RSU Sensor Fusion for High-Confidence Object-Actor Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current perception systems in vehicles lack a reliable method to establish ground truth and determine confidence levels in object-actor detection and classification, especially for non-V2X participants, leading to uncertainty and reduced accuracy in situational awareness.
Innovation Solution
A system that utilizes J2735 BSM/CAM and J3224 SDSM messages, RTK corrected GNSS positioning data, and sensor data to establish high confidence detection, classification, and tracking by correlating sensor data with ground truth, using machine learning models to determine confidence levels and reconcile perception outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If sensor-based perception is used to detect non-V2X object-actors, then coverage of non-participating object-actors is improved, but confidence level and accuracy of perception deteriorates due to lack of ground truth
Solution Approach 1:
The patent introduces cooperative awareness messages (BSMs/CAMs) as an intermediary source of ground truth information. These messages from V2X participants serve as a mediator between the perception system and non-V2X object-actors, enabling the system to establish confidence levels even when direct sensor-based perception is used. The ground truth data from cooperative awareness messages allows the perception system to validate and calibrate its detections of non-participating object-actors.
Solution Approach 2:
The patent implements a feedback mechanism where the perception system continuously compares sensor-based detections against ground truth information from cooperative awareness messages. This feedback loop enables the system to measure confidence levels by quantifying the deviation between perceived and ground truth positions, and to iteratively improve perception accuracy by learning from discrepancies between predicted and actual object-actor locations.
2Reliability
If consensus-based perception is used to establish object-actor awareness, then reliability of detection is improved through corroboration, but complexity of establishing ground truth and measuring confidence increases
Solution Approach 1:
The patent extracts the ground truth establishment function from the complex consensus-based perception process. By isolating cooperative awareness messages as a separate, trusted source of ground truth information, the system can simplify the confidence measurement process. Instead of requiring complex multi-source corroboration to establish ground truth, the system uses the extracted ground truth from BSMs/CAMs as a foundation for measuring confidence in individual perception outputs.
3Measurement precision
If RTK-corrected GNSS positioning is used to establish ground truth, then position accuracy is improved, but dependency on V2X participants and message availability increases
Solution Approach 1:
The patent applies preliminary action by pre-establishing ground truth datasets from cooperative awareness messages before the actual perception task. The system uses historical and pre-existing V2X data to create a foundation of ground truth information that can be referenced during operation. This preliminary preparation allows the system to maintain position accuracy even when real-time V2X participation is limited or unavailable.
Data Source
AI summary
Provided a Road Side Unit (RSU) transceiver enabled to receive authenticated participant Society of Automotive Engineers (SAE) standard J2735 Basic Safety Messages, RTK corrected GNSS positioning data, SAE standard J3224 Sensor Data Sharing Messages and equipped with a sensor suite consisting of one or more electro-optical sensors, camera sensors, thermal imaging sensors, lidar sensors, ultrasonic sensors, and or radar sensors; the system may include one or more processors programmed or configured to receive data from the system's own sensor suite, reporting transceivers and other network connected devices, to construct high confidence object-actor detection, classification, and tracking perception messages representing the object-actors in the system's field of view for use in other system internal processes and transmit to other external devices for use.