Track Segment Cleaning for False and Fragmented Object Tracks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately navigating due to segmented and false positive tracks in object detection and tracking systems, which can lead to unsafe decision-making.
Innovation Solution
A track segment cleaning system using a machine learning model identifies and eliminates or combines erroneous track segments, enhancing the accuracy of object tracking and reducing computational resources needed for movement planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If track segments are densely detected to improve coverage, then detection completeness improves, but false positive tracks and segmented tracks increase
Solution Approach 1:
A machine learning model is introduced as an intermediary component between track segment detection and movement planning. This model evaluates detected track segments and filters out false positives and fragmented tracks, allowing the system to maintain high detection completeness while improving tracking accuracy through intelligent filtering.
Solution Approach 2:
The system implements feedback by using detected track segments to train and refine the machine learning model, which in turn improves future detections. The model learns from accumulated data to better distinguish valid tracks from false positives, progressively improving both detection completeness and tracking accuracy over time.
2Productivity
If traditional track filtering methods are used, then computational resources are saved, but tracking accuracy deteriorates
Solution Approach 1:
Traditional mechanical filtering methods are replaced with a machine learning-based filtering system. The ML model processes track segments more intelligently than rule-based systems, achieving higher tracking accuracy while maintaining computational efficiency through optimized inference procedures and selective application of the model.
3Reliability
If all detected track segments are processed for movement planning, then navigation completeness improves, but computational load increases
Solution Approach 1:
The system extracts and processes only the most relevant and reliable track segments for movement planning, rather than processing all detected segments. The machine learning model identifies and prioritizes high-quality tracks, eliminating false positives and low-confidence detections, thereby reducing computational energy consumption while maintaining navigation completeness.
Data Source
AI summary
Provided are methods for track segment cleaning of tracked objects using neural networks, which can include detecting a first track segment and a second track segment. The method includes applying a machine learning model trained to determine if the first track segment and second track segment capture real objects and if the first track segment and the second track segment are representative of an identical object exterior to a vehicle. The method further includes combining the first track segment and the second track segment to form a single track segment having a single trajectory in response to the first track segment and the second track segment being determined to be representative of the identical object. Systems and computer program products are also provided.