Search

DE-102024210758-A1 - Procedures for perception plausibility and procedures for perception

DE102024210758A1DE 102024210758 A1DE102024210758 A1DE 102024210758A1DE-102024210758-A1

Abstract

The invention relates to a method for perceiving plausibility (36) a perception of at least one environmental object (12) in a vehicle environment (14) of a vehicle (16), comprising the steps of providing object parameters (18) determined during an environmental perception (32) including an object motion tracking (34) of at least the environmental object (12) for the environmental object (12) depending on sensor data (22) of at least one environmental sensor (24) of the vehicle (16) capable of detecting the environmental object (12), calculating (38) comparison object parameters (40) depending on the object parameters (18) by applying a physical motion model (42) of the environmental object (12), and calculating (50) at least one plausibility parameter (52) indicating a plausibility of the object parameters (18) depending on a comparison (53) of the object parameters (18) with the comparison object parameters (40). Furthermore, the invention relates to a method for perception (10).

Inventors

  • Wilhelm Christopher von Rosenberg
  • Christoph Kallenberger
  • Stephan Morgen
  • Nico Pfahl
  • Marie-Elise Janoir-Loesch

Assignees

  • Robert Bosch Gesellschaft mit beschränkter Haftung

Dates

Publication Date
20260513
Application Date
20241108

Claims (10)

  1. A method for the plausibility check (36) of a perception of at least one environmental object (12) in a vehicle environment (14) of a vehicle (16), comprising the steps: providing object parameters (18) determined during environmental perception (32) including object motion tracking (34) of at least the environmental object (12) for the environmental object (12) depending on sensor data (22) of at least one environmental sensor (24) of the vehicle (16) capable of detecting the environmental object (12); calculating (38) comparison object parameters (40) depending on the object parameters (18) by applying a physical motion model (42) of the environmental object (12); calculating (50) at least one plausibility check parameter (52) indicating a plausibility check of the object parameters (18) depending on a comparison (53) of the object parameters (18) with the comparison object parameters (40).
  2. Procedure for perception plausibility (36) according to Claim 1 , characterized in that as soon as the plausibility parameter (52) lies outside of a predefined validity range (60), an error message (66) is issued.
  3. Procedure for perception plausibility (36) according to Claim 1 or 2 , characterized in that a sensor data fusion of sensor data as first sensor data (22) of the environment sensor as first environment sensor (24) and second sensor data (28) of a second environment sensor (30) which can also detect the environment object (12) at least partially is carried out.
  4. Procedure for perception plausibility (36) according to Claim 3 , characterized in that at least one assignment parameter (80) describing a match of the environmental object (12) detectable via the first and second sensor data (22, 28) is calculated.
  5. Procedure for perception plausibility (36) according to Claim 4 , characterized in that as soon as the assignment parameter (80) is except If the value is half of a specified validity range (84), an error message (86) is issued.
  6. Method for perception plausibility (36) according to one of the preceding claims, characterized in that the object parameters (18) comprise an object position, an object velocity and/or an object acceleration and/or the comparison object parameters (40) comprise at least a comparison object position, a comparison object velocity and/or a comparison object acceleration.
  7. Method for perception plausibility check (36) according to one of the preceding claims, characterized in that the provision of the object parameters (18) takes place in temporally successive steps (20) and the comparison object parameters (40) of a current step are calculated at least depending on object parameters (18) of a preceding step.
  8. Method (10) for perceiving at least one environmental object (12) in a vehicle environment (14) of a vehicle (16), comprising the steps of determining object parameters (18) for the at least one environmental object (12) with an environment perception (32) including an object movement tracking (34) depending on sensor data (22) of at least one environment sensor (24) of the vehicle (16) capable of detecting the environmental object (12), calculating (50) a plausibility parameter (52) by a method for perception plausibility (36) according to one of the preceding claims, and carrying out a measure (70, 88, 94) influencing the processing of the object parameters (18) at least depending on the plausibility parameter (52).
  9. Procedure for perception (10) according to Claim 8 , characterized in that the measure is a deletion (88) of object parameters (18) and/or a change (94) of at least one vehicle control function of the vehicle (16).
  10. Procedure for perception (10) according to Claim 8 or 9 , characterized in that the measure (70) includes a filtering of the object parameters (18) and an output of filtered object parameters (74).

Description

The invention relates to a method for perception plausibility according to claim 1. Furthermore, the invention relates to a method for perception. State of the art In driver assistance systems and autonomous driving, perception—that is, the recognition and perception of objects in the vehicle's environment—depends on sensor data from environmental sensors and is crucial for vehicle safety. A key objective is to avoid false positives, i.e., false alarms that lead to unnecessary intervention. The safety objectives for perception primarily concern the control and limitation of vehicle deceleration. For example, the vehicle should only decelerate when this is intentional and safety-relevant. Deceleration should cease as soon as there is no longer any immediate danger. Furthermore, the vehicle must not brake when it is traveling faster than the maximum speed specified for the respective function. This ensures that the function is only activated under the correct conditions. To achieve these security goals, monitoring of environmental sensors is a well-known approach. Fault detection and protection mechanisms are designed so that if a sensor signal is faulty, for example from a camera in a video surveillance system, the affected sensor data is ignored. If sensor faults persist, a fault management system initiates a so-called functional degradation, ultimately leading to a complete shutdown. Some environmental sensors also have algorithms that detect failures or incorrect sensor orientations and implement appropriate corrective actions. Disclosure of the invention According to the present invention, a method for perception plausibility testing with the features of claim 1 is proposed. This allows implausible object parameters to be excluded and driving safety of the vehicle to be increased. The influence of false detections on the vehicle's operation can be reduced or prevented. The vehicle can be a motor vehicle, truck, two-wheeled vehicle, mobile device or mobile robot. Perception can be implemented in a driver assistance system and/or an automated driving function, in particular an autonomous driving function, of the vehicle. Environmental perception, including object tracking, or at least object tracking, can employ artificial intelligence algorithms and/or classical approaches such as Kalman filtering. The calculation of the comparison object parameters and/or the plausibility parameter can be performed during environmental perception, particularly during object tracking, or subsequently. The environmental sensor can be a camera, a LiDAR sensor, a radar sensor, or an ultrasonic sensor. The surrounding object can be another vehicle, a living being, a plant, a device, or a building. The object parameters can specify the presence and/or at least one property or state, for example a classification, size and/or distance, of the environment object. Determining the object parameters can include establishing the presence of at least one environment object. Object motion tracking can be the detection of a movement of the surrounding object, in particular including a path traveled by the surrounding object. The physical motion model can simulate the motion relationships of the surrounding object using physical laws. For example, the position of the surrounding object can be estimated, at least as a function of its detected velocity. The calculation of the reference object parameters can be performed in a coordinate system that is relative to the vehicle or absolute. The physical motion model can be applied during environmental perception, particularly during object motion tracking, for example, when using a Kalman filter. The comparison object parameters can be the parameters estimated by the Kalman filter. Depending on the innovation, the plausibility parameter can be calculated as the difference between the object parameters and the comparison object parameters. The parameter can be calculated depending on at least one predefined tolerance threshold of the innovation. For example, the plausibility parameter P = 1 can be if the innovation lies within a predefined tolerance range spanned at least by the tolerance threshold, and otherwise P = 0. The plausibility check of the object parameters can include an evaluation of the correctness, accuracy and/or consistency of the object parameters. The calculation of the comparison object parameters and the calculation of the plausibility check can be performed in real time. In a preferred embodiment of the invention, it is advantageous that an error message is issued as soon as the plausibility parameter lies outside a predefined validity range. The predefined validity range can be defined by a lower and/or upper limit. The error message can be output to a processing unit, in particular a driving state module of the vehicle. The driving state module can manage a driving state of the vehicle. In a specific embodiment of the invention, it is advantageous to perform sensor data f