Control procedure for an autonomous vehicle and computer program product

Abstract

The present invention relates to a control method for an autonomous vehicle, comprising the following steps: S5: While the autonomous vehicle is traveling straight ahead on the current road, an intersection area is detected ahead of the autonomous vehicle, formed by the current road and a cross street. A target vehicle on the cross street near the intersection area is identified, whose estimated route intersects or at least partially overlaps the navigation route of the autonomous vehicle. A priority coefficient is calculated, with a vehicle having priority if it has no priority. If the autonomous vehicle does not have priority, at least one road user is identified on the side of the autonomous vehicle facing the cross street area, whose predicted route crosses the intersection area parallel to the current road.S6: The autonomous vehicle is controlled to continue its journey along the navigation route. It also includes computer program products. The invention can optimize traffic flow, significantly reduce travel time, avoid traffic jams, and improve the driving experience.

Description

TECHNICAL AREA The present invention relates to autonomous driving technology and, in particular, a control method for autonomous vehicles. The present invention also relates to a corresponding computer program product. STATE OF THE ART In recent years, autonomous driving technology has gained increasing importance and application due to social progress and technological development. Autonomous vehicles can significantly improve the driving experience and reduce traffic accidents caused by human error or fatigue, thereby increasing road safety. For safety reasons, autonomous vehicles drive more cautiously than conventional drivers. In particular, if an autonomous vehicle is traveling straight ahead and approaches an intersection formed by its current lane and a cross street, and there is a vehicle on the cross street wishing to enter or exit the current lane, the autonomous vehicle will stop to yield the right-of-way. This can result in long waiting times for the autonomous vehicle, which can significantly increase travel time and cause unwanted traffic congestion, thus detracting from the driving experience. CONTENT OF THE INVENTION The aim of this invention is therefore to propose an improved control method for autonomous vehicles that enables autonomous vehicles to pass through intersections as quickly as possible while ensuring driving safety and compliance with laws and regulations, thereby optimizing traffic flow and significantly reducing travel time, thus avoiding traffic jams and improving the driving experience. According to a first aspect of the present invention, a control method for an autonomously driving vehicle is provided, the control method comprising the following steps: S1: When the autonomous vehicle is traveling straight ahead on the current road, it detects the intersection area ahead of the autonomous vehicle, which is formed by the current road and at least one intersecting road; S2: Identifies a target vehicle on the intersecting road near the intersection area, whose estimated travel path intersects or at least partially overlaps the navigation route of the autonomous vehicle; S3: The priority passage coefficients of the autonomous vehicle and the target vehicle are determined. The vehicle with the higher priority passage coefficient has the right of way.S4: If the autonomous vehicle does not have the right of way, at least one road user is identified on the side of the autonomous vehicle facing the intersecting road, with the predicted direction of movement of this road user across the intersection area in a direction parallel to the current road. S5: The autonomous vehicle is controlled to continue following the navigation route. In contrast to existing technologies, the control method for autonomous vehicles according to the present invention identifies a target vehicle on the cross street that could obstruct the autonomous vehicle as soon as an intersection between the current lane and a cross street is detected while the autonomous vehicle is traveling straight ahead. The priority clearance coefficients of the autonomous vehicle and the target vehicle are then evaluated to determine right-of-way. If the autonomous vehicle does not have right-of-way, a road user intending to cross the intersection is identified on the side of the autonomous vehicle facing the cross street. In this case, the target vehicle on the cross street must stop to avoid a collision with the road user. The autonomous vehicle is then controlled to pass through the intersection while the road user is crossing.This allows the autonomous vehicle to cross the intersection quickly and safely, even if it doesn't have the right of way. This can significantly reduce travel time, avoid traffic jams, and improve driving comfort. The priority passage coefficient is calculated based on factors such as: traffic signs, traffic lights, driving speed, distance to the intersection area and vehicle type. For example, if a traffic sign indicates that one direction of travel has priority and there is no traffic light, vehicles traveling in that direction have a higher priority coefficient. For example, the priority passage coefficient is higher the faster a vehicle is traveling and / or the shorter its distance from the intersection. For example, if it is a school bus or an emergency vehicle with warning lights, the vehicle has a higher priority factor. For example, each factor used to calculate the priority passage coefficient is assigned a weight value derived from empirical and / or experimental data. For example, in step S4 the position of the road user relative to the intersection area is additionally determined, and step S5 is only executed if the road user has entered a predetermined area of ​​the intersection; and / or in step S4 the speed of movement of the road user is additionally determined, and if the speed of movement drops to zero, the autonomous vehicle is controlled to brake and yield the right of way. The control procedure additionally includes, for example, the following steps: S6: If the autonomous vehicle does not have the right of way and has not identified the road user, it controls the autonomous vehicle to slow down and yield the right of way; S7: If the autonomous vehicle has the right of way, it controls the autonomous vehicle to continue its journey along the navigation route. For example, if the road has multiple lanes, the control method will only be applied if the lane in which the autonomous vehicle is located is on an edge lane adjacent to the transverse road. The estimated route of the target vehicle is determined based on factors such as the type of intersection, the direction of travel, the lane type, and the target vehicle's turn signal information. Intersections include junctions, T-junctions, and multi-lane intersections. The lane type can be a straight lane, a left-turn lane, or a right-turn lane. The estimated route of the target vehicle is selected from the following options: turning to merge into the current lane, turning to merge into oncoming traffic, or continuing straight ahead. Alternatively, the estimated trajectory of the road user can be calculated using an object tracking algorithm. For example, the intersection area is detected by the autonomous vehicle's environmental sensors and / or by the digital map stored on the autonomous vehicle; and / or the target vehicle and the road user are identified by the autonomous vehicle's environmental sensors in combination with image processing algorithms. For example, the environmental sensors include millimeter-wave radar and / or cameras and / or lidar, which are located on the front and sides of the autonomous vehicle; and / or the image processing algorithm includes an object classifier and / or a convolutional neural network algorithm; and / or the road user is a pedestrian or a driver of a two-wheeled vehicle. According to a second aspect of the present invention, a computer program product is provided which comprises a computer program which, when executed by one or more processors, enables the processors to perform the control method according to the present invention. DESCRIPTION OF THE FIGURES The present invention is described in more detail below with reference to the figures, so that the principles, features, and advantages of the present invention become even clearer. The figures include the following: Fig. 1 shows a flowchart of a control method for an autonomously driving vehicle according to an exemplary embodiment of the present invention. Figs. 2a and 2b illustrate driving scenarios for control methods for autonomous vehicles according to different embodiments of the present invention. DETAILED DESCRIPTION OF THE EXECUTION FORMS To better illustrate the technical problems to be solved, the technical solutions, and the advantageous technical effects of the present invention, the present invention is described in more detail with reference to the figures and exemplary embodiments. The specific embodiments described here serve only to explain the present invention and are not to be understood as limiting the scope of protection of the present invention. It should be clarified that in this document the terms "firstly", "secondly", etc. are used for descriptive purposes only and should not be understood as indicating relative importance or as implicitly specifying the number of technical features mentioned. Features designated as "firstly" or "secondly" may explicitly or implicitly indicate that at least one of these features is included. Fig. 1 shows a flowchart of a control method for an autonomously driving vehicle according to an exemplary embodiment of the present invention. Figs. 2a and 2b illustrate driving scenarios for control methods for an autonomous vehicle 1 according to various embodiments of the present invention. Within the scope of this invention, "autonomous vehicle" is understood to mean a vehicle that is conditionally automated (Level 3), highly automated (Level 4), or fully automated (Level 5). The control method according to the present invention is first explained with reference to the driving scenario shown in Fig. 2a. As shown in Fig. 1, the control method for an autonomous vehicle 1 according to the present invention comprises at least the following steps: S1: If the autonomous vehicle 1 travels in a straight line on road 10 according to the navigation route L1, an intersection area 30 is detected in front of it. This intersection area 30 is formed by road 10 and at least one cross street 20. For the purposes of the present invention, a "cross street" is understood to be a street that does not run parallel to road 10 and intersects the intersection area 30. The cross street 20 can run perpendicular to road 10 and thus form a T-shaped intersection, but it can also have another angle to road 20 that is considered reasonable by those skilled in the art, for example 45° or 60°, and thus form a Y-shaped intersection.Here, the navigation route L1 can be predetermined by the navigation unit of the autonomous vehicle 1, while the intersection area 30 can be detected by the environmental sensors of the autonomous vehicle 1, such as millimeter-wave radar and / or cameras and / or lidar, which are arranged on the front and sides of the autonomous vehicle 1. Alternatively, it is also possible for the intersection area 30 to be determined directly from the digital map stored in the navigation unit of the autonomous vehicle 1, in particular the high-precision digital map; S2: The environmental sensors of the autonomous vehicle 1 detect a target vehicle 2 on the cross street 20 near the intersection area 30. The estimated route L2 of the target vehicle 2 intersects the navigation route L1 of the autonomous vehicle 1 or overlaps with it at least partially.If neither the autonomous vehicle 1 nor the target vehicle 2 reacts by braking to yield the right-of-way, a collision is very likely. S3: The priority coefficients of autonomous vehicle 1 and target vehicle 2 are determined. The priority coefficient is a quantitative indicator for assessing the priority level of vehicles under certain conditions. It is calculated by combining several factors for dynamic priority assignment in traffic planning and autonomous driving decisions. Vehicles with higher priority coefficients have right-of-way. S4: If the autonomous vehicle 1 does not have right-of-way, i.e., if the priority coefficient of autonomous vehicle 1 in step S3 is lower than that of target vehicle 2, the environmental sensors of autonomous vehicle 1 identify at least one road user 3 on the side of autonomous vehicle 1 facing the intersecting road 20.The road user can be a pedestrian or a two-wheeler, e.g., a cyclist or motorcyclist. The predicted path of movement L3 of road user 3 runs parallel to the current road 10 through the intersection area 30. The predicted path of movement L3 crosses directly in front of the target vehicle 2, so the target vehicle 2 must stop and wait until road user 3 has passed. S5: Use the time that road user 3 needs to cross the intersection area 30 to control the autonomous vehicle 1 so that it continues its journey on the navigation route L1 through the intersection area 30 parallel to road user 3. In this situation, even if the target vehicle 2 has the right of way on cross street 20, the autonomous vehicle 1 can quickly pass through intersection 30 under the protection of road user 3. This significantly reduces unnecessary waiting times, shortens travel time, and largely avoids traffic jams, thereby improving the passenger experience of the autonomous vehicle 1. The method according to the present invention can, for example, further comprise the following steps: S6: If the autonomous vehicle 1 does not have the right of way and does not detect the road user 3 on the side of the autonomous vehicle 1 facing the intersecting road 20, the autonomous vehicle 1 steers itself so that it slows down and yields the right of way, so that the target vehicle 2, which has the right of way, can pass through the intersection 30 first, thereby avoiding a traffic accident and assuming responsibility for the accident. S7: If the autonomous vehicle 1 has the right of way, i.e., the priority coefficient of the autonomous vehicle 1 in step S3 is higher than the priority coefficient of the target vehicle 2, the target vehicle 2 must slow down on the intersecting road 20 and yield the right of way.Therefore, the autonomous vehicle 1 is controlled in such a way that it continues its journey on navigation route L1 directly through intersection area 30 without having to identify road user 3 on the side of the autonomous vehicle 1 facing cross street 20. For example, in step S3, the priority clearance coefficient of each vehicle can be calculated based on the following factors: traffic sign 4, traffic light, vehicle speed, distance relative to the intersection area 30, and vehicle type. Here, traffic sign 4 can be a stop sign or a major road sign. These signs clarify traffic flow in different ways to improve safety and traffic efficiency. A stop sign, for example, is red and octagonal, and is marked with the word "STOP". Vehicles must come to a complete stop before the stop sign and ensure that no other vehicles are approaching before proceeding. Furthermore, vehicles have priority when the traffic light is green, while vehicles must stop and yield when the traffic light is red.In particular, when traffic sign 4 designates a direction of travel as a priority direction and there is no traffic light, vehicles traveling in that direction have a higher priority clearance coefficient. Furthermore, the faster a vehicle is traveling and / or the shorter its distance from the intersection (zone 30), the faster it reaches and passes through the intersection (zone 30), and the higher its priority coefficient. Specifically, if the vehicle type is a school bus or an emergency vehicle with warning lights, such as a police car, fire engine, or ambulance, it has a higher priority factor. Each factor used to calculate the priority clearance coefficient is assigned a weighting value, which can be derived from empirical and / or experimental data.For example, the weighting value for traffic sign 4 and traffic lights is significantly higher than the weighting value for vehicle speed. For example, the target vehicle 2 and the road user 3 are identified by the environmental sensors of the autonomous vehicle 1 in combination with the image processing algorithm. The environmental sensors include millimeter-wave radar and / or cameras and / or lidar, which are located on the front and sides of the autonomous vehicle 1. The image processing algorithm comprises a pre-trained object classifier and / or a convolutional neural network (CNN) algorithm. In particular, the predicted trajectory L3 of road user 3 can be computed using a trained object tracking algorithm to predict the trajectory and future behavior of road user 3 in real time. For example, in step S4, the position of road user 3 relative to intersection area 30, specifically the pedestrian crossing, is additionally recorded. Step S5 is only executed if road user 3 is located within a predefined area of ​​intersection area 30, for example, in the middle section between 20% and 80% of the pedestrian crossing, so that the autonomous vehicle 1 continues its journey along navigation route L1. This ensures that the target vehicle 2 stops to allow road user 3 to cross the pedestrian crossing. For example, in step S4, the speed of road user 3 is additionally recorded. If the speed drops to zero, i.e., if road user 3 comes to a standstill in intersection area 30, this indicates that they may be yielding the right-of-way to the target vehicle 2. The autonomous vehicle 1 is then controlled accordingly to brake and yield, thereby reliably increasing driving safety. For example, as shown in Figures 2a and 2b, the estimated route L2 of the target vehicle 2 is determined based on the type of intersection 30, the direction of travel of the target vehicle 2, the lane type of the target vehicle 2, and the turn signal information of the target vehicle 2. The intersection type 30 can be a junction, a T-junction, or a multi-lane intersection; the direction of travel of the target vehicle 2 can be determined from the orientation of the vehicle's front; the lane type in which the target vehicle 2 is located can be a straight-ahead lane, a left-turn lane, or a right-turn lane, and this lane type can be determined from the lane markings or the digital map stored in the navigation system of the autonomous vehicle 1; and the turn signal information of the target vehicle 2 can be detected by the environmental sensors of the autonomous vehicle 1.In particular, the estimated route L2 of the target vehicle 2 can be selected from the following groups: turning to merge into the existing road 10, turning to merge into the opposite road from the existing road 10, or continuing straight ahead on the existing road 10. In Fig. 2a, the estimated route L2 of the target vehicle 2 merges into the existing road 10 by turning left. This estimated route L2 partially overlaps with the navigation route L1 of the autonomous vehicle 1. In Fig. 2b, the estimated route L2 of one target vehicle 2 is shown, which merges into the existing road 10 by turning right, while the estimated route L2 of another target vehicle 2 continues straight ahead on the existing road 10. This estimated route intersects with the navigation route L1 of the autonomous vehicle 1.However, it is also possible that the predicted route L2 of the target vehicle 2 merges into the oncoming lane of the current road 10 by turning left, and that this predicted route also intersects with the navigation route L1 of the autonomous vehicle 1. As shown in Fig. 2b, the control method according to the invention is applied, for example, only when the current lane in which the autonomous vehicle 1 is located is the edge lane of the transverse road 20. If the autonomous vehicle 1 is located in the middle lane of road 10, it can be controlled to pass directly through the intersection area 30, without regard to the target vehicle 2 on the intersecting road 20. The present invention further relates to a computer program product comprising a computer program which, when executed by one or more processors, enables the processors to perform a control method for an autonomous vehicle 1 according to the present invention. The computer program product according to the present invention is, for example, arranged in the vehicle control unit of the autonomous vehicle 1. The above descriptions of the embodiments only illustrate the present invention within the scope of the examples. Naturally, the various features of the implementation method can be freely combined with one another, provided this is technically feasible, without departing from the scope of the present invention. For those skilled in the field, further advantages of the present invention and alternative embodiments are obvious. Therefore, the present invention, in a broader sense, is not limited to the specific details, representative structures, and exemplary embodiments shown and described. Conversely, those skilled in the art can make various modifications and substitutions without deviating from the basic concept and scope of the invention.

Claims

A control procedure for an autonomous vehicle (1), characterized in that the control procedure comprises at least the following steps: S1: In the case that the autonomous vehicle (1) is traveling straight ahead on the current road (10), an intersection area (30) located ahead of the autonomous vehicle (1) is detected, wherein the intersection area (30) is jointly formed by the current road (10) and at least one cross road (20); S2: a target vehicle (2) located on the cross road (20) and in the vicinity of the intersection area (30) is identified, wherein the predicted route (L2) of the target vehicle (2), which intersects or at least partially overlaps the navigation route (L1) of the autonomous vehicle (1), is determined;S3: The priority passage coefficients of the autonomous vehicle (1) and the target vehicle (2) are each evaluated, with the vehicle having the higher priority passage coefficient having the right of way; S4: If the autonomous vehicle (1) does not have the right of way, at least one road user (3) located on the side of the cross street (20) facing the autonomous vehicle (1) is identified, wherein the road user (3) has a predicted trajectory (L3) that crosses the intersection area (30) in a direction parallel to the current street (10); S5: The autonomous vehicle (1) is controlled to continue following the navigation route (L1). Control method according to claim 1, characterized in that the priority passage coefficient is calculated jointly on the basis of the following factors: road signs (4), traffic lights, driving speed, distance to the intersection area (30), vehicle type. The control method according to claim 2, characterized in that, if the road sign (4) indicates a direction of travel as a priority direction and no traffic lights are present, the priority passage coefficient assigned to vehicles traveling along this direction of travel is higher; and / or the higher the speed of a vehicle and / or the shorter its distance to the intersection area (30), the higher the priority passage coefficient of this vehicle; and / or if the vehicle type of a vehicle is a school bus or an emergency vehicle with warning lights, the priority passage coefficient of this vehicle is higher; and / or each of the respective factors for calculating the priority passage coefficient is assigned a weighting value, wherein the weighting value is derived from empirical data and / or experimental data. Control method according to one of the preceding claims, characterized in that in step S4 the position of the road user (3) relative to the intersection area (30) is additionally detected and step S5 is only executed if the road user (3) has already entered a predetermined area of ​​the intersection area (30); and / or in step S4 the speed of movement of the road user (3) is additionally detected and, if the speed of movement decreases to zero, the autonomously driving vehicle (1) is controlled to brake and yield the right of way. A control method according to one of the preceding claims, characterized in that the control method additionally comprises the following steps: S6: If the autonomously driving vehicle (1) does not have the right of way and the road user (3) is not identified, the autonomously driving vehicle (1) is controlled to brake and yield the right of way; S7: If the autonomously driving vehicle (1) has the right of way, the autonomously driving vehicle (1) is controlled to continue following the navigation route (L1). Control method according to one of the preceding claims, characterized in that, if the current road (10) has several lanes, the control method is only executed if the autonomously driving vehicle (1) is on the current lane which is the edge lane and directly borders the cross road (20). A control method according to one of the preceding claims, characterized in that the predicted route (L2) of the target vehicle (2) is determined based on the type of intersection (30), the direction of travel of the target vehicle (2), the lane type used by the target vehicle (2), and the turn signal information of the target vehicle (2), wherein the type of intersection (30) is a junction, a T-junction, or a multi-lane junction, and the lane type is a straight-ahead lane, a left-turn lane, or a right-turn lane, wherein the predicted route (L2) of the target vehicle (2) is selected from the following group: turning into the current road (10), turning into the road opposite the current road (10), continuing straight through the current road (10); and / or the predicted trajectory (L3) of the road user (3) is calculated by an object tracking algorithm. Control method according to one of the preceding claims, characterized in that the intersection area (30) is detected by environmental sensors of the autonomously driving vehicle (1) and / or on the basis of a digital map stored in the autonomously driving vehicle (1); and / or the target vehicle (2) and the road user (3) are identified by environmental sensors of the autonomously driving vehicle (1) in combination with an image processing algorithm. Control method according to claim 8, characterized in that the environmental sensors comprise a millimeter-wave radar and / or a camera and / or a LiDAR arranged on the front and side parts of the autonomously driving vehicle (1); and / or the image processing algorithm comprises an object classifier and / or an algorithm based on a convolutional neural network; and / or the road user (3) is a pedestrian or a two-wheeler rider. Computer program product comprising a computer program, characterized in that, when the computer program is executed by one or more than one processor, the processor is able to execute a control method according to any one of claims 1 to 9.

Images