reach at the second time.In this predicted open space 151, the area 141, which the further vehicle 140 will occupy at the second time, is excluded.

[0046] In one embodiment, the area 141 of the further vehicle 140 is determined at a second time by using the environment sensor 130 to determine a speed and / or an acceleration and / or a trajectory of the further vehicle 140 relative to the vehicle 100 by means of several distance measurements between the vehicle 100 and the further vehicle 140, and taking into account the speed and / or the acceleration and / or the trajectory of the further vehicle 140 relative to the vehicle 100 when predicting the free area 151.

[0047] The Fig. 5 shows lane 101 of the Fig. 4 at the second time point. Vehicle 100 has moved on to location 160. The other vehicle 140 has also moved on in the right lane, but not far enough to reach the predicted area 141. The first vehicle 100 now again uses the environmental sensor 130 to determine the actual size of the open space 150 in front of the vehicle by measuring the distance to objects on the road. Now, the actual size of the open space 150 determined in this step is compared with the size of the open space (151) predicted for this time point. Fig. 3) compared, and this comparison is taken into account in the next prediction of the open space. Since the other vehicle 140 did not move as far along the right lane of roadway 101 as predicted at the first time, the distance between vehicle 100 and the other vehicle 140 is smaller at the second time than at the first time. When predicting the open space for a third time, the data processing device 110 now takes into account that the other vehicle 140 is slower than vehicle 100. When predicting the size of the open space at the third time, the data processing device 110 can predict an even smaller distance between vehicle 100 and the other vehicle 140. This means that the open space available to vehicle 100 on the right lane of roadway 101 is predicted to be smaller at the third time than at the second time.

[0048] Alternatively, the speed and / or acceleration and / or trajectory of the other vehicle 140 relative to the vehicle 100 can also be taken into account when predicting the open space 151.

[0049] The Fig. Figure 6 shows a vehicle 100 with an environmental sensor 130 in the left lane of a two-lane roadway 101. The vehicle 100 follows a trajectory 102 in the left lane. Within the defined area 131 of the environmental sensor 130, there is an object, in this case another vehicle 140. The other vehicle 140 is in the right lane of the two-lane roadway 101, with its direction of movement slightly to the left, i.e., towards the left lane. A data processing device 110 in the vehicle 100 now calculates two probabilities for different scenarios, i.e., for different possible trajectories of the other vehicle 140. These two scenarios are also shown in the Fig.Figure 5 is shown, with two arrows from the previously described part of the figure pointing to the two possible scenarios. In the first scenario 111, the other vehicle 140 changes lanes, i.e., it moves from the right lane to the left lane of the two-lane roadway 101. In the second scenario 112, vehicle 140 has swerved to avoid a smaller obstacle in the right lane of the multi-lane roadway 101, but the other vehicle 140 remains in the right lane. The moment vehicle 100 detects the initial steering movement of the other vehicle 140 to the left, it calculates a probability for the size of the clear area in each of the two scenarios 111 and 112.One way to calculate the probabilities of the open space is that the data processing device 110 contains data in a memory that indicates how likely a lane change of the other vehicle 140 is due to the fact that the direction of movement of the other vehicle 140 is slightly to the left, i.e. towards the left lane.

[0050] To determine the probabilities of an object's possible movements, a database of possible movements of objects and probabilities of object movements can be used in particular.

[0051] Alternatively, the calculation of the probabilities of the size of the open space in the two scenarios 111 and 112 can be based on the fact that the data processing device 110 has data on lane-change probabilities at specific locations. Based on these probabilities, the device for the automated execution of a driving function 120 brakes and / or accelerates and / or steers the vehicle 100. The greater the probability of a lane change by the other vehicle 140, that is, the greater the probability of the first scenario 111, the stronger the adjustment of the trajectory 102 of the vehicle 100 will be, in that the device for the automated execution of a driving function 120 brakes the vehicle 100. The greater the probability of the lane change by the other vehicle 140, the more likely the device for the automated execution of a driving function 120 will brake the vehicle 100.

[0052] In one embodiment, when evaluating the difference between the actual size of the open space and the predicted size of the open space, the variance of the difference is calculated for various objects. The variance for a deviation of the open space due to one of n objects can be calculated using the following formula: Var(xn)=(xn,vorh−xn,mess)2

[0053] where x n, vorh the size of the predicted and x n, mess The size of the actually measured open area due to the nth object. From the variance of the difference for n objects at i time points, the formula can be used. 1Q=∑nNn∑i(xni,vorh−xni,mess)2 A quality factor is calculated for the predictions, where Q is the quality factor of the predictions, n is a numbering of the tracked objects, N na normalization factor for each of the n objects, i a numbering of the times of comparisons between the actual and the measured size of the open space, x ni, vorh the size of the predicted and x ni, mess The size of the actually measured free space is determined by the nth object at the i-th time. Q increases the better the predictions of the free space size match the actual size of the free space. The n tracked objects are the objects within the measurement range of the environmental sensors 130 that may potentially interact with the vehicle 100.

[0054] It is also possible to use a weighting factor G. i to provide for the variance of the open area as specified in the following formula: 1Q=∑nNn∑iGi(xni,vorh−xni,mess)2

[0055] This allows, for example, the influence of the deviation on the variance to be weighted by the inverse of the time that has elapsed between the time of the forecast and the time of the quality calculation. Other weighting factors are also conceivable. This means that, when calculating the quality of the forecasts, deviations that occurred more recently between the forecast and the actual open space are weighted more heavily than deviations that occurred further in the past.

[0056] In one embodiment, a different method is used instead of variance to assess the quality factor, i.e., the quality of the prediction of the open area size. Another method could be, for example, calculating a mean absolute deviation or a mean absolute deviation with respect to the median.

[0057] In one embodiment, the device for the automated execution of a driving function 120 can control the vehicle 100 more conservatively when the quality factor falls below a predetermined value. In this case, "more conservatively" means that the vehicle 100 drives more slowly and / or maintains a greater safety distance and / or accelerates less and / or performs fewer lane changes or overtaking maneuvers than with a higher Q value.

[0058] Alternatively, the device for the automated execution of a driving function 120 can plan a more conservative trajectory 102 for the vehicle 100 if Q becomes smaller than a predetermined value. In this case, "more conservative" also means that the vehicle 100 travels more slowly and / or maintains a greater safety distance and / or accelerates less and / or performs fewer lane changes or overtaking maneuvers than with a higher value for Q.

[0059] In one embodiment, a faulty environmental sensor 130 is identified by evaluating the distance and / or velocity and / or acceleration and / or trajectory of the object relative to the vehicle 100 and / or by comparing the predicted size of the open space with the actual size of the open space. A faulty environmental sensor 130 then provides measurement data that results in an open space that does not correspond to the open space based on the measurement data from other environmental sensors 130 and / or the predicted size of the open space. If this is the case, the measurement data from the corresponding environmental sensor 130 is considered faulty, and the measurement data is no longer taken into account. It is also possible that, analogous to the previous embodiment, the device for the automated execution of a driving function 120 plans a more conservative trajectory 102 for the vehicle 100.

[0060] In one embodiment, the prediction of the size of the free space and / or the probability of the distance and / or speed and / or acceleration and / or the trajectory of an object relative to the vehicle 100 is passed to a method for the automated execution of at least one driving function.

[0061] In one embodiment, the method for the automated execution of at least one driving function takes into account at least one prediction of the size of the open space and / or a probability for the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle 100 when generating a trajectory for the automated execution of the driving function. The better the prediction of the open space size matches the actual size of the open space, the less conservative the trajectory for the automated execution of a driving function needs to be. In this case, less conservative means that the vehicle 100 will drive faster and / or maintain a smaller safety distance and / or accelerate more strongly and / or perform more lane changes or overtaking maneuvers than if the prediction of the open space size did not match the actual size of the open space.

[0062] Although the invention has been described in detail by means of the preferred embodiments, the invention is not limited to the disclosed examples and other variations can be derived from them by a person skilled in the art without leaving the scope of protection of the invention.

Claims

[1] Method for predicting the size of a free area (151) adjacent to a vehicle (100), wherein a distance between the vehicle (100) and an object is determined using an environment sensor (130) and an actual size of the free area (150) is determined from this, characterized by, that at a first time point, a first prediction (151) for the size of the open space is determined for a later second time point, that at the second time point, the actual size of the open space (150) is measured using the environment sensor (130), and the first prediction of the size of the open space (151) is compared with the actual size of the open space (150), and that a result of the comparison is taken into account when determining a second prediction for the size of the open space (151) for a still later third time point, whereby differences between the actual size of the open space (150) and the prediction of the size of the open space (151) are determined for several time points, and the determined differences are taken into account in the prediction of the open space (151) and / or in the calculation of the probabilities for different distances and / or speeds and / or accelerations of the object relative to the vehicle (100). [2] Method according to claim 1, wherein the comparison between the first prediction of the size of the open space (151) and the actual size of the open space (150) is taken into account when predicting the size of the open space (151) for the third time such that, if the actual size of the open space (150) deviates from the predicted size of the open space (151), the predicted size of the open space (151) for the third time is reduced if the actual size of the open space (150) at the second time is smaller than the size of the open space (151) predicted for the second time, and the predicted size of the open space (151) for the third time is increased if the actual size of the open space (150) at the second time is larger than the size of the open space (151) predicted for the second time. [3] Method according to one of the preceding claims, wherein the speed and / or acceleration and / or trajectory of the object relative to the vehicle (100) is determined using the environment sensor (130) and the speed and / or acceleration and / or trajectory of the object relative to the vehicle (100) is taken into account when predicting the size of the free area (151) by predicting a position of the object relative to the vehicle (100) for the third time point based on the speed and / or acceleration and / or trajectory of the object relative to the vehicle (100) and adjusting the prediction of the size of the free area (151) based on this predicted position by excluding the predicted position of the object from the predicted free area. [4] Method according to one of the preceding claims, wherein for predicting the size of the free space (151) a probability for the distance and / or speed and / or acceleration and / or trajectory of the object relative to the vehicle (100) is calculated, wherein the calculation of the probability takes into account the comparison between the predicted size of the free space (151) and the actual size of the free space (150) by determining different movement possibilities of the object in the event of a deviation of the actual size of the free space (150) from the predicted size of the free space (151) and assigning probabilities to the movement possibilities, wherein a database of movement possibilities of objects and probabilities of movements of objects is used. [5] Method according to any of the preceding claims, wherein at least two comparisons between the prediction of the size of the free area (151) and the actual size of the free area (150) are taken into account when predicting the size of the free area (151) and / or when calculating the probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle (100). [6] Method according to claim 1, wherein the differences between the actual size of the open space (150) and the prediction of the size of the open space (151) are weighted differently for different times. [7] Method according to claim 6, wherein the weighting of the difference between the actual size of the open area (150) and the prediction of the size of the open area (151) is lower the greater the time interval between the prediction of the size of the open area (151) and the current time. [8] Method according to any of the preceding claims, wherein a faulty environmental sensor (130) is identified by evaluating a distance and / or a velocity and / or an acceleration and / or a trajectory of the object relative to the vehicle (100) and / or comparing the prediction of the size of the free area (151) with the actual size of the free area (150), wherein the faulty environmental sensor (130) provides measurement data which, when determining the actual size of the free area (150), results in a different outcome than the actual size of the free area (150) determined based on measurement data from other environmental sensors (130) and / or the result for the actual size of the free area (150) based on the faulty environmental sensor (130) does not agree with the prediction of the size of the free area (151). [9] Method according to any of the preceding claims, wherein the prediction of the size of the free area (151) and / or the probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle (100) is passed on to a method for the automated execution of at least one driving function, wherein the method for the automated execution of at least one driving function controls a driving function based on the passed-on prediction of the size of the free area (151) and / or the probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle (100). [10] Method according to claim 9, wherein the method for automatically executing at least one driving function based on at least one prediction of the size of the free space (151) and / or a probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle (100) accelerates, decelerates and / or steers the vehicle (100). [11] Method according to one of claims 9 or 10, wherein the method for automatically executing at least one driving function takes into account at least one prediction of the size of the free space (151) and / or a probability for the distance and / or the speed and / or the acceleration and / or the trajectory of the object relative to the vehicle (100) when creating a trajectory (102) for automatically executing the driving function by placing the trajectory (102) in the predicted free space. [12] Method according to one of the preceding claims, wherein, when evaluating the difference between the actual size of the open space (150) and the prediction of the size of the open space (151), a quality factor for the prediction of the size of the open space is calculated, wherein, if a predetermined value for the quality factor is not reached, the method for the automated execution of a driving function is adapted such that the vehicle (100) drives more slowly and / or maintains a greater safety distance and / or accelerates less and / or performs fewer lane changes or overtaking maneuvers than if the predetermined value for the quality factor is exceeded. [13] Method according to one of the preceding claims, wherein the variance of the difference between the actual size of the open space (150) and the prediction of the size of the open space (151) is evaluated when calculating the quality factor. [14] Data processing device (110) configured to perform one of the methods of claims 1 to 13.

Images