Method and system for vehicle path planning

Abstract

A method for operating a vehicle having an ADS and a fallback stop feature is provided that determines, based on sensor data and positioning data, a plurality of candidate paths within a drivable area in a surrounding environment of the vehicle for a prediction time horizon. Each candidate path is associated with a nominal cost function value based on at least one cost parameter. Further, the method includes determining an expected trajectory of a target vehicle located in the surrounding environment of the vehicle for the prediction time horizon, and determining, for each candidate path, an overlap cost parameter regarding an overlap between the expected trajectory of the target vehicle and a set of stop locations of the vehicle based on a predicted execution of the fallback stop feature for the prediction time horizon.

Description

Technical Field

[0001] This disclosure relates to autonomous vehicles, and more specifically, to path planning for autonomous vehicles. Background Technology

[0002] The development of autonomous vehicles has exploded in recent years, with various solutions being explored. Today, within these diverse technological fields, both automated driving (AD) and advanced driver assistance systems (ADAS), or semi-autonomous driving, are constantly evolving. ADAS and AD can be collectively referred to as automated driving systems (ADS) with different levels of automation, as defined by SAE J3016 Driving Automation Levels. One such area is how to move vehicles accurately and consistently, as this is a crucial aspect of safety when vehicles are moving in traffic.

[0003] In other words, autonomous vehicles are developing rapidly, and there are often news reports and demonstrations of impressive technological advancements. However, one of the biggest challenges facing ADS is ensuring that autonomous vehicles can safely plan and execute paths and / or trajectories.

[0004] Generally, traditional path planning systems generate a target path for autonomous vehicles from a given drivable area, which is typically provided based on data from a perception system. The target path is then sent to the vehicle controller, which calculates the steering angle so that the vehicle follows that path. Summary of the Invention

[0005] Therefore, one object of this disclosure is to provide a method, a computer-readable storage medium, a control device, and a vehicle including such a control device for operating a vehicle having ADS and back-stop features, which mitigates all or at least some of the disadvantages of currently known systems.

[0006] More specifically, the purpose of this disclosure is to provide a path planning or trajectory planning solution that not only focuses on object avoidance and comfort, but also reduces the risk of collisions with other vehicles in the event that an autonomous vehicle comes to an unexpected stop.

[0007] This object is achieved by means of a method for operating a vehicle having ADS and back-stop features, a computer-readable storage medium, a control device, and a vehicle including such a control device, as defined in the appended claims. The term "exemplary" is understood in this context to mean instance, example, or illustration.

[0008] According to a first aspect of this disclosure, a method for operating a vehicle having ADS (Adaptive Dash) and back-stop features is provided. The method includes acquiring sensor data and positioning data (which includes information about the vehicle's surrounding environment), and based on the sensor data and positioning data, determining a plurality of candidate paths for a predicted time range within a drivable area in the vehicle's surrounding environment. Each candidate path is associated with a nominal cost function value based on at least one cost parameter. Furthermore, the method includes determining a predicted trajectory of a target vehicle located in the vehicle's surrounding environment for the predicted time range based on the acquired sensor data and positioning data, and, based on the back-stop feature, determining an overlap cost parameter for each candidate path regarding the overlap between the target vehicle's predicted trajectory and a set of stopping positions of the vehicle. The method also includes selecting a candidate path from the plurality of candidate paths by calculating a cost function based on the nominal cost function value and the overlap cost parameter, and generating a control signal at an output to control the vehicle to follow the selected candidate path.

[0009] Therefore, a method is provided for autonomous vehicles to execute paths where a minimum portion of the autonomous vehicle is in the lane of potentially oncoming traffic for a duration where unintended stopping could expose both the autonomous vehicle's passengers and those of oncoming vehicles to collision risk. This improves the overall safety of the autonomous vehicle's path planning and / or trajectory planning features.

[0010] According to a second aspect of this disclosure, a (non-transitory) computer-readable storage medium is provided that stores one or more programs configured to be executed by one or more processors of a vehicle control system, the one or more programs including instructions for performing methods according to any embodiment of the present disclosure. With respect to this aspect of the disclosure, there are similar advantages and preferred features as to the first aspect of the disclosure previously discussed.

[0011] The term "non-transitory" as used herein is intended to describe computer-readable storage media (or "memory") that do not propagate electromagnetic signals, but is not intended to otherwise limit the type of physical computer-readable storage device included by the terms computer-readable media or memory. For example, the terms "non-transitory computer-readable media" or "tangible memory" are intended to cover types of storage devices that do not necessarily permanently store information, such as including random access memory (RAM). Program instructions and data stored in a non-transitory form on tangible computer-accessible storage media can also be transmitted via transmission media or signals such as electrical, electromagnetic, or digital signals, which can be transmitted via communication media such as networks and / or wireless links. Therefore, the term "non-transitory" as used herein is a limitation on the medium itself (i.e., tangible, not signal-based), rather than a limitation on the persistence of data storage (e.g., RAM or ROM).

[0012] Furthermore, according to a third aspect of this disclosure, a control device is provided for operating a vehicle having ADS and backstop features. The control device includes control circuitry configured to acquire sensor data and positioning data (which includes information about the vehicle's surrounding environment), and based on the sensor data and positioning data, determine a plurality of candidate paths for a predicted time range within a drivable area in the vehicle's surrounding environment. Each candidate path is associated with a nominal cost function value based on at least one cost parameter. Furthermore, the control circuitry is configured to determine the expected trajectory of a target vehicle located in the vehicle's surrounding environment based on the acquired sensor data and positioning data, and, based on the execution of a prediction of the backstop feature within the predicted time range, determine an overlap cost parameter for each candidate path regarding the overlap between the expected trajectory of the target vehicle and a set of stopping positions of the vehicle. The control circuitry is also configured to select a candidate path from the plurality of candidate paths by calculating a cost function based on the nominal cost function value and the overlap cost parameter, and generate a control signal at an output to control the vehicle to follow the selected candidate path. Similar advantages and preferred features exist for this aspect of the disclosure as for the first aspect of the disclosure previously discussed.

[0013] According to a fourth aspect of this disclosure, a vehicle perception system is provided, comprising at least one sensor configured to monitor the vehicle's surrounding environment, a positioning system configured to monitor the vehicle's geographic map location, and a control device for operating a vehicle having ADS and back-stop features according to any of the embodiments disclosed herein. This aspect of the disclosure has similar advantages and preferred features to the first aspect of the disclosure previously discussed.

[0014] Other embodiments of this disclosure are defined in the dependent claims. It should be emphasized that the term "comprising / including" as used in this specification is used to specify the presence of a stated feature, integral, step, or component. It does not exclude the presence or addition of one or more other features, integrals, steps, components, or combinations thereof.

[0015] The above and other features and advantages of this disclosure will be further illustrated below with reference to the embodiments described herein. Attached Figure Description

[0016] Further objects, features, and advantages of embodiments of this disclosure will be described in detail below with reference to the accompanying drawings, wherein:

[0017] Figure 1 This is a schematic flowchart representation of a method for operating a vehicle having an automated driving system (ADS) and a back-stop feature according to embodiments of the present disclosure.

[0018] Figure 2 This is a schematic block diagram representation of a system for operating a vehicle having an automated driving system (ADS) and a back-stop feature, according to embodiments of the present disclosure.

[0019] Figure 3a This is a schematic top view of a vehicle having two candidate paths according to an embodiment of the present disclosure.

[0020] Figure 3b This is a schematic diagram illustrating a set of stopping positions for a vehicle within a predicted time range according to an embodiment of the present disclosure.

[0021] Figure 4a This is a schematic top view of the accumulation of the set of stopping positions of the vehicle and the first candidate path according to an embodiment of the present disclosure.

[0022] Figure 4b This is a schematic top view of the accumulation of the set of stopping positions of vehicles and second candidate paths according to embodiments of the present disclosure.

[0023] Figure 5a This is a schematic top view of a vehicle having two candidate paths according to an embodiment of the present disclosure.

[0024] Figure 5b This is a schematic diagram illustrating a set of stopping positions within a predicted time range for determining a vehicle according to an embodiment of the present disclosure.

[0025] Figure 6 This is a schematic side view of a vehicle according to an embodiment of the present disclosure. Detailed Implementation

[0026] Those skilled in the art will understand that the steps, services, and features described herein can be implemented using separate hardware circuitry, a microprocessor or general-purpose computer programmed with software features, one or more application-specific integrated circuits (ASICs), and / or one or more digital signal processors (DSPs). It will also be appreciated that, when this disclosure is described according to a method, it can also be implemented in one or more processors and one or more memories coupled to said one or more processors, wherein said one or more memories store one or more programs that, when executed by said one or more processors, perform the steps, services, and features disclosed herein.

[0027] In the following description of exemplary embodiments, the same reference numerals denote the same or similar parts.

[0028] In this disclosure, it is assumed that the Automated Driving System (ADS) requires user / driver supervision. Because it is assumed that automated driving capabilities are not available in every scenario, it is also assumed that the ADS is only available within a limited Operational Design Domain (ODD), therefore the user / driver's ability to take over driving tasks is always evaluated. When it is determined that the user is not ready to take over driving within a certain period after a handover request is issued (e.g., due to persistent negligence), the ADS (or any relevant safety system) performs a rollback measure by slowing the vehicle and eventually bringing it to a complete stop.

[0029] A rollback can also be performed when the autonomous driving system becomes inoperable due to a malfunction and the driver does not take over vehicle control (upon request). Furthermore, it is assumed that the rollback system and vehicle control remain available even if the remaining ADSs become inoperable due to safety design, which can be achieved through hardware redundancy and an "ASIL-based" design. Additionally, for the sake of simplicity and brevity in this application, the deceleration approximation obtained by the rollback stop execution described herein has a predefined curve, such as constant deceleration.

[0030] Figure 1 A schematic flowchart illustrates a method 100 for operating a vehicle with an Automated Driving System (ADS) and a back-stop feature. The back-stop feature may also be referred to as a "safe stop" feature. The ADS may, for example, be an automation level of Level 3 or higher as defined by SAE J3016 Automation Levels.

[0031] Method 100 includes acquiring 101 sensor data and positioning data, which includes information about the vehicle's surrounding environment. Sensor data can be acquired, for example, from the vehicle's perception system. The perception system is understood herein as a system responsible for acquiring raw sensor data from onboard sensors such as cameras, LiDAR and RADAR, ultrasonic sensors, and converting this raw data into scene understanding. Naturally, sensor data can be received directly from one or more suitable sensors (such as cameras, LiDAR sensors, radar, ultrasonic sensors, etc.). Positioning data preferably includes the vehicle's geographic location and heading, as well as map data (e.g., HD map data). In other words, positioning data indicates the vehicle's position and orientation on a map. For example, positioning data can be acquired from the vehicle's positioning system. The positioning system is a system configured to monitor the vehicle's geographic location and heading, and can be in the form of a Global Navigation Satellite System (GNSS), such as GPS. However, the positioning system can also be implemented as Real-Time Kinematic (RTK) GPS to improve accuracy. The term "acquisition" is interpreted broadly herein, including receiving, retrieving, collecting, obtaining, etc.

[0032] Furthermore, method 100 includes determining multiple candidate paths for a predicted time range (also referred to as a planning range) within a drivable area in the surrounding environment of vehicle 102 based on sensor data and positioning data. Additionally, each candidate path is associated with a nominal cost function value based on at least one cost parameter.

[0033] More specifically, in some embodiments, a series or set of waypoints are generated based on HD map data and vehicle locations, where the purpose of waypoints is to define the expected points the vehicle will pass through within a predicted / planned timeframe. Driving areas are then generated based on sensor data surrounding these generated waypoints. Given waypoints, candidate paths are planned by evaluating a cost function that quantifies a set of criteria (e.g., path proximity to waypoints, path smoothness, total path distance, etc.). Furthermore, each path is preferably planned such that it is constrained within the generated driving area. Additionally, in some embodiments, trajectories are planned, thus longitudinal planning such as speed curves or longitudinal future target locations may be considered. Then, under the cost function and constraints, conventional optimization methods, such as sample-based methods (e.g., fast exploratory random trees, RRT, RRT*) or optimal control methods (e.g., model predictive control, MPC), can be utilized. Therefore, the final product is a set of candidate trajectories with nominal cost function values.

[0034] Furthermore, method 100 includes determining, 103, the expected trajectory of a target vehicle (external vehicle) located in the vehicle's surrounding environment for a predicted time range based on acquired sensor data and positioning data. In other words, prediction 103 of the trajectory of the external vehicle relative to the autonomous vehicle is made based on sensor data and map data. More specifically, the step of determining the expected trajectory 103 may include determining the current state of the target vehicle based on sensor data and positioning data, wherein the current state of the target vehicle includes the geographic / map location of the target vehicle, the heading of the target vehicle, the speed of the target vehicle, the size of the target vehicle (e.g., the width of the target vehicle), and based on the current state and positioning data (e.g., lane geometry of the target vehicle's lane, obstacles, traffic rules, etc.). Then, based on the current state and the predicted future state, the expected trajectory can be determined.

[0035] For example, if a target vehicle is traveling in the opposite direction to the autonomous vehicle in an adjacent lane, the trajectory of the target vehicle in that lane is predicted given the initial state of the target vehicle (e.g., position, heading, speed) and environmental parameters such as traffic rules, lane geometry (of the target vehicle's lane), obstacles, etc. 103.

[0036] Given the expected trajectory determined 103, determine 104 the overlap cost parameter for each candidate path determined 102. The overlap cost parameter indicates the overlap between the expected trajectory of the target vehicle and a set of stopping positions of the autonomous vehicle, where the set of stopping positions is based on the execution of the prediction of the backstop feature within the prediction time range. Assuming multiple executions at multiple time points within the prediction time range, this set of stopping positions can be understood as the "cumulative stopping positions" of the autonomous vehicle. In this paper, it is assumed that the backstop feature executes the stop within the path generated by the vehicle; that is, we assume that there is no "safe zone" where the autonomous vehicle will move to the safe zone when the backstop function is executed. For example, if the prediction time range is 10 seconds, then the "set of stopping positions" could be the prediction of the stopping positions of the autonomous vehicle when the backstop feature is executed at 0 seconds, 2 seconds, 4 seconds, 6 seconds, 8 seconds, and 10 seconds. Therefore, within each candidate path, we have six stopping positions of the autonomous vehicle, and these stopping positions are compared with the expected trajectory of the target vehicle determined 103 to calculate the overlap cost parameter.

[0037] Furthermore, method 100 includes selecting a candidate path from multiple candidate paths by calculating a cost function based on a nominal cost function value and overlapping cost parameters, and generating a control signal at the output to control the vehicle to follow the selected candidate path. This can also be referred to as a constraint-controlled technique employing a cost-minimizing control strategy, i.e., given a set of constraints, generating multiple candidate paths or trajectories, and then selecting the candidate path with the lowest associated cost.

[0038] The effect is that the autonomous vehicle is configured to select a path or trajectory that results in a portion of the autonomous vehicle being in a lane of potentially oncoming traffic for the shortest duration (unexpected stopping could put passengers in the autonomous vehicle and passengers in oncoming vehicles at risk of collision). Furthermore, the inventors recognized that previous path planning solutions focused on avoiding objects, possible speed curves, and / or the efficiency of trajectory selection, and therefore provided a new solution that addresses these issues using additional parameters, namely, the risks associated with traversing unsafe parking areas or unexpected stopping from an observer's perspective.

[0039] Therefore, the solution proposed in this paper asserts that the risk of a malfunction occurring within an autonomous driving system at any given time is non-zero. In some cases, the ideal response to such a malfunction is to bring the vehicle to an immediate stop in a safe manner. Given these conditions, it is desirable to minimize the amount of time the autonomous system spends in a state where immediate stopping would be unsafe. Therefore, a means is proposed to influence the weighting factor (or cost function) of the candidate path (or trajectory) such that the penalty (or reward) corresponds to the duration for which the controlled vehicle is expected to be in this state (i.e., a state where immediate stopping would be unsafe).

[0040] Executable instructions for performing these features may optionally be included in a non-transitory computer-readable storage medium or in other computer program products configured to be executed by one or more processors.

[0041] Figure 2 This is a schematic block diagram describing the system architecture of a control device 10 for operating a vehicle with an automated driving system (ADS) and back-stop features. The ADS is shown as a path / trajectory planner 26 (hereinafter referred to as the path planner), an additional module having a path / trajectory selector 27 (hereinafter referred to as the path selector), and a motion predictor 28. In other words, the method disclosed herein can be implemented as a software update of conventional path / trajectory planning features. Therefore, the control device 10 includes control circuitry (implemented by means of multiple software / hardware modules 26, 27, 28).

[0042] Continuing, the system architecture also includes a waypoint generator 24, which can be understood as being configured to generate a sequence of waypoints expected to traverse within a predicted / planned timeframe. Waypoint generator 24 receives input in the form of HD map data 22 and autonomous vehicle location 21 to generate waypoints. Furthermore, based on HD map data 22 and sensor data 23, a "drivable area limiter" generates drivable areas around the desired waypoints. In other words, it generates "constraints" for path planner 26.

[0043] The system also includes a path planner 26, configured to determine multiple candidate paths for a predicted time range within a drivable area of ​​the vehicle's surrounding environment based on sensor data 23 and positioning data 21, 22. More specifically, given waypoints (generated by waypoint generator 24), path planner 26 determines multiple candidate paths by evaluating a cost function that quantifies a set of criteria, namely cost parameters, such as the path's proximity to waypoints, path smoothness, total path distance, etc. Furthermore, multiple candidate paths are determined such that they are constrained within the drivable area. In the case of trajectory planning, further criteria could be proximity to a set speed, speed profile, and longitudinal future target position. Numerical optimization methods are used under the cost function, cost parameters, and constraints. Therefore, the output from path planner 26 is a set of candidate paths, each associated with a nominal cost function value.

[0044] Furthermore, the motion predictor module 28 is configured to determine or predict the trajectory of a target vehicle within a prediction time range in the vehicle's surrounding environment based on sensor data 23 and positioning data 21, 22. More specifically, the prediction can be performed under conservative assumptions about the target vehicle's motion, such as by extending the target vehicle's velocity vector. While the target vehicle may perform slow deceleration or gradual evasive maneuvers, this assumption does not include sudden collision avoidance maneuvers with large accelerations. Additionally, the path selector 27 is configured to select a candidate path from multiple candidate paths by calculating a cost function based on a nominal cost function value and an overlap cost parameter. More specifically, the path selector 27 is configured to select a candidate path such that it minimizes the overlap between the target vehicle's expected trajectory and the cumulative stopping position of the autonomous vehicle caused by the back-stop execution. The path selector 27 is then configured to generate a control signal at the output to control the vehicle to follow the selected candidate path.

[0045] The generated control signals are received by the vehicle control system 30, which is configured to control the steering angle, acceleration, and braking, causing the autonomous vehicle to follow the selected candidate path. Furthermore, the vehicle control system 30 can be further configured to retain the most recently selected path / trajectory in memory, enabling it to track the lateral target position during backstop execution, even if the path planner cannot provide a target path in the event of a system failure. The system also includes a backstop actuator 29, which generates a command signal to the vehicle control system to perform a backstop, i.e., a predefined deceleration request. The backstop actuator 29 can be configured, for example, to generate the command signal if the driver / passenger determines that they cannot take over driving duties for a certain period of time, or in the event of a hardware / software failure in the ADS and the driver / passenger is unable to respond to a handover request. In some embodiments, the backstop actuator is part of the ADS software.

[0046] It should be noted that Figure 2 The system architecture modules shown are merely one example of several possibilities covered by the scope of this disclosure. For example, the path / trajectory planner 26 and the path / trajectory selector could be integrated into a single module. Thus, the combined module could obtain the nominal cost function value and overlapping cost parameters, and be configured to directly select the optimal candidate path (i.e., the candidate path associated with the lowest total cost function value).

[0047] Furthermore, the motion predictor module 28 may be a separate module external to the control device 10, such that the control device 10 is configured to "determine" the trajectory of a target vehicle located in the surrounding environment of the vehicle by acquiring the trajectory of the target vehicle from an "external" motion predictor module (not shown). The above and other obvious modifications, depending on the overall system architecture, specifications, and intended application, are considered readily understandable to those skilled in the art and are included within the scope of this disclosure.

[0048] The various details and features of the methods and systems disclosed above will be referenced. Figure 3a , Figure 3b , Figure 4a , Figure 4b , Figure 5a and Figure 5b The example scenario described further illustrates this, namely, when an autonomous vehicle is about to make a left turn at an intersection, thus crossing the lane of potential oncoming traffic.

[0049] Figure 3a This is a schematic top view of an autonomous vehicle 1 approaching intersection 32, with the intention of turning left to cross lane 31 with oncoming traffic. The autonomous vehicle 1 has an Adaptive Dashboard (ADS), a reverse stop feature, and a control device (not shown) including control circuitry configured to execute one or more programs comprising instructions for performing methods according to any of the embodiments disclosed herein. More specifically, the control circuitry is configured to acquire sensor data and positioning data (including information about the vehicle's surrounding environment). As previously described, the positioning data may be in the form of HD map data acquired, for example, by the vehicle 1's Global Navigation Satellite System (GNSS) and the vehicle's geographic location.

[0050] Furthermore, the control circuitry is configured to determine multiple candidate paths 33, 34 for a predicted time range based on sensor data, map data, and location data. A nominal cost function value calculated based on one or more cost parameters is provided for each candidate path 33, 34. In the example shown ( Figure 3aThe first candidate path 33 yields a smaller nominal cost function value than the second candidate path 34. Furthermore, the control circuitry is configured to determine the expected trajectory 39 of one or more target vehicles 31 located in the surrounding environment of vehicle 1 for a predicted time range, based on sensor data and positioning data. More specifically, in some embodiments, the prediction of the expected trajectory 39 of the target vehicles 31 is made under certain assumptions about the motion of the target vehicles, for example, by extending the velocity vector of the target vehicles 31.

[0051] Furthermore, the control circuit is configured to perform predictions within the prediction time range based on the back-off stop characteristics, and to determine an overlap cost parameter for each candidate path 33, 34 regarding the overlap between the expected trajectory 39 of the target vehicle 31 and a set of stopping positions of the autonomous vehicle 1.

[0052] Figure 3b This is a schematic diagram illustrating the determination of a set of stopping positions of vehicle 1 in the form of "cumulative stopping positions" 41 within a prediction time range 42 according to an embodiment of the present disclosure. The prediction time range 42 is denoted herein as extending from time t0 to time tN. More specifically, given prior knowledge of the rollback deceleration curve, the current vehicle speed, and the future target speed curve, the control circuit can be configured to predict the stopping position of the autonomous vehicle after performing a rollback at any point within the prediction time range [t0, tN]. Let t0 represent the current time; the stopping position resulting from the rollback starting at t0 is predicted as stopping position SP1. Similarly, for tN, the end of the prediction time range 42, a second stopping position SP2 resulting from the rollback starting at tN is predicted. As stated above, the prediction range 42 is a predefined constant. Therefore, the cumulative stopping position of the autonomous vehicle is defined as the range of stopping positions expected from the current time t0 to the time range t1 resulting from the rollback execution.

[0053] For each candidate path 33, 34, the expected trajectory of target vehicle 31 overlaps with a set of stopping positions (more specifically, cumulative stopping positions) in Figures 4a to 4b As shown in the image. More specifically, Figures 4a-4b It is the first candidate path ( Figure 4a The cumulative set of stopping positions 41a and the second candidate path ( Figure 4b A schematic top view of the cumulative set of stop positions 41b.

[0054] In addition, Figure 4a The diagram shows the first overlap 42a between the expected trajectory 39 of the target vehicle 31 and the cumulative stopping position 41a of the first candidate path, while... Figure 4bThe figure shows the second overlap 42b between the expected trajectory 39 of the target vehicle 31 and the cumulative stopping position 41b of the second candidate path. As shown, the overlap of the first candidate path is greater than that of the second candidate path, that is, the first overlap 42a is greater than the second overlap 42b (as shown by the bidirectional arrows 43a and 43b). Therefore, the overlap cost parameter of the first candidate path is correspondingly greater than that of the second candidate path.

[0055] Furthermore, the control circuit is configured to consider the dimensions of the autonomous vehicle 1 to determine multiple candidate paths 33, 34, and particularly when determining the overlap parameters for each candidate path. Similarly, the control circuit is configured to consider the dimensions / size of the target vehicle 31 when determining the expected trajectory 39 of the target vehicle. In other words, the dimensions of the autonomous vehicle 1 (or at least the length of the autonomous vehicle 1) and the dimensions of the target vehicle (or at least the width of the target vehicle) are used to determine the overlap cost parameters. By considering at least the width of the target vehicle 39 and the length of the autonomous vehicle 1, a more accurate estimate of the overlap 42a, 42b between the expected trajectory 39 of the target vehicle 31 and the candidate paths of the autonomous vehicle 1 can be achieved.

[0056] Alternatively, the expected trajectory 39 of the target vehicle 31 can also be considered to include a width defined by the width of the target vehicle 31. Similarly, each candidate path can include a width defined by the width of the autonomous vehicle 1.

[0057] Therefore, the control circuit of the control device is configured to select a candidate path from multiple candidate paths by calculating a cost function based on the nominal cost function value and the overlapping cost parameter associated with each candidate path (see reference). Figure 3a (33, 34). In other words, when selecting a candidate path, the control circuit is configured to select from multiple candidate paths the candidate path with the minimum overlap region 42a, 42b between the expected trajectory 39 of the target vehicle 31 and a set of stopping positions 41a, 41b of the vehicle 1, as a result of the execution of the prediction of the back-stop feature within the prediction time range. Therefore, in Figures 3a-3b and Figures 4a-4b In the example shown, the nominal cost of the first candidate path 33 is less than the nominal cost of the second candidate path 34. However, the total cost function value of the second candidate path 34 (i.e., nominal cost + overlap cost parameter) is less than the total cost function value of the first candidate path 33. Therefore, the control circuit is configured to select the second candidate path 33 and generate a control signal at the output to control the vehicle to follow the selected candidate path (i.e., the second candidate path 34).

[0058] Figure 5a This is a schematic top view of an autonomous vehicle 1 having two candidate paths 33, 34 according to another embodiment of this disclosure. More specifically, Figure 5aAnother example illustrating how to determine the overlap cost parameter between the expected trajectory 39 of target vehicle 31 and a set of stopping positions of autonomous vehicle 1 based on the execution of predictions within the prediction time range using back-off stopping characteristics. See reference... Figures 3a-3b and Figures 4a-4b As discussed in the previous examples, the same two example candidate paths 33 and 34 are used.

[0059] Figure 5b This is a schematic diagram illustrating the determination of a set of stopping positions 44 for an autonomous vehicle 1 within a predicted time range 42 [t0, tN] according to an embodiment of the present disclosure. More specifically, given prior knowledge of the rollback deceleration curve, the current vehicle speed, and the future target speed curve, the control circuit can be configured to predict the stopping position of the autonomous vehicle after performing rollbacks at multiple time points within the predicted time range 42; seven equidistant time points are selected herein. Let t0 represent the current time, and the stopping position resulting from a rollback initiated at t0 is predicted as the first stopping position SP1. Similarly, for a second time t1, a second stopping position SP2 resulting from a rollback initiated at t2 is predicted, and so on, until a seventh stopping position SP7 is determined as a result of a rollback initiated at tN.

[0060] exist Figure 5a The diagram shows a graphical representation of a set of stop positions 44 for each candidate path 33, 34. Therefore, to determine the overlap cost parameter, the number of stop positions 44 overlapping the expected trajectory 44 of the target vehicle 31 can be calculated for each candidate path 33, 34. For the first candidate path 33, there are four completely overlapping stop positions 44 and one partially overlapping stop position 44, while for the second candidate path 34, there are three completely overlapping stop positions 44. Therefore, the overlap cost parameter associated with the first candidate path 33 is greater than that of the second candidate path 34. As mentioned above, when determining the expected trajectory 39 of the target vehicle 31, the width of the target vehicle 31 can be considered. In other words, the expected trajectory 39 of the target vehicle 31 can have a width defined by the width of the target vehicle 31. Furthermore, the length of the autonomous vehicle can be used to calculate whether the stop positions 44 overlap with the expected trajectory 39 of the target vehicle 31. For example, if the reference point of autonomous vehicle 1 is the center point of the rear axle, the stopping position near the first boundary of the expected trajectory 39 (i.e., the stopping position near the center lane mark) may still result in overlap, while the stopping position near the opposite boundary of the expected trajectory 39 may be "clear" (i.e., non-overlapping).

[0061] Furthermore, while the above example embodiments have referenced scenarios involving intersections, more specifically left turns at intersections, other scenarios exist where the autonomous vehicle's planned candidate path intersects with the target vehicle's intended trajectory. For example, this could occur when the autonomous vehicle merges from an alley into the second lane of a main road while the target vehicle is approaching in the first lane of the main road. Another example scenario is when the target vehicle is following the autonomous vehicle on a road with multiple lanes for traffic traveling in the same direction.

[0062] Referring to the latter example scenario, based on the same principle of candidate path selection as described above, since the target vehicle is approaching from behind, the autonomous vehicle can select a candidate path that leads to a lane change. If the autonomous vehicle unexpectedly stops (i.e., if a reverse stop is performed), the lane change will reduce the risk of collision.

[0063] However, in some embodiments, the control circuitry is also configured to determine a scenario from multiple predefined scenarios (e.g., approaching an intersection, driving in a multi-lane road, lane merging, etc.) based on sensor and positioning data. In some embodiments, a machine learning model or predefined features configured to determine the current scenario based on sensor and positioning data can be utilized. For example, given the vehicle's location and HD map data, it can be determined whether the autonomous vehicle is in an "intersection scenario" or a "lane merging scenario," whereby the overlap cost parameter can be weighted differently depending on the current scenario. Therefore, for each candidate path, the overlap cost parameter can be further based on the determined scenario. In other words, the overlap cost parameter can be weighted differently depending on the vehicle's current scenario when calculating the total cost function value for each candidate path. This can potentially help avoid unnecessary lane changes in various situations, thereby improving passenger comfort.

[0064] Similarly, the overlapping cost parameter can be weighted differently based on traffic density, another "scenario parameter" used to define the vehicle's current scenario. More specifically, the cost parameter can have a greater impact on the total cost function value in scenarios with high traffic density compared to scenarios with low traffic density. Likewise, the path smoothness used to define the nominal cost function value can be weighted differently based on road surface conditions (another scenario parameter), such that path smoothness has a higher impact on the nominal cost function value under slippery conditions (e.g., ice on the road surface) compared to non-slippery conditions (e.g., dry road surface) to prevent lateral slippage of the autonomous vehicle.

[0065] continue, Figure 6This is a schematic side view of vehicle 1, which includes a control device 10 for operating vehicle 1 according to embodiments of the present disclosure. Vehicle 1 also includes a perception system 6 and a positioning system 5. The perception system 6 is understood herein to be a system responsible for acquiring raw sensor data from sensors 6a, 6b, 6c, such as cameras, LiDAR and RADAR, and ultrasonic sensors, and converting the raw data into scene understanding. The positioning system 5 is configured to monitor the vehicle's geographic location and heading, and may be in the form of a Global Navigation Satellite System (GNSS), such as GPS. However, the positioning system may also be implemented as Real-Time Kinematic (RTK) GPS to improve accuracy. Furthermore, the positioning system may also include map data in the form of HD map data (or be configured to acquire map data from an associated memory).

[0066] The control device 10 includes one or more processors 11, a memory 12, a sensor interface 13, and a communication interface 14. The processor 11 may also be referred to as control circuitry 11 or control line 11. The control circuitry 11 is configured to execute instructions stored in the memory 12 to perform a method for controlling a vehicle according to any embodiment disclosed herein. In other words, the memory 12 of the control device 10 may include one or more (non-transitory) computer-readable storage media for storing computer-executable instructions, for example, instructions that, when executed by one or more computer processors 11, cause the computer processors 11 to perform the techniques described herein. Alternatively, the memory 12 includes high-speed random access memory, such as DRAM, SRAM, DDR RAM, or other random access solid-state memory devices; and alternatively includes non-volatile memory, such as one or more disk storage devices, optical disk storage devices, flash memory devices, or other non-volatile solid-state memory devices.

[0067] Furthermore, vehicle 1 can connect to external network 2 via, for example, a wireless link (e.g., for retrieving map data). The same or other wireless links can be used to communicate with other vehicles or local infrastructure components in the vicinity. Cellular communication technology can be used for long-range communication, such as to external networks, and if the cellular communication technology used has low latency, it can also be used for communication between vehicles, vehicle-to-vehicle (V2V), and / or vehicle-to-infrastructure (V2X). Examples of cellular radio technologies are GSM, GPRS, EDGE, LTE, 5G, 5G NR, etc., and also include future cellular solutions. However, some solutions use short-to-medium range communication technologies, such as wireless local area networks (LANs), for example, solutions based on IEEE 802.11. ETSI is working on cellular standards for vehicle communication, and 5G, for example, is considered a suitable solution due to its low latency, high bandwidth, and efficient processing of communication channels.

[0068] This disclosure has been presented above with reference to specific embodiments. However, other embodiments besides those described above are also possible and are within the scope of this disclosure. Different method steps than those described above may be provided within the scope of this disclosure, and the method may be performed by hardware or software. Therefore, according to an exemplary embodiment, a non-transitory computer-readable storage medium is provided storing one or more programs configured to be executed by one or more processors of a vehicle control system, the one or more programs including instructions for performing the method according to any of the above embodiments. Alternatively, according to another exemplary embodiment, a cloud computing system may be configured to perform any of the methods presented herein. The cloud computing system may include distributed cloud computing resources that collectively perform the methods presented herein under the control of one or more computer program products.

[0069] Generally, computer-accessible media can include any tangible or non-transitory storage medium or storage medium, such as electronic, magnetic, or optical media, like a disk or CD / DVD-ROM coupled to a computer system via a bus. The terms “tangible” and “non-transitory” as used herein are intended to describe computer-readable storage media (or “memory”) that do not transmit electromagnetic signals, but are not intended to otherwise limit the types of physical computer-readable storage devices included by the terms computer-readable media or memory. For example, the terms “non-transitory computer-readable media” or “tangible memory” are intended to cover types of storage devices that do not necessarily permanently store information, such as random access memory (RAM). Program instructions and data stored in a non-transitory form on tangible computer-accessible storage media can also be transmitted via transmission media or signals such as electrical, electromagnetic, or digital signals, which can be transmitted via communication media such as networks and / or wireless links.

[0070] Processor 11 (associated with control system 10) may be or include any number of hardware components for performing data or signal processing or for executing computer code stored in memory 12. Apparatus 10 has an associated memory 12, and memory 12 may be one or more means for storing data and / or computer code used to perform or implement the various methods described herein. Memory may include volatile or non-volatile memory. Memory 12 may include database components, object code components, script components, or any other type of information structure for supporting the various activities described herein. According to exemplary embodiments, any distributed or local storage device may be used with the systems and methods described herein. According to exemplary embodiments, memory 12 may be communicatively connected to processor 11 (e.g., via circuitry or any other wired, wireless, or network connection) and includes computer code for performing one or more processes described herein.

[0071] It should be understood that sensor interface 13 can also provide the possibility of acquiring sensor data directly or via dedicated sensor control circuitry 6 in the vehicle. Communication / antenna interface 14 can also provide the possibility of transmitting output to a remote location (e.g., a remote operator or control center) via antenna 8. Furthermore, some sensors in the vehicle can communicate with the control system using local network settings such as CAN bus, I2C, Ethernet, or fiber optics. Communication interface 14 can be arranged to communicate with other control features of the vehicle and can therefore also be considered a control interface; however, a separate control interface (not shown) can also be provided. Local communication within the vehicle can also be wireless, using protocols such as Wi-Fi, LoRa, Zigbee, Bluetooth, or similar medium / short-range technologies.

[0072] Therefore, it should be understood that parts of the described solution can be implemented in vehicle 1, in a system located outside vehicle 1, or in a combination of inside and outside vehicle 1; for example, in a server communicating with vehicle 1, i.e., a so-called cloud solution. For example, sensor data can be sent to an external system, and this system performs some or all of the necessary steps to select the best candidate path. Different modules and steps of the embodiments can be combined in other combinations different from those described.

[0073] Although the accompanying drawings may show a specific order of method steps, the order of steps may differ from that described. Furthermore, two or more steps may be performed simultaneously or partially simultaneously. This variation will depend on the chosen software and hardware system and the designer's choices. All such variations are within the scope of this disclosure. Similarly, software implementation can be accomplished using standard programming techniques with rule-based logic and other logic to implement various connection steps, processing steps, comparison steps, and decision steps. The embodiments mentioned and described above are given by way of example only and should not be limited to this disclosure. Other solutions, uses, objectives, and features within the scope of the disclosure claimed by the following patent embodiments should be apparent to those skilled in the art.

Claims

1. A method for operating a vehicle having an autonomous driving system, ADS, and a fallback stop feature, the method comprising: obtaining sensor data and localization data, the sensor data and localization data comprising information about a surrounding environment of the vehicle; determining, based on the sensor data and the localization data, a plurality of candidate paths within a drivable area in the surrounding environment of the vehicle for a prediction time horizon, each candidate path being associated with a nominal cost function value based on at least one cost parameter; determining, based on the obtained sensor data and localization data, an expected trajectory of a target vehicle located in the surrounding environment of the vehicle for the prediction time horizon; determining, for each candidate path, an overlap cost parameter regarding an overlap between the expected trajectory of the target vehicle and a set of stop positions of the vehicle based on a predicted execution of the fallback stop feature within the prediction time horizon; selecting a candidate path from the plurality of candidate paths by computing a cost function based on the nominal cost function value and the overlap cost parameter; and generating a control signal at an output to control the vehicle to follow the selected candidate path. The predicted execution of the fallback stop feature comprises a predicted execution of the fallback stop feature within each candidate path.

2. The method of claim 1, wherein, The set of stop positions comprises at least two stop positions based on a predicted execution of the fallback stop feature within the prediction time horizon.

3. The method of claim 1 or 2, wherein, The at least one cost parameter is based on at least one of a proximity to a waypoint, a path distance, and a path smoothness.

4. The method of claim 1 or 2, wherein, The step of selecting the candidate path from the plurality of paths comprises:

5. The method of claim 1 or 2, wherein, determining a total cost function value for each candidate path based on the nominal cost function value and the overlap cost parameter; selecting the candidate path associated with a lowest total cost function value.

6. The method according to claim 1 or 2, further comprising: determining a scenario from a plurality of predefined scenarios based on the obtained sensor data and localization data; wherein the step of determining the overlap cost parameter for each candidate path is further based on the determined scenario.

7. A computer-readable storage medium storing one or more programs configured to be executed by one or more processors of a vehicle control system, the one or more programs comprising instructions for performing the method according to any one of claims 1 to 6.

8. A control device for operating a vehicle having an autonomous driving system, ADS, and a fallback stop feature, the control device comprising a control circuit configured to: obtain sensor data and localization data, the sensor data and localization data comprising information about a surrounding environment of the vehicle; determine, based on the sensor data and the localization data, a plurality of candidate paths within a drivable area in the surrounding environment of the vehicle for a prediction time horizon, each candidate path being associated with a nominal cost function value based on at least one cost parameter; ​ determining, based on the acquired sensor data and positioning data, an expected trajectory of a target vehicle located in the surrounding environment of the vehicle for the prediction time horizon; determining, for each candidate path, an overlap cost parameter regarding an overlap between the expected trajectory of the target vehicle and the set of stop locations of the vehicle based on a predicted execution of the fallback stop feature within the prediction time horizon; selecting a candidate path from the plurality of candidate paths by computing a cost function based on the nominal cost function value and the overlap cost parameter; and generating a control signal at an output to control the vehicle to follow the selected candidate path. The predicted execution of the fallback stop feature comprises a predicted execution of the fallback stop feature within each candidate path.

9. The control device of claim 8, wherein, The set of cumulative stop locations comprises at least two stop locations based on a predicted execution of the fallback stop feature within the prediction time horizon.

10. The control device according to claim 8 or 9, wherein The at least one cost parameter is based on at least one of proximity to a waypoint, path distance, and path smoothness.

11. The control device according to claim 8 or 9, wherein The control circuit is configured to select the candidate path from the plurality of paths by:

12. The control device according to claim 8 or 9, wherein determining a total cost function value for each candidate path based on the nominal cost function value and the overlap cost parameter; selecting the candidate path associated with the lowest total cost function value. The control circuit is further configured to determine a scenario from a plurality of predefined scenarios based on the acquired sensor data and positioning data; and 13. The control device according to claim 8 or 9, wherein wherein, for each candidate path, the overlap cost parameter is further determined based on the determined scenario.

14. A vehicle comprising: a perception system comprising at least one sensor configured to monitor a surrounding environment of the vehicle; a positioning system configured to monitor a geographic map position of the vehicle; a control device according to any one of claims 8 to 13. ​ ​

Images