Computer-implemented method for evaluating the driving skills of a driver of a vehicle during a parking maneuver performed manually or with steering assistance, system, vehicle and computer program product

Abstract

A computer-implemented method for evaluating the driving skills of a driver of a vehicle during a parking maneuver performed manually or with steering assistance, the method comprising: (10) acquiring sensor data during the parking maneuver, wherein the sensor data represent the vehicle's surroundings; (20) detecting an available parking space based on the acquired sensor data; (30) determining a target parking position within the detected available parking space; (40) acquiring operating parameters of the vehicle during the driver's execution of the parking maneuver to the determined target parking position; (50) determining at least two key performance indicators (KPIs) based on the acquired sensor data and / or the acquired operating parameters of the parking maneuver;and determining (60) a driver skill score based on a comparison between at least two key performance indicators (KPIs) and at least one rating set (S), wherein the rating set (S) represents key performance indicators (KPIs) of several other drivers in the same vacant parking space and / or similar vacant parking spaces.

Description

[0001] The present invention relates to a computer-implemented method and a system for evaluating a driver's driving skills during a parking maneuver. The invention also relates to a vehicle with the system according to the invention and a computer program product corresponding to the method according to the invention. State of the art

[0002] Document DE 10 2022 213 674 A1 discloses a method for displaying an evaluation of a vehicle's driving maneuver, in particular a parking maneuver, comprising the acquisition of sensor data on the vehicle's surroundings and / or operating parameters of the vehicle.

[0003] Document DE 10 2015 205 156 A1 discloses a method for driver assistance in which a driving maneuver of a vehicle, performed without the use of a driver assistance system, is detected using the vehicle's sensors, and an equivalent driving maneuver, which can be performed with the support of a driver assistance system, is calculated. The detected driving maneuver and the equivalent driving maneuver are evaluated based on at least one criterion, and a result of the evaluation of the detected driving maneuver and the equivalent driving maneuver is output.

[0004] The object of the present invention is to improve the assessment of the driving skills of a driver of a vehicle during a parking maneuver performed manually or with steering assistance. Disclosure of the invention

[0005] The above problem is solved according to the invention in accordance with independent claims 1, 10, 11 and 12.

[0006] The present invention relates to a computer-implemented method and system for evaluating a driver's driving skills during a parking maneuver. The parking maneuver can be performed manually or with steering assistance. The invention enables an objective assessment of a driver's parking abilities and provides the basis for personalized assistance functions. This increases comfort and safety when parking and contributes to improving driving skills.

[0007] The system captures sensor data representing the vehicle's surroundings. Various sensor types, such as LiDAR, camera, radar, and / or ultrasonic sensors, can be used for this purpose. This sensor data forms the basis for precisely identifying available parking spaces in the vehicle's vicinity and for determining performance indicators.

[0008] Based on the acquired sensor data, the system detects an available parking space. The automatic detection of parking spaces simplifies the parking process for the driver and saves time. A target parking position is then determined within the detected available parking space. The target parking position and / or the availability of the parking space are preferably displayed to the driver, particularly via a display device in the vehicle. The target parking position advantageously represents or defines a desired parking position with a desired parking orientation, which lies within the area of ​​the available parking space and which the vehicle should ideally reach. Determining or defining the target parking position advantageously allows for a precise evaluation by comparing the currently achieved parking position with the target parking position as a reference.Preferably, the target parking position within the detected free parking space can also be entered by the driver. For example, a suggested target parking position, positioned in the center of the free parking space, is displayed, and the driver moves and confirms it. In the simplest case, the target parking position is determined or defined as being located or positioned in the center of the free parking space, so that the desired parking position and orientation result in the car being centered and aligned with the boundaries of the free parking space. However, this is not mandatory; a more flexible target parking position, for example, positioned diagonally and off-center within the parking space, is also possible.

[0009] During the manual or steering-assisted parking maneuver, vehicle operating parameters are recorded. These parameters can include, for example, the steering angle, vehicle speed, vehicle acceleration, and / or the braking behavior of the vehicle or the driver. The recording of these parameters provides information about the driver's behavior during the parking maneuver, which is used for subsequent process steps.

[0010] Based on the recorded sensor data and operating parameters, the system determines at least two key performance indicators (KPIs) for the driver during or after the parking maneuver. These KPIs are calculated, in particular, until the vehicle has come to a complete stop within the parking space. Furthermore, the calculation of these KPIs is based on the recorded sensor data and / or operating parameters related to the parking maneuver, specifically until the vehicle has come to a complete stop within the parking space. The current traffic density and / or weather conditions in the vicinity of the vehicle can also be taken into account when calculating these KPIs.The performance indicators are advantageously determined from the moment the available parking space is detected, and / or when the vehicle's movement towards the available parking space is detected, and / or when a parking maneuver is initiated by the driver, and / or when reverse gear is engaged, and / or when the available parking space is automatically detected, particularly by a neural network. The determined performance indicators advantageously quantify the driver's parking performance based on objective criteria. The use of at least two performance indicators advantageously enables a comprehensive evaluation of parking performance.

[0011] The driver's skill score is then determined by comparing the calculated performance indicators with at least one rating set. This rating set contains performance indicators from other drivers who have parked in the same or similar available parking spaces, particularly under essentially the same traffic density and / or weather conditions in the vicinity of the vehicle. The comparison can be performed, for example, using Pareto dominance. The skill score preferably lies between 0 and 1 and represents the driver's parking abilities compared to other drivers. This comparison with other drivers allows for an objective and relative assessment of driving skills. Multiple rating sets may exist, with each rating set representing a specific driving skill class or the performance indicators of other drivers.

[0012] In a preferred embodiment, at least one end-position accuracy is determined as a performance indicator. The end-position accuracy is determined as a function of a deviation between the vehicle's stationary or parked position, in particular the vehicle's center point, and the target parking position, and / or between the current vehicle orientation or parked orientation and the target parking orientation of the target parking position. Preferably, the deviation is determined as the sum of the squared deviations of the current vehicle position from the target parking position and of the current vehicle orientation from the target parking orientation of the target parking position. In other words, the end-position accuracy is advantageously determined based on a deviation between the actual stationary vehicle position and the target parking position.Alternatively or additionally, at least one maneuvering efficiency can be determined as a performance indicator. Maneuvering efficiency is determined in particular as a function of the time required for the parking maneuver and / or the number of changes of direction during the parking maneuver, the distance traveled during the parking maneuver, the total force required for steering during the parking maneuver, and / or steering movements while stationary. In other words, maneuvering efficiency is advantageously determined as a function of the time required for the parking maneuver, the number of changes of direction during the parking maneuver, and / or the distance traveled during the parking maneuver.Alternatively or additionally, a performance indicator can be vehicle safety and / or the safety of other road users, which is determined as a function of at least one distance to a stationary or dynamic object and / or at least one determined time to collision (TTC) with the stationary or dynamic object during the parking maneuver. In other words, vehicle safety and / or the safety of other road users can be determined based on the recorded sensor data, in particular the distances to dynamic and / or static objects derived from it, and represent the risk of collisions during the parking maneuver with stationary and / or dynamic objects, such as people or pedestrians in the vicinity of the vehicle.Performance indicators can therefore include, for example, final position accuracy, maneuver efficiency and / or the safety of the vehicle and / or other road users.

[0013] Preferably, the determined performance indicators are quantified by assigning them to discrete values ​​within predefined ranges. Quantization simplifies the processing and comparison of the performance indicators and enables a more efficient assessment of driving skills.

[0014] Furthermore, the comparison is preferably carried out using Pareto dominance. Pareto dominance is a concept from game theory and optimization, named after the Italian economist Vilfredo Pareto. Thus, a driver's skill is advantageously rated higher than that of another driver skill class if the driver's performance indicators are at least as good as, and better than, the rating set of the other driver skill class in at least one criterion of the rating set.

[0015] Alternatively, the comparison can be performed as the scalar product of a vector of performance indicators and a vector of the rating scale of another driver skill class. Depending on the resulting scalar product, which represents the angle between the vectors, the driver's skill is advantageously assessed as better or worse than that of other drivers in the other skill class.

[0016] In one embodiment of the invention, the display for the driver is adapted based on a specific skill level. For example, a suggested parking trajectory to reach the target parking position can be determined and superimposed on camera images from a vehicle camera. The level of detail of the displayed parking trajectory and / or the display of learning tips to improve the skill level can also be adapted to the skill level. This personalized display assists the driver with parking and promotes learning.

[0017] Preferably, a semi-automatic or automatic parking function is activated and / or its parameters are adjusted based on a specific skill score and / or at least one determined performance indicator, particularly if the skill score is below a predefined threshold. Automatic activation of the semi-automatic or automatic parking function relieves the driver in difficult parking situations and increases safety. Adjusting the parameters results in a personalization of the semi-automatic or automatic parking function, allowing the driver, for example, to find the parking speed to be comfortable.

[0018] It may be provided that the driver is shown a recommendation to activate the semi-automatic or automatic parking function, in particular including a service fee payable for this purpose, and advantageously by means of a display device. The recommendation to activate the semi-automatic or automatic parking function is displayed, in particular, depending on the skill score and / or at least one determined performance indicator, advantageously when the skill score and / or the performance indicator is less than or equal to a predefined threshold. In particular, the driver is also shown a service fee payable for activating or performing the semi-automatic or automatic parking function. Subsequently, an input from the driver is recorded to confirm the activation of the semi-automatic or automatic parking function and preferably to pay the service fee.The activation of the semi-automatic or automatic parking function then occurs based on the driver's recorded input for confirmation. Advantageously, activation can also be contingent upon receiving a server signal confirming payment of the displayed service fee. This occurs when, prior to the confirmation, a radio signal is sent to the server, representing the payment transaction for the service fee. This design offers the advantage that the driver only pays for actually activated semi-automatic or automatic parking functions or for actually completed assisted parking maneuvers.

[0019] In a further configuration, the performance indicators are sent to a server device, where they are compared with the rating set. Alternatively, the rating set can be received from a server device. Using a server device allows access to a large and up-to-date database of parking operations and improves the accuracy of the evaluation by comparing the performance indicators with the rating set.

[0020] The invention also includes a system for evaluating driving skills, consisting of a sensor device and a computing unit. The sensor device acquires sensor data from the environment, and the computing unit is configured to execute the described method. The system can be integrated into a vehicle to continuously evaluate the driver's driving skills.

[0021] Finally, the invention also includes a computer program product that implements the described method. The computer program product enables the method to be executed on a computer and thus offers a flexible solution for evaluating driving skills.

[0022] Further advantages will become apparent from the following description of exemplary embodiments with reference to the figures. Fig. 1: Flowchart diagram of the procedure Fig. 2: Quantization Examples of implementation

[0023] In Fig. Figure 1 is a schematic block diagram illustrating the process for evaluating a driver's driving skills during a parking maneuver. The individual blocks represent the process steps described above and their logical relationships as examples. It is therefore possible that individual process steps may be carried out in parallel or in a different order, where feasible. The arrows between the blocks thus illustrate an exemplary sequence of the process, from the acquisition of sensor data to the determination of the skill score and the adjustment of the display or activation of the automatic parking function. In other words, in Fig. Figure 1 schematically illustrates the flowchart of the inventive method for evaluating a driver's driving skills during a parking maneuver. The method begins with the acquisition 10 of sensor data, in particular by means of a vehicle sensor, wherein the acquired sensor data represents the vehicle's surroundings, i.e., for example, distances to dynamic and / or static objects in the vehicle's vicinity. For example, the acquisition 10 of the sensor data is carried out using several vehicle sensors and / or different sensor types, such as at least one LiDAR, camera, radar, and / or ultrasonic sensor. The acquisition 10 of the sensor data forms the basis for precise detection of the surroundings, in particular of an available parking space or a parking gap, and enables an accurate analysis of the parking maneuver. Based on the acquired sensor data, an available parking space or a parking gap is detected in the next step 20.The detection of a parking space (20) is performed, for example, by an algorithm when, after passing the space, it is identified as free and sufficiently large. Alternatively, the detection of a free parking space (20) can be performed, for example, by a neural network based on sensor data, where the sensor data is primarily camera data, when a parking space in the vicinity of the vehicle is detected as free and sufficiently large. The detection of a free parking space (20) thus occurs through an analysis of the sensor data to identify free areas in the vehicle's vicinity. Detected parking spaces are advantageously displayed to the driver, for example, via a display overlaid with a camera image or a synthetic environment model (not shown here as a process step).The automatic detection and display of parking spaces simplifies the parking process for the driver, as they can be certain that the vehicle will fit, especially when a parking space is detected. Within the detected free parking space, a target parking position is determined in the next step. This target parking position is determined, for example, as being centered in the detected parking space and aligned with the parking space boundary. This specific target parking position represents the intended parking position the vehicle should reach and advantageously includes a target parking position and / or a target parking orientation. Determining the target parking position allows for a precise evaluation of the parking process by comparing the actual position with the target parking position. Furthermore, in step 40, at least one operating parameter of the vehicle is recorded, such as the vehicle's speed.The at least one operating parameter advantageously includes a steering angle of the vehicle's steering system, or its progression and / or changes over time; a vehicle speed, or its progression and / or changes over time; a vehicle acceleration, or its progression and / or changes over time; and / or braking behavior, brake activation, or brake pressure of the vehicle's braking system, or its progression and / or changes over time. Recording the at least one operating parameter provides important information about the driver's behavior during the parking maneuver. Subsequently, in step 50, at least two key performance indicators (KPIs) are determined based on the recorded sensor data and the at least one recorded operating parameter.These performance indicators quantify the driver's parking performance based on objective criteria. Using at least two performance indicators allows for a comprehensive and differentiated assessment of parking performance. In step 60, the driver's skill score is then determined by comparing the calculated KPIs with at least one rating set S. The rating set S comprises performance indicators from other drivers who have parked in similar situations. It is possible to compare the calculated KPIs with multiple rating sets S, where each rating set S represents a different driving skill class of other drivers.In an additional optional step 55, the performance indicators can be sent to a server device. The server device then determines the skill score by comparing it with at least one rating set S and sends it to the vehicle. Alternatively, the rating set S can be received from the server device. Using a server device provides access to a large and up-to-date database of parking maneuvers and improves the accuracy of the assessment. Comparison with other drivers allows for an objective and relative evaluation of the driver's skill. Based on the determined skill score, an optional display is customized for the driver in step 70. This can include the display of a suggested parking trajectory and / or learning tips. The personalized display assists the driver with parking and promotes learning.Furthermore, optional activation (step 80) of a semi-automatic or automatic parking function can be provided, depending on the determined skill score. For example, if the skill score falls below a predefined threshold, the semi-automatic or automatic parking function is activated and / or its parameters are adjusted. This activation of the semi-automatic or automatic parking function can initially be displayed to the driver as a recommendation in optional step 75, so that in this configuration, activation (step 80) is also based on a recorded manual input from the driver. Activation (step 80) can also be linked to costs or a service fee and a wireless, server-based payment process, with the costs preferably being displayed to the driver in step 75.In other words, in this configuration, activation 80 occurs depending on an input recorded in step 76 confirming the driver's consent to a paid parking maneuver and, advantageously, a received server signal. Prior to activation 80, a radio signal is transmitted to a server device, based on the recorded driver's confirmation of payment of the service fee (which is displayed), and the server signal is received to confirm payment of the service fee. The automatic activation of the semi-automatic or automatic parking function can occur during the parking maneuver, thus relieving the driver in difficult parking situations, for example, after the driver has begun or realized the parking maneuver is highly difficult, and increasing the safety of the vehicle and the safety of other vehicles and road users.Furthermore, in optional step 90, it can be provided that parameters of a semi-automatic and / or automatic parking function are adjusted based on a specific skill score, particularly if the skill score is less than or equal to a predefined threshold. Advantageously, adjustment 90 is performed depending on at least one determined performance indicator, so that different parameters of the semi-automatic and / or automatic parking function can be adjusted for different driving skill scores or performance indicators. This results in a highly personalized semi-automatic parking function.The semi-automatic parking function, for example, is a pure steering assistance feature that is activated depending on the driver's skill level, with the driver manually controlling the vehicle's drive, speed, and any necessary gear changes. It may also be possible for the driver to manually control the vehicle's drive, speed, and any necessary gear changes or changes of direction during the parking maneuver. With the semi-automatic parking function, a braking system can also be activated, for example, based on at least one determined performance indicator and / or a specific driver skill level, particularly to prevent collisions with static and / or dynamic objects in the vehicle's vicinity.

[0024] In Fig. Figure 2a schematically depicts a two-dimensional diagram 200, or an abstraction of two performance indicators (x1, x2), whereby in principle any number of performance indicators or dimensions can be provided. The vectors between the origin and any point in the circled evaluation set, which is formed by performance indicators of other drivers, can be used to assess a driver's driving skills. The evaluation sets or performance indicators of other drivers, represented as gray circles 210 in the diagram, therefore represent driving skill reference points, each of which can be assigned to specific driving skill classes.

[0025] In Fig. 2b is a schematic quantification of the performance indicators or evaluation sets from Fig.Figure 2a illustrates this. To reduce the required amount of data, the target space is quantized into small spaces or areas 220 or cells. For each area 220, a mean score set, for example, relative to the center of the respective area, is determined and assigned to the corresponding area 220, thus reducing the number of possible score sets S or score vectors and driving skill classes. This is useful because small changes in performance indicators are not perceived by the driver anyway, or hardly at all, and can also occur in a real-world system due to measurement inaccuracies. All points or score sets S or vectors between the origin and an endpoint within an area 220 or a cell can be interpreted as repeated occurrences of a parking process or parking maneuver with equivalent driving skill on the part of the driver.Quantization allows for the formulation of the conditions of a region 220, or space, of an ideally assessed driving skill, or a space of ideally assessed rating sets, or the estimation of the distance between the driver's performance indicators and the ideally assessed driving skill or the space of ideal rating sets. Advantageously, the driver's driving skill can be determined as the scalar product between the vector of performance indicators and the mean rating set of a quantized region 220, which is assigned to a specific predefined driving skill class. In this way, a large number of different performance indicators can be considered for assessing the driver. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2022 213 674 A1

[0002] DE 10 2015 205 156 A1

[0003]

Claims

[1] Computer-implemented method for assessing the driving skills of a driver of a vehicle in a parking maneuver performed manually or with steering assistance by a driver, the method comprising: • Acquisition (10) of sensor data during the parking process, wherein the sensor data represent the vehicle's surroundings, • Detection (20) of a free parking space depending on the sensor data acquired, • Determination (30) of a target parking position within the detected free parking space, • Recording (40) of at least one operating parameter of the vehicle during the execution of the parking maneuver by the driver of the vehicle, • Determination (50) of at least two key performance indicators (KPIs) depending on the recorded sensor data and / or the recorded operating parameter for the parking process, and • Determination (60) of a driver skill score as a function of a comparison between at least two key performance indicators (KPIs) and at least one rating set (S), wherein the rating set (S) represents key performance indicators (KPIs) of several other drivers in the same free parking space and / or similar free parking spaces. [2] Method according to claim 1, wherein the key performance indicator (KPI) is at least i. an end-position accuracy, which is determined depending on a deviation between the vehicle's stationary position and the target parking position and / or between the current vehicle orientation and the target parking orientation of the target parking position; and / or ii. a maneuvering efficiency which is determined as a function of the time required for the parking maneuver and / or the number of changes of direction, the length of the distance traveled during the parking maneuver, the total force required for steering during the parking maneuver and / or a steering movement while stationary; and / or iii. the safety of the vehicle and / or the safety of other road users, which is determined as a function of at least a distance to a stationary or dynamic object and / or as a function of at least a determined time to collision (TTC) with the stationary or dynamic object during the parking maneuver. [3] Method according to one of claims 1 or 2, wherein the following additional step is carried out: • Quantization of the at least two identified key performance indicators (KPIs) by assigning them to discrete values ​​within predefined ranges. [4] Method according to any of the preceding claims, wherein the determination (60) of the skill value is carried out by comparison using Pareto dominance. [5] Method according to one of the preceding claims, wherein the determination (60) of the skill value is carried out by comparison as a scalar product of a vector of performance indicators and a vector of the rating set. [6] Method according to any of the preceding claims, wherein the following additional step is carried out: • Adaptation (70) of a display for the driver based on the specified skill score, wherein the display includes in particular the overlay of a suggested parking trajectory into a sequence of camera images captured by a rear-facing vehicle camera and / or the level of detail of the overlaid parking trajectory and / or learning tips to improve the driver's skill score, each depending on the specified skill score, in particular when the skill score is less than or equal to a specified threshold. [7] Method according to any of the preceding claims, wherein the following additional step is carried out: • Activation (80) of a semi-automatic or automatic parking function and / or adjustment (90) of the parameters of the semi-automatic and / or automatic parking function, each based on the determined skill value and / or at least one determined performance indicator, in particular if the skill value and / or the performance indicator is less than or equal to a predetermined threshold. [8] The method of claim 7, wherein the following steps are carried out • Display (75) of a recommendation to activate the semi-automatic or automatic parking function depending on the skill score and / or at least one determined performance indicator, in particular with a service fee, especially if the skill score and / or the performance indicator is less than or equal to a specified threshold, • Recording (76) an input from the driver to confirm the activation of the semi-automatic or automatic parking function, and • Activation (80) of the semi-automatic or automatic parking function additionally based on the driver's input recorded, wherein the activation (80) is in particular additionally dependent on the receipt of a server signal to confirm that the displayed service fee has been paid. [9] Method according to one of the preceding claims, wherein the performance indicators are sent to a server device and the determination (60) of the skill value is carried out by comparison between the performance indicators (KPIs) and the at least one rating set (S) by a server device or the at least one rating set (S) is received from a server device by a radio link. [10] System for assessing a driver's driving skills when parking a vehicle, comprising at least the following components: • a sensor device designed to acquire sensor data from the vehicle's surroundings during a parking maneuver, in particular comprising a LiDAR, camera, radar and / or ultrasonic sensor, • a computing unit configured to perform a method according to any one of claims 1 to 9, wherein the sensor data acquired by the sensor device are processed to determine at least two performance indicators for the parking process and to determine a skill score of the driver as a function of a comparison of the at least two performance indicators (KPIs) with at least one rating set (S). [11] Vehicle comprising a system according to claim 10. [12] Computer program product comprising a computer program stored thereon which, when executed on a computer, causes the computer to execute the method according to any one of claims 1 to 9.

Images