Parking Assistance Trajectory Retracing False Positive Curb Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Parking assistance systems equipped with ultrasonic sensors often incorrectly detect curbs or thresholds as obstacles during retracing of learned trajectories, leading to abortion of the retracing process, which is annoying for users.
Innovation Solution
The method involves using inclination angle sensors to identify positions on the trajectory where false-positive obstacle detections are likely, such as curbs, and adjusting the system's behavior during retracing by switching to a second rule set that increases the threshold for obstacle detection reliability, suppressing false-positive signals, and allowing the vehicle to drive over detected obstacles if they are non-hazardous.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ultrasonic sensors are used to detect obstacles during retracing, then obstacle detection capability is improved, but false-positive detections of curbs as obstacles occur
Solution Approach 1:
The patent introduces an intermediary evaluation mechanism that mediates between the ultrasonic sensor's obstacle detection and the actual obstacle avoidance action. The system uses additional sensor data (cameras, LIDAR, radar) and environmental context information as intermediaries to verify whether a detected obstacle is a true obstacle or a false positive (such as a curb or threshold), thereby resolving the contradiction between detection reliability and precision.
Solution Approach 2:
The system dynamically changes detection parameters based on the vehicle's state and environment. When the vehicle is in retracing mode and approaches identified curb locations (from the learned trajectory), the system adjusts the obstacle detection thresholds and confidence levels, making it less likely to trigger obstacle avoidance for objects that are known to be drivable features like curbs and thresholds.
2Reliability
If the system aborts retracing upon detecting obstacles, then collision avoidance is improved, but smooth retracing is deteriorated due to false positives
Solution Approach 1:
The system performs preliminary identification of curbs and thresholds during the learning phase of trajectory acquisition. These identified features are stored as part of the trajectory data. During retracing, when the vehicle approaches these pre-identified locations, the system proactively adjusts its obstacle detection and response behavior, preventing false-positive abortions before they occur.
Solution Approach 2:
The system implements a feedback mechanism where the outcome of obstacle detections near known curb locations is used to adjust future detection behavior. When the vehicle successfully passes over a detected obstacle without issue during manual driving, the system learns from this feedback and adjusts its obstacle detection sensitivity for similar future situations, reducing false positives while maintaining collision avoidance.
3Reliability
If the system uses strict obstacle detection rules, then safety is improved, but false-positive detections increase
Solution Approach 1:
The patent applies different detection rules and sensitivity levels in different spatial locations along the trajectory. In areas where curbs and thresholds are known to exist (identified during learning), the system uses more lenient detection rules with higher thresholds for triggering obstacle avoidance. In other areas without known drivable features, the system maintains strict detection rules to ensure safety, thus achieving both safety and reduced false positives through location-specific quality adjustment.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces the frequency of erroneous obstacle detections near curbs while maintaining reliable detection of actual obstacles, ensuring smooth and accurate retracing of learned trajectories.
Implementation Method 1
receiving an inclination angle sensor signal from a sensor unit of the vehicle, which signal is indicative of an inclination angle of the vehicle or a roadway and/or a change in the inclination angle of the vehicle or the roadway
Implementation Method 2
receiving an environment sensor signal from an environment sensor unit of the vehicle, which signal is indicative of an obstacle arranged in the direction of travel of the vehicle
Data Source
AI summary
A method for operating a parking assistance system for a vehicle is disclosed. The parking assistance system is configured to learn a manually traveled trajectory in a learning mode, and to automatically retrace the learned trajectory with the vehicle in a retracing mode. The learning mode includes receiving an inclination angle sensor signal from a sensor unit of the vehicle, ascertaining a curb at a specific position along the trajectory depending on the received inclination angle sensor signal. The retracing mode includes receiving an environmental sensor signal from an environment sensor unit of the vehicle, ascertaining the position of the obstacle on the learned trajectory depending on the received environmental sensor signal, and performing the retracing according to a first or second rule set depending on a comparison of the ascertained position and the position determined in the learning mode.