Temporal Data Association for Autonomous Vehicle Sensor Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in synchronizing data from different onboard devices operating at varying sampling rates, which affects correlation accuracy and precision in navigating and controlling the vehicle.
Innovation Solution
A method and system where a controller onboard the vehicle selects and correlates data from devices like cameras and lidar based on temporal associations, aligning data sets from different devices to improve object detection and classification, and autonomously operates actuators based on these correlations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data from multiple devices operating at different sampling rates are correlated without temporal synchronization, then the system can process data from all devices, but the correlation accuracy deteriorates due to temporal misalignment
Solution Approach 1:
The system performs preliminary temporal association by selecting data from the first device based on the timestamp and the sampling frequency of the second device before correlation. This preliminary selection ensures that only temporally aligned data points are correlated, preventing temporal misalignment from degrading correlation accuracy.
Solution Approach 2:
The timestamp serves as an intermediary that mediates between devices with different sampling rates. By using the timestamp and the sampling frequency of the second device to select corresponding data from the first device, the system establishes a temporal reference framework that enables accurate correlation despite frequency differences.
2Measurement precision
If the system synchronizes data from devices with different sampling rates, then correlation accuracy improves, but the system complexity increases due to additional processing requirements
Solution Approach 1:
The system extracts only the necessary temporal information (timestamp and sampling frequency) from the devices to perform synchronization. Rather than processing all data from all devices simultaneously, the controller extracts and uses only the temporal parameters needed for alignment, reducing the overall processing complexity while maintaining correlation accuracy.
Solution Approach 2:
The system changes the parameter used for data selection from simple sequential processing to timestamp-based selection. By using the timestamp and sampling frequency as selection parameters, the system achieves accurate temporal alignment without requiring complex synchronization protocols or additional hardware, thus managing complexity through parameter-based control.
3Measurement precision
If data selection is based on temporal associations, then object detection accuracy improves, but the processing time increases due to additional selection steps
Solution Approach 1:
The temporal association and data selection are performed as preliminary steps before the main correlation and object detection processes. By pre-selecting temporally aligned data points based on timestamps and sampling frequencies, the system avoids the need for complex temporal analysis during the detection phase, thereby improving detection accuracy while managing processing time through staged computation.
Data Source
AI summary
Systems and method are provided for controlling a vehicle. In one embodiment, a method includes: selecting, by a controller onboard the vehicle, first data for a region from a first device onboard the vehicle based on a relationship between a time associated with the first data and a frequency associated with a second device, obtaining, by the controller, second data from the second device, the second data corresponding to the region, correlating, by the controller, the first data and the second data, and determining, by the controller, a command for operating one or more actuators onboard the vehicle in a manner that is influenced by the correlation between the first data and the second data.