Computer-implemented method, processing device and vehicle control device

Abstract

The invention relates to a computer-implemented method (10) for behavior planning of a vehicle (12), comprising at least acquiring and processing (16) environmental sensor data (18) of at least one environmental sensor (22) acquiring a vehicle environment (20) of the vehicle (12) for environmental perception (24), including creating an environmental model (30), providing (32) safety requirements (34) for environmental perception (24) at a first time, providing (36) safety capabilities (38) of environmental perception (24) estimated at the first time, calculating (40) at least one proposed operating parameter (42) of the vehicle (12) that reduces a deviation between the safety capabilities (38) and the safety requirements (34), and planning (44) at least one operating parameter (45) of the vehicle (12) depending on the proposed operating parameter (42).Furthermore, the invention relates to a processing device (14) and a vehicle control device (50).

Description

[0001] The invention relates to a computer-implemented method for behavior planning of a vehicle according to claim 1. Furthermore, the invention relates to a processing device and a vehicle control device. State of the art

[0002] In the field of autonomous vehicles, the continuous detection and interpretation of the vehicle's surroundings is crucial. Fundamental to this is environmental sensing, which uses sensors such as LIDAR, radar, ultrasound, and cameras to collect information about the vehicle's surroundings. This information is contained in environmental sensor data and is processed by environmental perception, which allows objects to be detected, classified, and their movements to be predicted. Environmental perception enables a comprehensive and detailed picture of the vehicle's surroundings, which is crucial for safe vehicle operation.

[0003] The functions of autonomous driving can be divided into three levels. The perception level focuses on environmental perception, including the interpretation of the vehicle's surroundings using an environmental model. Based on this, the planning level uses a behavior planner to make decisions about vehicle behavior and a trajectory planner to determine the vehicle's movement trajectory based on the perception of the surroundings. Finally, the action level implements the planned actions in the vehicle, for example, by controlling actuators.

[0004] The behavior planner plans the tactical behavior, for example actions such as "follow lane" or "change lane" for longer periods of time, for example 8 seconds. The trajectory planner plans the operative implementation of the specific trajectory with destination and speed for shorter periods of time, for example 3 seconds. Both planners calculate several alternatives over these periods and evaluate them according to predefined cost functions. One criterion that is checked here is the dynamically safe driveability of a driving maneuver. This is often determined in the downstream "motion control system" functional module for implementing the planned actions, taking into account, among other things, the current and future dynamic state of the vehicle, the states of the actuators and the nature of the environment, e.g. friction values.

[0005] However, one of the biggest challenges of automated driving is managing unavoidable functional deficiencies in environmental perception, particularly according to ISO / PAS 21448. ISO / PAS 21448 relates to road vehicles and the safety of the intended functionality. It is related to SOTIF (Safety of the Intended Functionality). It addresses safety risks that can arise when the vehicle functions as intended but may still act unsafely due to unexpected events or scenarios in the environment. This is particularly relevant for driver assistance systems, especially ADAS and autonomous driving systems, where the complexity and potential for unforeseen scenarios are higher.

[0006] Such deficiencies include, for example, an inability of the function to correctly detect and safely process the situation, including functions that use machine learning algorithms, or insufficient robustness of the function with regard to sensor input fluctuations or different environmental conditions.

[0007] A common approach is to identify potential limitations and vulnerabilities through analysis or by recording data from real-world driving, and to minimize them by adapting the environmental sensors and the perception system, including the algorithms. Furthermore, approaches are known to identify limitations during operation, for example, due to adverse weather conditions, and to mitigate them adaptively or rule-based, for example, by reducing speed. Disclosure of the invention

[0008] According to the present invention, a computer-implemented method for planning the behavior of a vehicle is proposed, comprising the features of claim 1. This allows safer vehicle behavior to be planned even with limited environmental perception. The planned operating parameter can be calculated more easily and quickly. The search space for an optimal planned operating parameter to reduce a discrepancy between safety requirements and safety capabilities can be narrowed down more precisely.

[0009] The vehicle can be a motor vehicle, a truck, a two-wheeled vehicle, a mobile robot, or an industrial robot. The vehicle can assume at least a semi-autonomous or autonomous operating state during driving.

[0010] The behavior planning may include vehicle behavior planning of a vehicle behavior of the vehicle and / or trajectory planning of a movement trajectory of the vehicle.

[0011] The environment sensor can be a LIDAR sensor, a radar sensor, an ultrasonic sensor, and / or a camera. The environment sensor data can be provided by multiple environment sensors. The environment sensors can all be assigned to the same or at least two different sensor instances.

[0012] The vehicle environment can include, preferably exclusively, the immediate surroundings of the vehicle.

[0013] The sensing may be environmental sensing. The environmental sensor data may be provided by the environmental sensing. During environmental sensing, raw data from the physical environment may be acquired using the at least one environmental sensor. The capability, reliability, and / or integrity of the environmental sensing may depend on a range, a resolution, and / or a weather condition. The environmental sensing may at least partially involve the processing of the environmental sensor data. In this regard, ASIL (Automotive Safety Integrity Level) according to ISO 26262, which refers to the probability that internal processing is faulty due to an electrical / electronic failure, may be applicable.

[0014] The vehicle environment detected by the environment sensor may include at least one environment object. The environment object may be a building, a living being, a device, an object, another vehicle, a plant, or a road surface.

[0015] Environment perception can process and interpret the environment sensor data provided by the environment sensor. This can include attempting to understand the meaning of the information contained in the environment sensor data about the vehicle's surroundings and the at least one environment object. For example, it can be recognized whether the environment object is a car, a pedestrian, or a bicycle and / or whether it is moving or stationary. Machine learning and computer vision techniques can be used in environment perception, in particular to recognize patterns in the environment sensor data and / or to correctly classify the at least one environment object. The goal of environment perception is a reliable and accurate representation of the environment in real time. Environment perception can comprise object detection and / or object tracking of at least one environment object.

[0016] The environment model can be created and / or modified using the environment perception. The environment model can be a real-time representation of the vehicle's environment. The environment model can include an environment map or a 3D model of the vehicle's environment. The environment model can contain information about stationary environmental objects (e.g., buildings, trees, roads) and / or about moving objects (e.g., other vehicles, pedestrians, animals) in the vehicle's environment.

[0017] The planned operating parameter can be an emergency operating parameter applied in a fallback solution to maximize driving safety, for example, during a Minimal Risk Maneuver (MRM). The planned emergency operating situation can relate to the current situation or future situations. The planned operating parameter can be the operating parameter to be applied at the initial point in time or at a different point in time, particularly one preceding or following the initial point in time.

[0018] The planning of the operating parameter can be part of behavioral planning and / or trajectory planning. The planning of the operating parameter can be part of behavioral planning and / or trajectory planning.

[0019] The operating parameter can be a vehicle parameter, a driving operation parameter of the vehicle's driving operation, and / or a movement parameter of the vehicle. The operating parameter can actually be adjustable and / or variable by at least one actuator. The operating parameter can be a longitudinal speed, longitudinal acceleration, lateral speed, lateral acceleration, and / or a vehicle position. The operating parameter can be a signal, in particular a visual, acoustic, and / or vibrating signal.

[0020] The planned operating parameter may depend on a current environmental perception of the currently existing vehicle environment.

[0021] The proposed operating parameter can be a speed, in particular a maximum speed or a minimum speed, a lateral lane position, for example, left, center, or right, a vertical position, a relative position to a surrounding object, a vehicle angle, and / or a maneuver to be avoided. The proposed operating parameter can comprise a sequence of actions for several operating parameters. The proposed operating parameter can be a target value of the operating parameter to be planned.

[0022] The safety requirements for environmental perception can be predefined safety conditions for environmental detection by the at least one environmental sensor, for processing the environmental sensor data, and / or for environmental perception and the environmental model. The safety requirements can be derived from an "abstract scenario" that is recognized as present from the environmental model. The safety requirements can be derived from a scenario database. For example, a complex scenario such as an intersection with oncoming traffic can be composed of sub-processes, such as "intersection," "left turn," "oncoming traffic," "pedestrian crossing," and "obstacle in the target lane," for which specific safety requirements, such as the visibility ranges to be guaranteed, can then be derived based on rules.

[0023] The safety capabilities of the environment perception can be the performance, reliability, and / or availability of the environment detection by the at least one environment sensor, the processing of the environment sensor data, and / or the environment perception and the environment model. The safety capability can include or be safety integrity (in particular according to ISO 26262). The safety capabilities can include classifying and / or quantifying statistical parameters, for example, defined by the relevant standard ISO / PAS 8800.

[0024] The safety requirements and / or safety capabilities may include or be completeness of the environmental information, in particular of all relevant environmental objects and environmental situations, accuracy, reliability, in particular correct classification, status detection and / or movement detection of the environmental object, robustness against external influences, for example against weather influences or lighting conditions and / or redundancy.

[0025] The safety requirements and / or safety capabilities can be limited to a predefined area of ​​the vehicle's surroundings. The surrounding area can be a sub-area of ​​a 3D grid that divides the vehicle's surroundings, particularly regularly.

[0026] The security capabilities may depend on a state of environmental perception, possible constraints, such as occlusions, environmental conditions, such as weather conditions, visibility conditions, resource capacity, system states, errors, and / or out-of-distribution detection or anomaly detection. The security capabilities may be calculated and provided taking into account at least one of these parameters.

[0027] The safety capabilities and / or safety requirements may include the quality of environmental perception, the redundancy of the environmental sensor data, and / or the integrity of the at least one environmental sensor. The safety capabilities and / or safety requirements may include the availability of the at least one environmental sensor and / or the signal density of the environmental sensor data.

[0028] The safety capabilities and requirements can be provided depending on the environmental perception. The safety requirements can be provided depending on the planned operating parameters.

[0029] Security capabilities may be spatially dependent.

[0030] The planned operating parameter can be the same as or close to the proposed operating parameter. Proposed operating parameters can be calculated, for example: a maximum speed depending on the current visibility of individual or multiple environmental sensors, in particular with regard to the braking distance with a known coefficient of friction, a relative or absolute position and / or pose to be avoided, in particular an Euler angle of the vehicle and / or the environment sensor, preferably at the first point in time or a point in time deviating therefrom, in order to avoid impairment due to glare or obscuration under given environmental conditions, for example position of the sun, reflective surfaces (puddles), reflections, radar / lidar-reflecting surfaces and / or obscuration by stationary or dynamic objects, a track position, for example, where the sun is obscured by other objects, especially houses, a final position before a stop line in which the visibility of a traffic light or the view of an intersection is given with sufficient redundancy and quality, a lateral lane position from which the visibility behind surrounding objects at the roadside, such as parked vehicles, advertising columns, vegetation, is improved, for example for pedestrians possibly approaching from this area, a vertical vehicle position, especially with active suspension, in which visibility is sufficient for the planned vehicle speed, for example on a given road gradient, a relative vehicle position to a stationary or dynamic surrounding object, in particular to avoid obscuring, for example, a traffic light and / or a curve trajectory at a confusing intersection with many pedestrians at the side of the road.

[0031] The proposed operating parameter can be calculated model-based, in particular depending on the environmental model and / or by applying machine learning.

[0032] In a preferred embodiment of the invention, it is advantageous if a safety comparison of the safety requirements with the safety capabilities is carried out. This allows the deviation between the safety capabilities and the safety requirements to be objectively recorded and evaluated. The safety comparison can provide an evaluation measure of a deviation between the safety requirements and the safety capabilities. The evaluation measure can be the degree of deviation between the safety requirements and the safety capabilities.

[0033] In a preferred embodiment of the invention, it is advantageous if the planning of the operating parameter depends on the safety comparison and the proposed operating parameter. The proposed operating parameter can limit the range of possible operating parameters to be planned. The proposed operating parameter can be a suggestion for an operating parameter to be planned, in particular a target value or setpoint value to which the operating parameter to be planned can be based.

[0034] In a preferred embodiment of the invention, the proposed operating parameter is calculated based on the safety comparison. The proposed operating parameter can be calculated iteratively. If the deviation between the safety capabilities and the safety requirements falls below the threshold, the proposed operating parameter can be omitted. Preferably, the proposed operating parameter can only be calculated when the deviation between the safety capabilities and the safety requirements is reached or exceeded.

[0035] A preferred embodiment of the invention is advantageous in which the planning of the operating parameter is a planning of an operating parameter to be set at the first time. The planning of the operating parameter can also be a planning of an operating parameter to be set at a time different from the first time.

[0036] In an advantageous embodiment of the invention, the proposed operating parameter is calculated based on the environmental model. This allows the proposed operating parameter to be calculated according to the situation. The proposed operating parameter can take dynamic processes in the vehicle's environment into account.

[0037] In a specific embodiment of the invention, it is advantageous if the proposed operating parameter is calculated based on an environmental map. The environmental map can be a three-dimensional map. The environmental map can include dynamic environmental objects.

[0038] In a specific embodiment of the invention, it is advantageous if the proposed operating parameter is calculated depending on the location of the vehicle in the surrounding map. The proposed operating parameter can be dependent on the location of the vehicle in the surrounding map at the first time or at a different time.

[0039] According to the present invention, a processing device for a vehicle is further proposed, which is configured to execute the computer-implemented method with at least one of the previously described features. The processing device can have at least one processor for at least partially executing the method.

[0040] According to the present invention, a vehicle control device for a vehicle is further proposed, which is configured to adjust the planned operating parameter calculated using a computer-implemented method having at least one of the previously described features. The vehicle control device can be a control unit or a part of a control unit.

[0041] At least one method step of the computer-implemented method can be performed in the vehicle using edge computing and / or outside the vehicle using cloud computing. At least one method step of the computer-implemented method can be processed using machine learning. The machine learning can use a neural network. The machine learning can include reinforcement learning.

[0042] Furthermore, the invention relates to a non-transitory, computer-readable medium containing instructions, the execution of which causes at least one processor to carry out said computer-implemented method with at least one previously described feature.

[0043] The planned operating parameter can be set at the first time.

[0044] Further advantages and advantageous embodiments of the invention emerge from the description of the figures and the illustration. Character description

[0045] The invention is described in detail below with reference to the figure. Fig.1 shows a computer-implemented method in a specific embodiment of the invention. The computer-implemented method 10 of a vehicle 12 can be executed by a processing device 14 of the vehicle 12 and comprises acquiring, as environment sensing, and processing 16 of environment sensor data 18 from at least one environment sensor 22 acquiring a vehicle environment 20 of the vehicle 12, a subsequent environment perception 24 of the vehicle environment 20 including at least one environmental object 26, here a pedestrian 28 in front of the vehicle 12, including the creation of an environment model 30 depending on the environment sensor data 18, providing 32 safety requirements 34 valid at a first time to the environment perception 24, and providing 36 estimated safety capabilities 38 of the environment perception 24 effective at the first time.a calculation 40 of at least one proposed operating parameter 42 of the vehicle 12 reducing a deviation between the safety capabilities 38 and the safety requirements 34 and a planning 44 of at least one operating parameter 45 of the vehicle 12, in particular at the first time, depending on the proposed operating parameter 42.

[0046] The safety capabilities 38 and safety requirements 34 are provided depending on the environmental perception 24. By comparing the safety requirements 34 with the safety capabilities 38, the planning 44 of the operating parameter 45 can be carried out depending on the safety comparison 46. The safety requirements 34 are, in turn, provided depending on the planned operating parameter 45, which in turn is based on the environmental model 30.

[0047] The planning 44 of the operating parameter 45 depends on the safety comparison 46 and the proposed operating parameter 42. The proposed operating parameter 42 is calculated based on the safety comparison 46. The safety comparison 46 can, for example, output a capability comparison map 48 on which the proposed operating parameter 42 depends. The proposed operating parameter 42 is further calculated based on the environmental model 30, in particular, based on an environmental map and a localization of the vehicle 12 in the environmental map.

[0048] The planned operating parameter 45 is actually adjustable and / or changeable by a vehicle control device 50 with a motion control of at least one actuator.

Claims

[1] Computer-implemented method (10) for behavior planning of a vehicle (12), at least comprising Acquiring and processing (16) environmental sensor data (18) of at least one environmental sensor (22) that detects a vehicle environment (20) of the vehicle (12) for Perception of the environment (24) including the creation of an environment model (30), Providing (32) security requirements (34) to the environmental perception (24) at a first time, Providing (36) safety capabilities (38) of the environment perception (24) estimated at the first time, Calculating (40) at least one proposed operating parameter (42) of the vehicle (12) that reduces a deviation between the safety capabilities (38) and the safety requirements (34) and Planning (44) of at least one operating parameter (45) of the vehicle (12) depending on the proposed operating parameter (42). [2] Computer-implemented method (10) according to claim 1, characterized by that a safety comparison (46) of the safety requirements (34) with the safety capabilities (38) is carried out. [3] Computer-implemented method (10) according to claim 2, characterized by that the planning (44) of the operating parameter (45) depends on the safety comparison (46) and the proposed operating parameter (42). [4] Computer-implemented method (10) according to claim 2 or 3, characterized by that the proposed operating parameter (42) is calculated depending on the safety comparison (46). [5] Computer-implemented method (10) according to one of the preceding claims, characterized by that the planning (44) of the operating parameter (45) is a planning (44) of an operating parameter to be set at the first time. [6] Computer-implemented method (10) according to one of the preceding claims, characterized bythat the proposed operating parameter (42) is calculated depending on the environmental model (30). [7] Computer-implemented method (10) according to one of the preceding claims, characterized by that the proposed operating parameter (42) is calculated depending on an environment map. [8] Computer-implemented method (10) according to one of the preceding claims, characterized by that the safety requirements (34) are provided depending on the environmental perception (24) and the planned operating parameter (45). [9] Processing device (14) for a vehicle (12) which is arranged to carry out the computer-implemented method (10) according to one of the preceding claims. [10] Vehicle control device (50) for a vehicle (12) which is configured to adjust the operating parameter (45) of the vehicle (12) planned using a computer-implemented method (10) according to one of claims 1 to 8.

Images