LiDAR Tracking Verification Under Occlusion and Multiple Tracks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle tracking systems using LiDAR technology face challenges in maintaining accurate object recognition and driving performance when target objects are occluded or multiple tracks are generated, leading to errors in information output and instability in driving control.
Innovation Solution
A method and system for verifying tracking information that updates verification information based on output data from LiDAR, generates reliability scores, and performs post-processing on abnormal tracking channels, including adjusting reliability, switching to memory tracks, or deleting channels, to maintain accurate object classification and tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If LiDAR-based tracking is used to monitor target objects, then object recognition capability is improved, but tracking accuracy deteriorates when objects are occluded or multiple tracks are generated
Solution Approach 1:
The system performs preliminary actions by maintaining historical track information and prediction models before occlusion occurs. When occlusion happens, the system can continue tracking using pre-established prediction models and historical data, avoiding complete loss of tracking accuracy during occlusion events.
Solution Approach 2:
The system implements feedback mechanisms by continuously verifying tracking information against prediction models and historical data. This feedback loop allows the system to detect and correct tracking errors caused by occlusion or multiple track generation, maintaining measurement precision despite LiDAR limitations.
2Adaptability or versatility
If multiple tracking channels are used to monitor different objects, then object recognition coverage is improved, but system complexity increases
Solution Approach 1:
The verification unit serves multiple functions: it verifies tracking information, generates prediction models, maintains historical data, and performs quality assessment across all tracking channels. This multi-functionality allows the system to handle multiple objects without proportionally increasing overall system complexity.
Solution Approach 2:
The verification unit acts as an intermediary between multiple tracking channels and the central processing system. It consolidates verification tasks and manages track information centrally, reducing the complexity burden on individual tracking channels and the overall system architecture.
3Measurement precision
If tracking information is continuously verified and updated, then tracking accuracy is improved, but processing time increases
Solution Approach 1:
The system performs partial verification by focusing computational resources on tracks that show anomalies or deviations from prediction models. Instead of verifying every track point equally, it applies verification selectively to maintain accuracy while reducing overall processing time through targeted assessment.
Solution Approach 2:
The system changes verification parameters dynamically based on track quality, occlusion probability, and deviation from prediction models. By adjusting verification frequency and depth according to situational parameters, the system maintains high tracking accuracy while optimizing processing time through adaptive resource allocation.
Data Source
AI summary
A method for verifying tracking information may include updating verification information of each of the tracking channels based on output information corresponding to object tracking information of each of the tracking channels, generating reliability of the verification information of each of the tracking channels, generating a verification protocol for verifying an output of at least one tracking function of the first tracking channel among the tracking channels based on the reliability of the verification information, and performing output verification of at least one tracking function of the first tracking channel.