Vehicle Stationary Object Detection for Pre-Filtering Roadside Barriers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for detecting stationary objects in vehicles are inefficient in distinguishing truly stationary objects from temporarily stationary objects, leading to increased data processing loads and potential errors.
Innovation Solution
A computer-implemented method that analyzes groups of sensor detections to identify truly stationary objects by leveraging their regular physical attributes, such as those of guard rails or road barriers, and removes them in a batch-processing pre-processing stage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all sensed detections are processed in detail to distinguish stationary objects from temporarily stationary objects, then detection reliability is improved, but data processing time and computational load increase significantly
Solution Approach 1:
The patent segments the detection process into two distinct stages: a pre-processing stage that quickly identifies and removes clearly stationary objects (like guard rails and barriers) using simplified criteria, and a main processing stage that performs detailed analysis only on remaining detections. This segmentation allows the system to handle the large volume of stationary object detections efficiently without compromising the reliability of distinction between stationary and temporarily stationary objects.
Solution Approach 2:
The patent applies preliminary action by performing a quick pre-processing analysis before the main detection process. In this pre-processing stage, the system uses basic criteria (such as lateral position relative to the vehicle path and detection consistency) to identify and remove obviously stationary objects like guard rails and barriers. This preliminary removal reduces the data volume for subsequent detailed processing, thereby reducing overall processing time while maintaining detection reliability.
2Productivity
If simplified detection methods are used to reduce processing load, then processing efficiency is improved, but detection accuracy and reliability deteriorate
Solution Approach 1:
The patent divides the detection system into two processing tiers: a pre-processing component that applies simplified rules for quick stationary object identification and removal, and a main processing component that performs comprehensive analysis on remaining detections. This segmentation enables the system to achieve high processing efficiency for the majority of stationary objects while maintaining high detection accuracy for distinguishing temporarily stationary objects through the more rigorous main processing stage.
Solution Approach 2:
The patent applies partial action by using simplified detection criteria for the pre-processing stage, which handles the bulk of stationary object detections (like guard rails and barriers). The main processing stage then applies full, rigorous analysis only to the reduced set of remaining detections. This partial application of simplified methods to appropriate cases maintains overall processing efficiency while preserving detection accuracy where needed.
3Measurement precision
If detailed analysis is performed on each detection to avoid miscategorization, then detection precision is improved, but the volume of data requiring processing increases
Solution Approach 1:
The patent extracts and removes clearly stationary objects (such as guard rails and barriers) from the detection data stream in a pre-processing stage using simplified criteria. By taking out these obvious stationary objects before the main processing stage, the system reduces the volume of data that requires detailed analysis, thereby decreasing the computational burden while maintaining high detection precision for the remaining temporarily stationary objects.
Solution Approach 2:
The patent segments the detection data into two categories: stationary objects identified in pre-processing (which are removed) and other detections requiring detailed analysis. This segmentation reduces the quantity of data that needs high-precision processing, as the simplified pre-processing stage handles the majority of stationary object detections efficiently, leaving only a smaller subset for the more computationally intensive main processing stage.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
The present disclosure relates to a method of detecting stationary objects implemented by a control system (12) of an ego-vehicle (10), the method comprising: receiving (700) sensor signal data comprising stationary and non-stationary detections from a surrounding environment of the ego-vehicle; determining (702) at least one group of stationary detections which meet one or more lateral position selection criteria based on the lateral position of each stationary detection from a direction faced by the ego-vehicle; determining (704), at least one group of stationary detections which meet one or more group regularity selection criteria based on the regularity of the differences in position between pairs of sequentially positioned stationary detections in the group in the direction faced by the ego vehicle; determining (706) one or more groups of stationary detections which meet the lateral selection criteria and the group regularity selection criteria for being a group of stationary detections corresponding to at least one stationary object; and removing (708) the stationary detections in corresponding to at least one stationary object from the sensor signal data output to one or more data processing components (22) of the control system (12).