Vehicle Occupant Position Prediction from Incomplete Biometric Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control systems fail to accurately identify the spatial position of vehicle occupants due to incomplete image data, leading to unreliable vehicle control.
Innovation Solution
A method utilizing two machine learning algorithms to enhance occupant spatial positioning by detecting missing biometric data points and predicting occupant movements, enabling precise control of vehicle safety systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image-based biometric data is used to identify occupant spatial position, then occupant identification is enabled, but measurement precision of spatial position is insufficient
Solution Approach 1:
The patent introduces an intermediary imputation algorithm that acts as a mediator between incomplete image-based biometric data and the required complete spatial position data. This algorithm fills missing key points by leveraging correlations between visible and hidden body parts, effectively bridging the information gap without requiring additional direct measurements.
Solution Approach 2:
The patent creates virtual copies of missing biometric key points by imputing their spatial positions based on patterns learned from complete biometric datasets. These imputed key points are essentially copied spatial information that reconstructs the complete occupant model even when parts of the original image data are missing.
2Adaptability or versatility
If static biometric data is used for vehicle control, then system complexity is reduced, but adaptability to occupant movement is insufficient
Solution Approach 1:
The patent transforms the static biometric data system into a dynamic one by continuously updating the occupant spatial model based on real-time image data and predicted movements. The system adapts to changing occupant positions by repeatedly imputing missing key points and updating the control parameters, making the previously static system responsive to dynamic conditions.
Solution Approach 2:
The patent applies preliminary action by predicting future occupant positions before actual movement occurs. The machine learning model forecasts occupant trajectory and pre-adjusts safety system parameters in anticipation of the occupant's next position, enabling proactive rather than reactive control.
3Reliability
If traditional camera systems are used to monitor occupants, then device complexity is minimized, but reliability of spatial positioning is insufficient
Solution Approach 1:
The patent replaces the mechanical limitation of physical cameras with computational intelligence. Instead of adding more physical sensors to capture all body parts, the system uses machine learning algorithms to computationally infer missing spatial information, substituting mechanical complexity with intelligent data processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The disclosure relates to a method (100) for providing control data configured to control a vehicle (11), comprising: providing (110) biometric data of an occupant of the vehicle based on image data; utilizing (120) a first machine learning algorithm, wherein the first machine learning algorithm is at least configured to detect missing values in the provided biometric data and to determine the missing values in the provided biometric data; utilizing (130) a second machine learning algorithm, wherein the second machine learning algorithm is at least configured to provide dynamic occupant data for the occupant based on the provided biometric data and the determined missing values, the dynamic occupant data comprising a prediction of a movement of the occupant; and providing (140) control data configured to control the vehicle (11) based on the provided dynamic occupant data.