In-Vehicle Semantic Sensing for Accurate Occupant Motion Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently processing and accurately understanding the vast amount of data collected from multiple sensors in motor vehicles to enhance occupant safety.
Innovation Solution
An in-vehicle information processing system utilizing a semantic model to generate an aggregated value from sensing data, including both observation and latent data, to analyze the motion of vehicle occupants, reducing errors through confidence value computation and disparity identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple sensing devices are used to collect data, then the quantity and coverage of sensing data increase, but the complexity of processing and analyzing the data increases
Solution Approach 1:
The patent segments the processing of sensing data by introducing a semantic model that divides data into observation sensing data and latent sensing data. This segmentation allows the system to process different types of data through appropriate methods (aggregation for observation data, extraction for latent data), reducing overall processing complexity while handling multiple sensor inputs.
Solution Approach 2:
The semantic model acts as an intermediary between the multiple sensing devices and the motion analysis system. It generates aggregated values that simplify the raw sensing data from multiple sources into a form that is easier to process for motion detection, thereby reducing the complexity burden of handling data from multiple sensors.
2Measurement precision
If all sensing data is processed in detail, then the accuracy of understanding occupant behavior improves, but the processing time and computational load increase
Solution Approach 1:
The patent extracts only the essential features from sensing data through the semantic model. By generating aggregated values that capture the most relevant information from multiple sensors, the system achieves accurate understanding of occupant behavior without processing every detail of the raw data, thus reducing processing time while maintaining precision.
Solution Approach 2:
The system transforms raw sensing data into aggregated values with different parameter representations. This parameter transformation allows the system to work with simplified data structures that retain the essential information needed for accurate behavior understanding while requiring less computational resources and time.
3Measurement precision
If aggregated values are generated from multiple sensing data sources, then the accuracy of motion analysis improves, but the complexity of data aggregation increases
Solution Approach 1:
The patent segments the aggregation process by distinguishing between observation sensing data (which is aggregated) and latent sensing data (which is extracted). This segmentation simplifies the aggregation complexity by applying different processing strategies to different data types, while still generating comprehensive aggregated values for accurate motion analysis.
Solution Approach 2:
The semantic model serves as a universal processing component that handles multiple types of sensing data through a unified aggregation framework. This multi-functional approach simplifies the overall system complexity by providing a single aggregation mechanism that works across different sensor types and data formats.
Data Source
Figure 1
Figure 2a
Figure 2b
AI summary
Disclosed is an in-vehicle information processing system and method thereof. The in-vehicle information processing system comprises at least one sensing device operable to obtain sensing data in relation to an observation scene. The system may further comprise a processor including a memory having a set of instruction stored thereon, the set of instructions stored thereon retrievable by the processor. The processor may further comprise a semantic model operable to generate an aggregated value in relation to at least one subject observed in the observation scene. In response to the aggregated value generated by a semantic model, the processor may be further operable to analyse a motion of the at least one subject observed in the observation scene. A computer product and a computer-readable medium for executing the computer product is also disclosed.