Asynchronous Sensor Data Synchronization by Interpolation and Resampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to synchronize data from multiple asynchronous sources, such as sensors in vehicles, which collect data at different rates, leading to desynchronization and complicating analysis and data usage in applications like collision reconstruction and machine learning model training.
Innovation Solution
A method and system for synchronizing data by accessing input video data and telematic data at different rates, generating reduced telematic data by selecting data points based on differences to iteratively-defined reference lines, and interpolating both video and telematic data to a synchronized rate, followed by sampling to produce synchronized data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data from multiple sensors is collected at different rates, then each sensor can operate at its optimal data rate, but the data becomes desynchronized and difficult to analyze
Solution Approach 1:
The patent introduces an intermediary processing system that receives data from multiple asynchronous sensors, synchronizes them to a common time base, and outputs aligned data streams. This intermediary layer allows sensors to operate independently at their optimal rates while providing synchronized data for analysis, resolving the contradiction between data rate adaptability and synchronization reliability.
Solution Approach 2:
The system dynamically adjusts data sampling rates and timing parameters to synchronize multiple data sources. By changing the temporal parameters of data collection and processing, the system maintains synchronization across sensors with different native data rates, enabling both adaptability and reliability.
2Loss of information
If all data points from high-rate sensors are retained, then data completeness is maintained, but data processing and storage requirements increase significantly
Solution Approach 1:
The patent extracts and retains only the essential data points needed for accurate analysis while discarding redundant information. By selectively extracting critical data features and events rather than processing all raw data points, the system maintains data completeness for analysis purposes while significantly reducing processing and storage complexity.
Solution Approach 2:
The system applies partial action by processing only the necessary portion of data at full detail, while summarizing or aggregating other portions. This approach maintains data completeness for critical analysis needs without requiring full processing complexity across all data, balancing information retention with system complexity.
3Reliability
If data is synchronized using traditional interpolation methods, then data alignment is achieved, but computational overhead and processing time increase
Solution Approach 1:
The patent performs preliminary synchronization actions during data collection phases, establishing time references and alignment markers before full data processing begins. This preliminary action reduces the computational burden during main processing by having basic synchronization already in place, thereby maintaining data alignment while reducing overall processing time.
Solution Approach 2:
The system uses efficient algorithms that skip unnecessary computational steps in the synchronization process. By identifying and bypassing redundant calculations in traditional interpolation methods, the system achieves accurate data alignment faster, reducing processing time while maintaining reliability.
Data Source
AI summary
A synchronized data set is generated from multiple asynchronous data sets. An interpolated data set is generated for each asynchronous data set, at a higher data rate than recorded data rates for the asynchronous data sets. The interpolated data for each data set is then sampled at a lower data rate which is common to all the interpolated data sets. The sampled data points are synchronized between the different data sets, such that the sampled data points represent a synchronized data set. Input telematic data can be reduced to inflection points in the telematic data, and interpolation and sampling of asynchronous data sets can be limited to threshold time periods around the inflection points.